Black Hole Movies, Digital Heart Twins, and World Cup Tech

Cold open

0:00Yeah, >> if you took like one of those mini donuts from uh uh Dunkin Donuts >> and you put it on the moon, that's how big those black hole donuts would be. What they've done is create a digital twin, a personalized 3D computer simulation of the patient's heart. Okay, here's how they do it. But like it's also like fairly clear from the video that it hit something, right? How do we explain? It's like it's like it's I don't know like it's it's fairly clear from the video that there's a change in momentum. How is it that the I am you on the ball did not pick it up. Both things can be true. Hello internet. This is your captain speaking Lester Narre

Intro

0:40joined as always by my co-host and our resident PhD Chowdery. We are back for one of our first rundown episodes after all of our specials. If you haven't caught our America 250 or our World Cup episode, be sure to check that. This week, we're going to cover some of our favorite discoveries and stories in science over the last few months while we've been away, including astronomers who caught a super massive black hole reversing its magnetic field. Researchers developing a digital twin for your human heart to

1:20kind of help physicians with diagnostics. a new ultra sonic CT scanner that uses nothing but sound waves and a pool of water. And we have two follow-up stories on the World Cup, including the Norway vers England ball tracking controversy. And we may sneak in a little segment of Are You Smarter Than the Scientists? Oh, and the other World Cup story is the science behind favoritism and homefield advantage, which was clearly not in effect for any of the host nations in this World Cup. As always, we are going to talk about the science from the ground up today

2:01because this is from first principles.

Event Horizon Telescope and M87*

2:13[music]

2:19So, our first story comes from the Event Horizon Telescope. Um, I'm sure you remember back in 2019 they imaged a black hole for the first time and there were so many memes, cat memes. Um, it was the first black hole that was ever photographed. I think the first image actually is the cat meme that I want to show everyone. It's it's the black hole and then as you zoom out it's it's it's the cat eyes. I thought that was pretty funny. It was all over social media because it's such a big deal, right? Um, for the longest time, black holes are things of theory. It's confined to the theory departments in physics,

2:59and no one else really like messed with it. In the late 1990s, Andrea Gz among others started imaging the black hole at the center of our Milky Way. But what I mean by that is they were really imaging the stars around the black hole, right? And they were they were looking at the star trails and you could see that there was some mass that was massive, millions of times the mass of the sun that was ringing these stars around in these capillarian orbits. And so there had to be a black hole in there. But to actually image a black hole, right, where you've got an event horizon, you've got the accretion disc around it, that is going to take an insane amount

3:40of technology and an insane amount of coordination from the entire globe. So this is a follow-up story to that 2019 photograph where now you can imagine in 2019 they photographed for the first time the black hole. It's been about 7 years. You can image

Making a movie of a black hole

4:01over and over and create a movie >> of a black hole of the accretion disc revolving around a black hole and all of the weird physics that happens around that black hole. Right? So now you can start testing the theories and it's it's such a cool thing that we are able to do because of all the technology that we have. So if you start from the beginning right the most of the photos that you see of the black hole like that cat meme that is of M87 which is this galaxy 55 million lighty years away. It was first discovered in 1781 as part of the Messier catalog. Charles Messier was a French comet astronomer and he was basically like trying to hunt for

4:43comets. That was his big passion. And when you hunt for comets what do you do? You basically point the telescope at the sky and you look for fuzzy objects. But those fuzzy objects have to move. So in order to distinguish the ones that are moving from the ones that are not, he created a catalog of all the all the things that he should be ignoring. And that became the Messier catalog, which is now much more important than all of his comet. >> That's so >> research. You know, he was trying to basically find comets. And he he made a list of stuff to ignore. like you know when you're like coding or something and it's like ignore this flag, ignore this flag. That that that was him with the

5:23entire night sky. He created this giant catalog and turns out he cataloged I think something like on the order of hundreds if not maybe a thousand objects that are now like nebula galaxies other things that are close to the Milky Way and close to close to the sun. This is like a classic story in science which is kind of like our Hubble story with the deep field we did a few episodes ago which is let's point this at a >> a space of nothing >> and see what happens and and everyone's like what and it became one of the most important things ever and it's like sometimes you know >> you stumble upon discoveries and just being serendipitous is really important in the process because it it it does

6:04still again in this case it's a little bit annoying because he's like I I cared about this thing and I'm known for this other thing. But it's like you still did the work. >> No. And and I I don't think he would mind if he was alive today that all of these objects have the prefix M and then and then and then the object right in front like right everybody knows M87 in >> astrophysics because this is the closest large super massive active galactic nuclei to the Milky Way. It's only 55 million lighty years away which is actually quite close. Like the Andromeda galaxy is 2 million lighty years away. So this is only like what 20 to 30 times as distant as the Andromeda galaxy. So it's really in our local, you know,

6:45we're talking about tens of millions. The universe is tens of billions, right? So this is this is very local to us. It's it's in the constellation Virgo. And um you know to your point it's it's one of those things where cataloging negative results was actually a good thing in science everybody if you want to publish in science and nature and cell you have to come up with results and positive results. It's like oh I did an experiment I had a hypothesis the hypothesis worked here. He was just cataloging like negative things because he thought these are the stuff that I want to ignore. And it turns out the negative results in this particular scientific endeavor was the most

7:26important thing that he did in his life. Right. Kind of cool, >> right? >> Um so anyways, it turns out to be um one of the largest galaxies in our neighborhood and um it's the most studied super massive black hole in history. [snorts] If you go back to that photo, actually photo number two. >> Yes, photo number two. Notice you've got the it's an elliptical galaxy, so it doesn't have a lot of structure, which means it's full of old stars and it's massive. Okay? It's not a spiral galaxy in the sense that the Milky Way or the Andromeda galaxy is. And also, there's this giant jet. It's it's like a it's like a hose coming out of that center, right? >> Mhm.

8:07When people saw that using the radio astronomy that was built up after World War II, people were like, "That's really weird." Because one, you can actually measure the speed of that jet and it's very, very close to the speed of light. At some point, people were actually saying super luminous. It was higher than the speed of light. Turns out it was like it's not an error, but there's like some weird physics going on that like makes it seem like the stuff is moving faster than the speed of light, but really it's like kind of a group velocity type thing. Um the the actual stuff is not actually moving faster than the speed of light, but maybe some phenomenon is. Um

8:48when when people saw that, they were like, there's got to be something really energetic at the center of that thing that is causing a giant jet to form. That's like many light years in size. >> What what is creating the the inciting what is creating the force that is enabling us to image this particular situation? >> Yeah. Like if you've got a hose, right, and and the hose produces a jet that's light years long, the the hose has to be insanely strong. And so that's where we get the idea of active galactic nuclei. the idea that there could be a super massive black hole at the center that is creating so much energy from accreting

9:31matter into the black hole that that heat and that magnetic field and that plasma is shooting out a bunch of material at this incredibly high speed. So this [clears throat] is the first sort of indication that yes black holes are real and perhaps um we can study it in a really really precise way. Okay. Um then in 2019 we got two images these amazing images that popped up all over social media took over the entire globe. On the left is the black hole of M87. The center is probably where the event horizon is. Okay. That's where >> the the dark spot in the middle.

10:11>> Yeah. That's that's really what the black hole is. But obviously a black hole is somewhere is is a region of space where nothing can escape, not even light. So if light can't escape, how do you image it? Well, what you're really imaging is the accretion disc around the black hole. And I want to take a quick moment for our audio listeners. You know, many of you will have seen this image. It looks like a glazed donut, right? You have that orange glow, the black center, and and it was very big in this time period. But that that is the image we're looking at two, but that is one of the two that we're looking at. And then you'll get to the second here. >> Exactly. Okay. And and so the one on the left is the one at M87. That's 55

10:53million lighty years away. >> Okay. >> The one on the right >> and and you can see the size of that relative to the one on the left. Right. The the the one on the right you can see the inset. It it it would fit very neatly inside the black hole inside the event horizon of M87. The one on the right is our puny little black hole of the Milky Way. >> Ah okay. Yeah. >> Okay. >> Yes. And it's it's quite amazing that the M87 black hole is so much bigger than the Milky Way, >> right? We're it's probably in the way these images are scaled, we would say maybe it's >> two and a half to three and a half

11:33times. >> Yeah. >> Scale-wise, >> I will say that this the the R is is I think it's uh is it Sagittarius A? >> Sagittarius A. >> Sagittarius A. Kind of looks like a bagel. Yeah. >> A little more bagel shaped. >> Yeah. Compared to the donut shape. [laughter] Yeah. Yeah. Yeah. And and that's probably um >> it's probably a feature of the fact that M87 is so big >> that like the accretion disc is like more uniform and it's like farther away so you're not getting the nitty-gritty details. Um now the relative size of these photos you can see that like you know in terms of the relative size they look about the same. >> Yeah. Yeah. Yeah. Yeah, that's because in the sky, relative size-wise, they

12:15look about the same. It's just that the Milky Way black hole, Sagittarius a star, by the way, it's not Sagittarius A. The star means that we're talking about the the black hole. That's that's always like that's all the M87, you see the M87 star. So, that means we're talking about the black hole. >> And when we say star, we we mean asterisk. >> Asterisk. Yeah. Yeah. But in in colloquial language, >> 100%. Just for people who might be viewing it, that's what we're referring to as the asterisk. >> Exactly. Yeah. So whenever you see like that, that means we're talking about the black hole at the at the center of the galaxy. And the two are about the same size in the sky. So this is how how big they would look if we could like zoom in zoom in zoom in. >> It's just that the Milky Way, even though it's so much smaller, yes, >> it's so much closer to us. It's only

12:56100,000 lighty years, right? Compared to the 55 million lightyear. So even though the M87 is actually way bigger, it would it could fit the entire Milky Way black hole inside because it's so far away. The size of it looks about the same. So now that brings me to my second point.

How small is the black hole image?

13:12>> That photograph is really really small >> in the night sky. Okay, that thing is about 50 microarch seconds across. So let's get to how how actually big that is. Right. >> So um >> this is from the Babylonian time. >> I want to quickly 40 micro, not 50 micro. >> Oh yeah. Well 40 to 50. There's like a scale bar in there that says 40. It's like it's nebula. It's like, okay, if I measure this way, it's like 60. If I measure this details, details, you know what I mean? No. No. Totally fair. Totally fair. >> Yeah. But at the end of the day, it's it's something like tens of micro arcseconds. Yes. Across, right? >> Yes. >> Now, how big actually is that in the

13:53night sky? So, um, astronomers use the degree system that we have inherited from the Babylonians and they just like took that to the >> Shout out to the Babylonians. >> Yeah. Shout out to the Babylonians. way back they said a full circle is 360°. So the astronomers said okay a full circle in the night sky is 360°. So a a single degree is 1 360th of a full circle. That makes sense. The moon is about half a degree which means it would take 720 moons lined up to make a full circle. Does that make sense? Right. So the moon is about half a degree. Each degree has 60 arc minutes

14:33and each minute has 60 arc seconds, right? The same way that we use hours and minutes and and seconds in like normal timekeeping. So that's the idea. You've got um you got you got half a degree. So that's 30 ark minutes is the moon. >> Mhm. >> Right. Because an hour would be a full degree. >> Yes. >> Okay. So the moon is 30 ark minutes which means it's um 60 arcsec to a minute. So that's 1,800 arcseconds. >> Yes. >> Right. So imagine taking the moon and dividing it up into 1,800. That's that's a single arcsec. And then now a micro arcsec is a millionth of that.

15:13>> Okay. To give you an idea, you see that donut? You remember that donut?

A donut on the Moon

15:17>> Yeah. If you took like one of those mini donuts from uh uh Dunkin Donuts >> and you put it on the moon, >> that's how big those black hole donuts would be. >> Jesus. >> Does that make sense? That's crazy. [laughter] >> That's crazy that that's what we're imaging. Like if you if you want to put it into terrestrial terms and and just you're saying we put a munchkin >> Mhm. >> on the moon. >> Yeah. And then we imaged it >> from from here from ground and we're looking at a munchkin on the moon. >> Yeah, >> that's what we're talking about. >> That's what we're talking about. And if you want to put it into terrestrial terms, um it's

15:58like an ant >> on the Empire State Building viewed from Los Angeles, >> which for those who are not familiar with the US geography is >> that's across the continent. >> It's across the continent. >> Yeah. [laughter] Yeah. Okay. or like or like from New York to Paris. Okay. You could do you could do that if you wanted, >> which is so I think the point you're trying to point out here is the level of precision over the scale of distance is outrageous on its face. >> Yes. On its it's it's incredible. And Right. And the thing is that donut, it wasn't just a donut. We could see details in the donut. So the resolution we're getting is like insane. Right. We're getting sprinkles on the donut.

16:39Like which side has sprinkles? Is it Boston cream? Is it sprinkles? Like, is it what is the one that has the little crumpled edges on it? Like, >> oh, I hate those. [laughter] >> I only like glazed. Plain plain glazed. >> But it's it's it's incredible that we're able to do this, right? And to give you an idea, right? >> Um, we've actually covered the um the tyranny of resolution when it comes to astronomy plenty of times on this podcast. Um there's a fundamental physics issue with trying to resolve stuff. Okay. Yeah. Um light at the end

The Rayleigh limit

17:14of the day is a wave, right? And so so if I have two, let's say I have two light bulbs that are right next to each other, if they're right in front of my face, of course I can discern them, right? Because my eye acts like a tiny little telescope and it creates an image in the back of my retina. And if the light bulb A goes to a certain part of my retina and light bulb B goes to another part of my retina then I can discern that there are two light bulbs because these rods and cones are saying hey I'm seeing a light bulb over here. These rods and cones are seeing saying hey I I see a light bulb over here right and so the two the two light bulbs are not overlapping in my detector and

17:55therefore I can discern the two. Now there there comes something called the rally limit which is if I take those two light bulbs those two points of light let's say tiny little LEDs and I take them really really far away all of a sudden they merge into one LED why >> right >> why because because the LEDs the the image in my retina is now getting overlapped because of the wave nature of light my pupil is only this big and so the the the the wave nature of light is going to like defract in my pupil and that's going to cause a blur uring effect that is just purely due to the wave nature of light. It's there's nothing I can do about it. There's no like oh put it in a vacuum, put it in

18:36cold, nothing because of the distance and the fundamental uh uh the fundamentals of how light travels over distance. As you increase that distance away from our own detectors, which are our eyes, >> the farther the distance away, the more overlap will happen between two independent light sources. >> Exactly. >> Uh they could be very far away. >> Yeah. >> But the farther you go away, they become it becomes hard to discern that there's a gap between these two. >> Yes. And so how do how do we how do we do anything, right? Um well, we make the the the detector bigger, right? We make our telescope bigger. The eye is a terrible detector. It's only about like

19:17less than a centimeter in diameter. Um that's why we have large telescopes. The larger the telescope, the the smaller the angular resolution that you can have, right? Um or I should say the greater the angular resolution, the smaller the angular distance you can discern between two independent points. >> Which is why we continue to Why do we need to make bigger and bigger telescopes, right? Because the bigger we make them, this is why, for example, between Hubble and JWST, we have fuzzy versus less fuzzy. There's different wavelengths in there and there's details. Yes. But fundamentally speaking, >> we have giant Mellin that's coming out versus regular Mellin. Why do you need

20:00to make a bigger Mellin? Because we like 4K TV. Yes. >> We don't like 720p. No, >> we don't like fuzzy CRT TV. We want to see every blade of grass. Exactly. >> We want to see every tear drop that comes from the face. And you can get that from making your detector larger in order to resolve light to more of a level a better level of granularity. >> Exactly. And if we're not going to settle for 720p on our TVs, we shouldn't settle for it [laughter] >> in astronomy. You know, you know what I'm saying? >> Yes. >> Come on, guys. 100%. >> Um, so, so the rally limit, that's that's that's the fundamental trade-off that we have. Okay. Now there's two ways to make your resolution better. Okay. Um

20:42the the the the smallest angular resolution that the smallest angular distance that you can discern is given by the rally limit. It's effectively the wavelength of the light divided by the diameter of your aperture. The diameter of your telescope. This is something that we've been over a lot, right? And it kind of makes sense. The smaller the wavelength of the light, the less it's going to like bend around your detector. And so you can maybe discern smaller and smaller things. The bigger the detector, the smaller the thing you can discern. That's why the detector is in the denominator, the wavelength is in the numerator. Okay? So suppose we go for like radio wavelengths like millimeters wavelength. So it's not it's not quite radio, it's like millimeter. So it's it's um it's like shortwave radio. Okay.

21:24So about like 1.3 millimeters. Um and I'll get into why we want to do radio and not optical which is at the hundreds of nanometers, right? Um so suppose we

Why we need an Earth-sized telescope

21:34want to do radio. um wavelengths and we want to image um a donut on the moon. Okay, we can calculate okay that's uh we want to image at the level of like single arc single micro arc seconds um plug that into the theta of the rally criterion the lambda is 1.3 mm so how big should my diameter be the diameter be turns out to be 10,000 kilometers [laughter] that's it's quite large >> quite large um what it's saying is in order To image a donut on the moon with radio

22:16wavelengths, I need a detector, a radio telescope the size of the Earth. >> 10,000 km. Take it or leave it. That's what physics says. >> Let's get a Dyson sphere going, baby. [laughter] >> Well, these guys created like a like a telescope sphere around the Earth effectively. Okay. The next video shows you how they did it. They said, "Okay, this this is what it is, right? We need a telescope the size of the Earth. >> Yes, >> we can't build a single telescope. What we're going to do is rig up a bunch of radio telescopes all across the globe and they're all going to communicate with each other and take independent images of this thing. Now, our light gathering power is not going to be that

22:57high because the detector is not the size of the Earth. Right. >> Right. But our resolution capacity is going to be the size of the earth because I'm able to take light from this part of the detector and this part of the detector. So there's um there's um two detectors in Chile, there's one in um Tucson, there's one in Hawaii, there's one in northern Africa, and as the Earth rotates, you use the Earth's rotation to image the black hole. >> Isn't that sick? And and and so this is

Radio interferometry

23:29what is so brilliant about this is how resourceful it is, which is we're saying we don't have Dyson spheres. We we're not going to build a >> 10. We're not civilization stage 4, whatever, >> whatever it is, right? >> In the comments, tell me what it is. >> We don't have enough political social collaboration to pull something like that off at this moment in humanity. So we say okay we have these detection platforms all over the surface of the earth and because they are covering the surface of the earth in these different latitude longitude positions we can image the same object at distance. >> Yeah. But from these different points on

24:09the planet. >> Yeah. >> And then we can basically blend >> Yeah. >> the results from that to create what would effectively be the same thing. Yes. >> As having a 10 km wide singular device. You would have to take more time because you don't have as much light gathering power, right? Like you don't have as many photons coming in cuz you've only got these individual detectors. If you had something the size of the Earth, then you could just boom take an image. But we're not we're not we're not making, >> right? We're not we're not harvesting all of Mercury's mass to create a giant radio telescope. Maybe in our future as humanity, we will do that. We will destroy Yeah, we will destroy the planet of Mercury [laughter] and create a giant

24:50telescope. Um I'm sure there's going to be people that really don't want that to happen, [laughter] right? Mercury retrograde is a very big thing. >> It's very popular. It's very I heard with the horoscopes. Yeah, it's it matters. >> But that's effectively what's happening. I mean, you've been to the very large array in New Mexico, right? That's the same thing. It's an interferometer. That's what this is called. This is called radio interferometry. The idea is to use interference between all of your detectors to create an image with an exquisite resolution that is that is effectively given by the size of how far apart your farthest telescopes are. >> Right. >> Right. And in this case, it's the Earth. Um, one more thing I want to point out

25:30with that with that video actually, if you don't mind. Um,

The South Pole Telescope

25:33>> so notice the Earth is rotating. Right. And this is from the vantage point of um M87. Okay. So it's like it's like I'm at M87 looking at the Earth. Yes. >> Now the Earth is tilted. >> Yes. >> And so notice Antarctica is always facing >> M87, right? Because because the Earth is tilted. So Antarctica is like always facing the constellation of Virgo, right? So the the the telescope in Antarctica on the South Pole is called the South Pole telescope. It's just always staring at M87. [laughter] And then and then the other guys are like are like picking and choosing like the filling in the data >> at time in the day things. >> Yeah. But it's just cool that like no matter what part of the year, right, because the Earth's tilt doesn't doesn't

26:13change. Yes. >> So no matter what part of the year, it's just going to keep staring at that spot in the sky. And fortunately with radio astronomy, it doesn't really mind staring at the sky during the daytime. >> Right. Right. So it's the the sun isn't really affecting radio astronomy as much unless optical optical optical obviously we can't see any stars during the day. Radio if you look you'll you'll be able to see stuff. >> We have our atmosphere and there's blocking happening and light and it's interfering. >> Yeah. Like the the sun is still a source of radio waves don't get me wrong but um it's not like getting scattered as much by the atmosphere as compared to you know like optical light. So, as long as you don't like just stare directly at

26:54the sun, you're good to go. >> Um, I just think that's it's such a cool thing that we've done as humanity. >> It's also a cool thing that we've figured out, right? Because you could just be like, "Oh, we can't build a 10 km wide telescope. I'm going to give up because this problem is insurmountable for us because we're not going to have the engineering capacity, the funding, and whatever." And it's like this is I think one of the beauties of the process that we talk about so much on the show is how can you do more with less? >> Yeah. >> And the ingenuity that arrives from if we can't do X, can we maybe do Y ultimately to get the same end result? And this is a perfect example of how,

27:34you know, we're not going to get the funding to do what we want to do, but we already have the resources to be pragmatic in getting to the same end result, which is this imaging task. >> Exactly. >> Uh which has given us this M87 um this beautiful, >> beautiful, beautiful image. >> Beautiful image. >> Right. And so now the other point that I just want to make is why radio, right? If we can do this with like just light at the end of the day, right? Light is light. Why don't we just like wait for the nighttime and then image with optical, >> right? Because the argument would be like some of our optical stuff is pretty tough. >> It's pretty tough, right? But we don't

28:14have optical interferometers. Not really. We've got a few >> which is meaning more like this large array of optical telescope optical telescopes. We have a lot. So the VA being a larger very the very large array of radio telescopes which are lower cost right uh and maybe you could argue lower data stream complexity and downstream

Why radio works better than optical

28:38>> that's the main one okay >> it's the data stream not the complexity in the sense of like computational complexity >> volume >> uh no in terms of literal physics of um radio is slow >> ah okay that makes sense that makes sense radio is slow the radio wavelengths are longer which means the cycle of the light is slower. Okay. Now, here's the key thing with interpherometry. What I need to do is interfere the light signals, which means a wave is coming in over here, a wave is coming in over here. I need to know the exact timing of these waves so that I can add them up >> and destructively and constructively interfere them. Mhm. >> Now with radio astronomy like something

29:18like the VA right um the you're looking at about the the cycle the cycle between these things is going to be about 10 -12 to 10 - 10 seconds. So like um you know 09 or 10 zeros and then 1 second is going to be the cycle time of this thing. Okay. This is in the this is in the megahertz to gigahertz range. And that's something that we know how to do, >> right? Radio, like literally the radio in your car, right? We know how to tune stuff at that range. Okay. >> Um and what that means is we've got the timing just right that we can add them up and subtract them and so on and so

29:59forth. >> Um so what you can do with radio astronomy is you can take data at the South Pole and take data in Spain and take data in Hawaii and stuff like that. rig it up to a clock that is accurate enough and then when you put it all together with like your pabytes and pabytes of data, you have the tag of like when the signal came and you can you can you can match and you can do the computational >> you can make the math math. You can make the math math, right? With um with optical now, you run into the block that the cycle time is faster than your clock. Like the light that's coming in is faster than the reliable clock that you have to record the thing. So you

30:41can't actually line up the signals. >> That makes it makes total sense, >> right? And this is why so there is a

Optical interferometry and atomic clocks

30:47there is a famous um optical interferometer in Chile called the very large telescope VLT compared to the VLA but and what they do in order to do the interference they don't do computational interference they literally have tunnels underneath the ground that are routing the light >> physically >> so that the light physically interferes >> right >> right because that's like okay you can't you can't record the data and do do it computationally hardware is king, right? And even when it comes to astronomy, hardware is king, right? >> As we've discussed multiple times on the show. >> And so VT does it that way. But if if I wanted to rig up a a optical array

31:27that's the size of the Earth, currently it's not possible because we just don't have the timing complexity. Now, there are optical clocks that are getting to that level of accuracy, but like now deploying it across all of these telescopes and it it's it's a challenge, right? because it's still a very new technology >> but this is just to be clear this is an engineering problem >> not necessarily a fundamental science or physics problem at this point like we would know we we have potential paths well identified >> yes the potential paths are well identified I think realizing those paths there still might be some physics issues okay you know like like creating an optical clock that um that that is

32:08stable enough like I don't know much about and I think we're going to do future episodes on optical clock right this is a Comment below if you want this one because I'm because we we people always talk about oh we have atomic clocks but what we're saying is we they're not enough atomic clocks. Yeah. There's not enough decimal places. >> Yeah. Yeah. And if there aren't enough decimal places it's like only in Boulder, Colorado, right? There's like there's like a few groups in Boulder, Colorado that have this thing working. >> We have it for the nukes [laughter] but not for science. Let's make let's make the clocks for nukes work for science. That's what we want. So, so and and I I do also want to shout out there's um there's a team at Mount Wilson that we covered that also had um uh an optical

32:50interpherometer and they're the guys that like found sunspots on other stars. We covered that episode, but they also do the hardware thing. They've got like a tunnel with a bunch of lasers and like they're they're interfering it optically. Yes. So, so doing this which is really computational challenge we haven't been able to figure out for atomic for um optical >> optical telescope >> which is which is and the the description is really important because basically the what we're saying is radio waves travel through space >> slower >> than light does which is what we say when we mean when we say optical and so we we we have a >> no I should be clear radio waves travel at the same speed as optical the cycle

33:31the cycle the cycle by which >> the the wave like goes up and down slower slower, >> right? The speed of light is the same there. People are going to come at us. >> No, no, no, no, no, [laughter] no. That that's a fair distinction and excuse my excuse my misnomer there. That's a fair distinction. And like as a crude analogy, it's like we're a police with a police radar traing people go by, but the the rate at which we can see the up and down go is is too slow for the for the Dodge Charger, whatever the the Dodge the Bugatti. >> Yeah. Right. Right. The Dodge Charger. It's going to be fine. >> Wait, no, wait, hold on. Some people are going to be mad cuz there's there's the Trackhawks. There's some label of them that are fast, but the point >> Yeah, you got to like rig it up. [laughter]

34:11>> Um, but but we we just are not able to move as quickly to have enough enough >> uh discrete granularity from a measurement and timing to be do anything valuable. Yes. And that's kind of the And once we could be able to do that, it's it's a very big deal. Um

The future of high-resolution astronomy

34:30>> because then we can just res we can start resolving like crazy things. Not just the black hole, right? But like just imagine like like um resolving features in Andromeda. Like all of the nebula that we have like the the Horsehead Nebula, the Orion Nebula, the Pillars of Creation, all of that is in the Milky Way, right? Imagine being able to resolve those types of star creating regions in Andromeda for example and seeing like similar structures uh in M87 even you know that I think that would be it would be insane. So that's like kind of a frontier of astronomy that is um that is just like ripe but but it it involves like physics and engineering to

35:11happen first. Um anyways it was a long sojourn. So now we've got the event horizon telescope 2019 it came out with that photo but it's been seven years. So it's been taking data this whole time >> and now finally >> it has released >> a video of our black hole >> which >> the M87 black hole >> which looks beautiful. >> Yeah, this is a video this is a video over 6 years of the black hole at the center of the M87 galaxy. >> This is so outrageous. >> Billions of light years. This is so outrageous. >> I mean, sorry. Billions of solar masses. >> Right. Right. >> Right. Um 55 million lighty years away.

M87’s changing magnetic field

35:50>> When you say a solar masses, >> yeah, it's the sun. >> Our sun. >> Billions of suns. >> Billions of our sun >> inside a black hole that is like accreing. And all of those lines that you see, this is this is what's um what's very special about this new finding. Those lines are the polarization of the radio waves from that part of the accretion disc. Now, why is that important? Polarization is a property of light that has to do with the fact that, you know, light is a wave. So, well, it could oscillate this way or it could oscillate this way or it could oscillate anywhere in between, right? There's a there's a 2D degree of freedom. >> Mhm. >> The direction of the polarization tells you something about the direction and

36:31the strength of the magnetic field. >> Right. >> Okay. And black holes have intense magnetic fields around them because you've got all this plasma which is a bunch of charged particles. Charged particles moving around Maxwell's equations you get giant magnetic fields. The giant magnetic fields are the reason why M87 has that giant jet that we saw earlier. And so and so studying that magnetic field is extremely important for figuring out the properties of not only the black hole but also the properties of elliptical galaxies in general. Um, I don't know if you remember the Carnegie episode that we had. Yes. With Michael Blandon. >> Yes. >> Um, we asked him, "What's the one like question that you would want answered

37:12>> at the towards the end of the episode?" 100%. >> And he said, "I'd really like to know why these large elliptical galaxies and like these large old galaxies don't create any new stars." >> He he was like, "For some reason." >> Yeah. For some reason. >> For some reason. And there's like [laughter] there's like 10 different theories and everyone's got an idea but no one knows this happens. >> Yeah. >> But it's just for some reason. >> Yeah. And every single one of those theories, the culprit is the black hole. Okay. The black hole does something and then and then and then all of the stars stop forming. No new babies in the galaxy because of the black hole, >> right? um studying the black hole like M87 is going to give us clues as to

37:55whether one theory is correct versus the other. So I mean first of all just like the technological achievement of creating this video of a of a black hole that's 55 million years 55 u million light years away um is is insane. You can see the magnetic field changing. It's even flipping some which is kind of crazy which means that like >> you know the stuff in the accretion disc is moving at near the speed of light >> and the size of that accretion disc is about where like if the sun was at the center Voyager would be where the main donut is. So it's like it's like like outside the the realm of Pluto, but it's

38:35not that big. Which means if this stuff is going at near the speed of light, this is a highly dynamic system where we are going to see changes on a yearly basis. And if we can get that time down to even monthly, we are going to see changes on a monthly basis. >> It the system is highly turbulent. It's it there it's it's a highly active >> relatively maybe not dense is the right word. >> No, it's [clears throat] of course it's dense relatively like like like dense and the amount of activity that's happening. We've never been able to visualize in motion. >> No, we never have in motion. >> Right. And most things in astronomy just stay put. They like the pillars of

39:16creation that were photographed by Hubble and then we photographed with James Webb. They look the same, right? And [laughter] it's like 20 years later, but they it looks about the same because the stuff is not moving at relativistic speeds and the size of the stuff is like light years across. But here you've got something that's like light hours to light days and the stuff is moving at light speed. So you're going to have a lot of motion. It's extremely exciting for theorists, for astronomers, for people in AI that are like like putting algorithms to the test with like how to make this data better because you've got pabytes and pabytes of data, right? It's um it's a it's a really cool cool thing. >> If you're looking for motion in the

39:58universe, check out M87 because that's where the motion is. >> M87 has motion, as the kids would say. Yeah, it's it's I think it's it's it's an incredible story. Okay, so moving on

Digital heart twins

40:11to our second story. >> Yeah, this is a good one. So, that was our first one on on the sort of magnetic field related to black holes and our ability to actually see it. Now, >> yeah, >> we're moving on to our digital heart twin. >> Yes. >> Uh not to be confused with the FIFA twin that's created when you get scanned in for V. >> The V the very not truthful. Um, so we talked about in our previous episode, you get every player gets scanned into this FIFA system. So when they do the V animations, it's not just a random stick figure. It's their actual scan. >> Yeah. >> We're doing something not dissimilar for our hearts. >> Not dissimilar. Exactly.

40:52>> Uh for for some medicinal benefit and to assist with how physicians do different diagnosis. I'm actually very curious about this because obviously in recent news unfortunately we've seen um the passing of Senator Lindsey Graham and he seems to have had a cardiac related um uh cause of death for those who are not looking at conspiracy theories. >> Oh, okay. >> And so there's this this >> my Instagram is only conspiracy theories. No, I'm just [laughter] kidding. But yeah, >> like so precision medicine for the heart. >> Yes. being one of the number one and the number one and two and three maybe leading causes of death in America at

41:33least >> definitely >> hugely hugely beneficial. >> Yeah. And I think this is a really cool um sort of symbiosis between um data and simulation and like actual practice in medicine. So the digital twin that they're creating this is out of John's Hopkins University um precision medicine for the heart. What they've done is effectively create a digital twin for the heart to try and solve one of the most challenging conditions in medicine, which is um ventricular tachicardia, a VT. It's basically um you've got like a

Ventricular tachycardia and ablation

42:07fast heart rhythm that originates somewhere in your heart, usually in your lower chambers, okay, in the ventricles like the two that are on the bottom. And historically what doctors do when they want to treat this is they stick a catheter and they do a catheter ablation. So basically you thread a wire into the heart into the pulmonary arteries like the the arteries that are around the heart that are keeping the heart going right because the heart is a muscle so it needs blood. The blood comes from these veins and arteries. So you thread a a a wire through these arteries and you effectively try to look for where the arrhythmia is happening. Okay? And it's like it's like a needle

42:49in a haststack type of problem. You're trying to find this thing and then when you finally find it, you burn you ablate a tiny patch of that tissue and that hopefully stops the erratic signal, the erratic electrical signal. >> You're trying to disturb the system. >> Yeah. Yeah. You're you're trying to just like there's like some cells in there. There's like some like pacemaker neurons or like whatever neurons that are part of the contracting of the heart that are like doing weird things. You burn them off. Yeah. >> You kill them and then the rest of the heart kind of just like moves on. >> Okay. Um it but as I said, it's like finding a needle in a haststack, right? So this new procedure um it was published in the New England Journal of Medicine. We love to see it.

43:30>> Yes. Still waiting for our subscription. >> Yeah. I I still um had to [laughter] had to had to call in favors from my doctor friends to get access to this this um journal article. But at the end of the day, the procedure that usually takes hours cuz the doctors are literally proddding and looking for where this thing is happening. Now, they've brought it down to 30 minutes. >> And if the doctor misses the exact spot, this thing finds it. What they've done

Building a personalized heart simulation

43:56is create a digital twin, a personalized 3D computer simulation of the patient's heart. Okay, here's how they do it. They first take a bunch of really highresolution MRI scans and they create a 3D digital twin of the patient's heart. So, every patient gets his or her own digital twin. Okay? Then what you do is you reconstruct that in your software. Okay? All of those 2D images from the MRI scans become this 3D really high resolution image and then you simulate what is happening in that heart because now you know enough about heart cardiology, about heart physiology,

44:37about how these neurons are working and the model is then assigned these electrical properties and this the the researchers pace that digital heart in the same way that the pacemaker cells create this electrical gradient that causes the heart to contract and then and then you know push out blood to the lungs, take in blood from the lungs, so on and so forth. And what you can do here is you can do a virtual surgery. So here here's what's happening in in the video. Okay, >> that's fascinating. >> So you've got you've got the heart. You've got you've created this 3D model. Now you're simulating the electrical signals in that heart and you're trying to figure out where just according to

45:19the physiology of the heart there should

Virtual surgery before real surgery

45:22be that arhythmia and there you see it in the VT that circle that you see you're seeing the current kind of loop >> around that that center. It shouldn't be doing that. >> It's the eye of the hurricane. >> Yes, it's an eye of the hurricane. It shouldn't be doing that. There should be no hurricane, >> right? It should be cleanly diffusing and creating the sort of contracting that we all know and love that creates our that like keeps our blood going. Right? So in the simulation you can identify exactly where it's happening. >> And then in the simulation you can ablate that part >> and see what the result >> and see what the resulting pattern of electrical stimulation is going to is going to do. Fascinating, >> right? And if you ablate that part, is it going to work? If it works, then

46:03finally you go to the surgeon and you're like, "Yep, we've got a best strategy out of all of the different strategies that we've tried. We've got a best strategy that we found on the computer. Now you're the surgeon. Go in >> and just replicate >> and just do it. >> This is this is >> 30 minutes instead of hours. >> This is and and I think what we're sort of saying is particularly for having a patient under during that time period where you have to be searching doing search and rescue. You no longer have to have the patient under to do search and rescue. You've theoretically drisk >> but it's like way lower the the d-risk

Trial results

46:37it's it's way lower you you're you're you've you've been able to do enough pre-work here where the the the patient risk profile is also lower and it's so fascinating because we've been studying the heart >> vigorously for a very very long time now and so we have a lot of knowledge >> as it relates to this particular organ >> and it builds This is built on all of that knowledge out of thin air, right? And I think that's a really important point here, which is like we can do things like this now because we've spent so much time. >> Yeah. >> Gathering data about the heart, building simulations around the heart and then so you can now go to an individual patient

47:18and then apply all of that that history of knowledge, create a personalized solution to them and and basically run the surgery in in what uh in in uh in silica. Yeah, exactly. Uh ver before you go >> invivo. Invivo. [laughter] >> Yes, that's exactly right. >> That's incredible. >> Yeah. And in in a 10patient trial, so this was all FDA approved, by the way. Um the the trial was FDA approved. And in the 10 patient trial, all participants were free of the sustained arhythmias at the follow-up. >> Incredible. >> And that's far exceeding the typical rate, which is like 60% success rate, right? So for 10 people to have success like at 60% that's 6 to the 10th power

48:01which is point about 06. So like that happening by chance among like a normal w with normal procedures is like um less than 1%. But here it was happening right >> 100% of the time. Yeah. Every time. >> Exactly. [laughter] So here you don't even Yeah. This isn't you can't really argue on it. The other thing that was crazy is that um eight of the patients were entirely off anti-arithmia medications. They didn't even have to take medications. >> That's two of them had lower doses, but eight of them were just like done. >> We're good. >> Yeah, >> they did it. And I I you know, I think one of the things technology understandably is very complicated in our modern era in consumer applications

48:43and a variety of other contexts and the economics that relates to people who make technological discoveries. But the biomeds is hugely patient outcomes. >> The health system is not what we're talking about when we talk about just the the raw ability to get the solution. And then we have to deal with how do we make this more accessible and there's sort of the public health healthcare insurance paradigm that's needs to be dealt with. But our ability to have real solutions to these problems that are this effective is is so again you know

49:26I'm sure many people who are listening us personally others have had friends family who have had maybe not this specific heart related issue but adjacent related heart related issues and it's the number one killer of people in the US >> um and we're slowly chipping away at the tool set to start to solve these problems. Again, I and it's it's

From hours to 30 minutes

49:52from hours to 30 minutes. If anyone's been to the hospital, you don't want to be in there. >> No, that's insane. I mean, the surgeons don't want to do a surgery for hours when they can do something in 30 minutes, right? Yeah. The the one funny thing I will say is like so I saw an interview with um Natalia Treyo Nova who is the professor of biomedical engineering at Johns Hopkins and um her team they're I think they're like trying to work on a desktop app now that is um that is going to make this accessible on a desktop. >> That's cool. >> And so doctors can have this information in minutes. Pretty soon they're going to make a app store like on the iPhone. [laughter] I think that's I think that was that was kind of funny. It's like

50:32now it's just like we're just going to miniaturize it and put it on a laptop. >> Yeah. >> And like there's no like that's kind of the I mean ultimately eventually basically you'll have the tooling >> and then your end surface a laptop or a phone >> and right now some of us have watches or aura rings that do very rudimentary. >> Yeah. It's just Yeah. The monitoring is going to go up, the technology is going to go up. This is great. >> The time is going to go down >> down >> which which matters in terms of saving lives. >> Yeah, frankly. Um

Midjourney Medical

51:07>> this is very cool. >> Super super fascinating. Great story. That's our story number two. I'm going to ceue up a story number three for you which interestingly is related >> to our story number two. >> We didn't plan this but we should. >> Right. Right. Right. [laughter] Right. This ended up being convenient even though so as you guys know with the rundown we're not going super deep into any particular story except the black hole story that was good but we >> there's still so much more I could talk about like we only spent 30 minutes on it. [laughter] I could have spent an hour and a half. You guys know me if you're regular listeners. But we try to use the rundown as a way in which to give the listeners a little

51:51taste of many of the things happening at the frontier. As we mentioned at the beginning of this episode, we're trying to cover cover uh cover a couple of stories we weren't able to get to earlier in the year as we get back into now doing our normal weekly rundown. So, we're hitting what's happening as it's happening because it's happening very fast. And one story that I saw that was fascinating to me was the introduction of Midjourney Medical. And some of you who are listening may be familiar with Midjourney. For those who don't know, Midjourney is primarily known for being an image generation tool. It initially started um as a Discordon image

52:35generation thing. So they had a bot on Discord. Yeah, I remember using your Discord account to like try to mess around. There was no website. There was no interface and they were very early even before a lot of the big >> uh model providers OpenAI anthropic in terms of leaning into um image generation. So the question might be what does MidJourney Medical have to do with AI image generation? And the underlying company has created this what they're calling an ultra CT scanner. Um they've launched a healthc care division and they've now announced this fullbody ultrasound scanner that aims to image

53:17the entire body in about 60 seconds using sound waves and water uh without radiation or magnetic fields which would be the the MRI would be the comparative here which we have I talked about a lot recently on the podcast and so a lot of people's uh spidey senses went up historically a lot of tech companies make a lot of claims there's not been a single of the modern tech architecture that has moved into medical imaging >> right >> uh at any time and the founder of midjourney has done a couple of different endeavors that he's really tried to lean into saying how can we utilize something like AI image

53:58generation to create a funding vehicle for greater good purposes right instead of having to raise money all the time so this is kind of what his vision has been for some time and midjourney broke through. So the the main idea here is it's a first generation uh prototype with no FDA regulatory clearance yet and it's built on licensed technology from butterfly network rather than obviously majour's generative AI but at launch they want to offer fullbody composition maps rather than diagnostics with a first location planned for late 2027 in San Francisco. So, the idea here [music]

54:38is Midjourney wants to create these Midjourney spas. We're now looking at this video of their their prototype that they brought forth at this demo that they brought [music] for other people where you stand in what is like a circular platform that lowers you into a pool of water. [music] You're surrounded in a sea by these circular micron level size sensors. There's like thousands, millions of these sensors that [music] are sending these radio waves at you. And you're getting this real time because it's ultrasound versus MRI, you're getting this real time sort of cross-sectional scan of your body that they then through [music] a few algorithmic systems then compose a 3D image afterwards. And [music] we're

55:19going to get into some of the technical details here in a second. When you step in the water, you're staying on the platform. It's connected directly [music] to these rails and begins moving into descending down here and gently lowering about 2 in or 5 cm [music] out of the water and they have a [music] very brief technical uh video breakdown of what they're doing here. [music] So as you descend, you pass through made of about half a million tiny squares, each the size of a fine grain of sand, [music] each capable of both being a tiny speaker and a tiny microscope. So you're looking [music] top down at a human body

56:02how these uh individuals [music] sending these waves and they're measuring. >> Oh, very nice. In in in all [music] these directions, they're doing this at super super high frequency. Very nice. >> Right. And so this is now how they are you trying to push [music] um I don't want to say more different implementation of the underlying concept. No. So that's really cool. Yeah. You've got um it's the microphone and speaker that you were talking about, right? Yeah. You're sending out ultrasound and then the Yeah, it's the It reminds me of, you know, how um we mapped the Earth's

56:43interior using seismic waves. You know what I mean? like like we we figured out that there's a mantle and a core and a um inner and outer core and all that other kind of stuff because we looked at how seismic waves, earthquake waves bent around in the interior of space [music] and we deduced that there has to be like these layers in the earth's crust or under the earth's crust. And here what you're doing is effectively [music] that but you're creating fake earthquakes, right? You're creating fake little earthquakes where you can control the size and the frequency of them [music] and then you have a bunch of like sensors. That's crazy. So that's so clever. Each of the squares creates

57:24ultrasonic waves of course [music] ripples back millions of times per second and they produce obviously terabytes of >> Yeah. per second. So part of the challenge is [music] the processing pipeline. I can imagine it's not just the sensing but it's also how you process the data. So take a look here now at the next one

How whole-body ultrasound imaging works

57:44which is so what we're looking at is is sort of one of the 3D images that it creates. And on the labels on each side you kind of see the different parts that's that it's identifying as waves travel through the water your body and your body they kind of change shape and the shapes change. But whenever they change, it helps them understand the density and stiffness going from water to fat to muscle to bone of what the underlying structure is. By looking at how the shape uh the shapes of all the wave change, they can reconstruct right this detail map. It's not dissimilar to other medical imaging technologies. It's just kind of a little bit of a different medium. And then so all of these images

58:24will then come together to create this 3D uh uh mockup of your body. So we we see in this this graphic here they're you're looking at two legs. >> Right. >> Right. And then they're basically

Prototype claims and limitations

58:39stretching out these two legs >> spreading out to the different slices that are these individual images. And so they're what they're putting they're sort of saying their piece is the software pipeline to go from the slices to the 3D imaging the time span to be able to do so the processing system and they you know their their goal is is quite interesting here. So what let's talk about like what exists now. Yeah. >> And and what does it so they have a working Gen One prototype demonstrated live as we saw >> and that's what we saw. >> That's what we saw. It does body composition mapping. That's the stated launch capability and they did this with a a licensing deal with butterfly

59:19network which has some of the underlying image sensing technology pieces. They kind of put the pipeline together to make it this end solution and the public announcement and marketing site. What is claimed or planned from Midjourney Medical is image quality superior to MRI which many uh in the radiological and scientific community are have some feedback about which I'll talk about in a second. >> Diagnostic capability pending FDA submissions. They want to have 50,000 scanners worldwide over a six-year period. And they want to be doing a billion full body scans per month. That's their stated goal for this system

1:00:02and platform. So like what is the so what here? Like okay great. Like why does this kind of like really matter?

What ultrasound can—and cannot—replace

1:00:08>> Ultrasound is you know generally excellent for like a wide range of problems including you know assessing abdominal organs, blood vessels, the heart pregnancy >> any other kind of soft tissue problems right? Uh in real time without radiation. Okay great. uh what it doesn't do straightforwardly and can't replace kind of what MRI does currently uh is the soft tissue contrast for the brain and the spinal cord not quite as good in those context characterization of joints and certain tumors and then validation pathways for certain platforms like DEXA for bone density okay >> so it's not quite as as uh able to be great in those areas and when I looked at some of the back and forth people

1:00:50said medical imaging technology is about a tool toolbox of instruments. It's not a one instrument to rule them all. And depending on your use case, you have better options than others. So the idea is a whole new body ultrasound system can do some things very well and it might not answer every problem, but doing some things really well that makes it more accessible. Yeah. Is super valuable >> in and of uh itself. And so that's kind of I thought this was interesting because it's this, you know, we have a lot of the tech folks come in to say we want to help save the world. This seems to be an actual pract something that could be really practical um if it can

FDA approval and adoption

1:01:30go through the FDA process and get approval here and fit into the toolbox of medical imaging. It would be the first new medical imaging device. I believe it's in like over 50 years plus. That's really, you know, it has to gain adoption and get buy in. But I think there's some interesting >> concept that it that is that is really cool >> in the fundamentals of of what they're doing here. >> Yeah. Yeah, definitely. I mean the only ultrasound that I've ever really experienced is like you know for like pregnancies and things like that, right? So so but obviously yeah doing doing soft tissue probing is going to be huge. The one one the one thing that I'm like

1:02:11kind of skeptical about I would say is the billion full body scans per month. There's only 7 billion people on the earth. So like what I guess I guess a single person could have multiple full body scans cuz this is as you said real time thing >> and they're trying to make it a lifestyle spa concept like >> but then they got to make it cheap >> which I agree with you. No, which I >> like a billion a month is like that that already you're um you're you now have to you're saturating the developed world. >> You know what I mean? Now you got to tap into like a lot a lot of countries that don't have the kind of capital for the

1:02:51healthcare that we do here in the states. >> I don't know if we want up and things like that. >> We don't need a spa like this in in Zimbabwe for example. I don't think that's the right entry point. So, so a billion full body scans per month. What? So, a single person is doing multiple per >> month. Yeah. >> And then also it's like, well, if one scan is really good, why do I need to be doing multiple? >> Exactly. Like, how much is my body really changing month per month? >> Um, anyways, that that's the only number that I'm like kind of skeptical about. Um, 50,000 scanners worldwide, that makes sense. It's not that bad. There's a lot of cities in the world that are rel that have a relatively wellto-do enough population to like want something

1:03:32like this in one of their hospitals. >> We could do a dozen to two dozen in LA alone. >> Exactly. So, um yeah. No, I I still think Yeah. the fact that it's a new medical imaging device. I think that's what the interesting thing is is just at least we're we're pushing you know again >> there's both this balance between what's tried and true >> and then also integrating like crisper right as an example it's a new medical technology in a totally different context and we're now trying to figure out how do we make this safe how do we scale it and I think it is good to push the envelope >> the the weird part about this story is at least in part the source which is coming from a non-traditional Yeah.

1:04:14Yeah. Yeah. In the in the medical imaging space which has created a lot of skepticism. >> Uh but again the proof's in the pudding.

Support the show

1:04:21It's falsifiable. They can put it to the test and if the FDA approves it, >> we will have a great new solution. We are going to take this moment after our first three stories. We still have I believe it is two stories left. >> Yeah. But those are really kind of fake stories. [laughter] >> They're going to be fun because we're going to end on some World Cup hype. But before we get there, I'm going to do a little bit of housekeeping for those who are listening. As you know, with myself and Krishna, it is us two on the pod. No network, no extra people coming around. We really do this, just the two of us. We really appreciate the support from our patrons who help support bringing

1:05:02you the best entertaining science and deep dive content that you can find in the world in English language. you get it here on FFP. If you want to help support the podcast, there's a couple ways you can support. You can become a donor by going to ffpod.com/donate. We have a variety of different options, one-time and monthly. If you cannot support monetarily in this economy, please share it with a friend, bring it to Journal Club, post it on your Instagram, your socials. Any kind of promotion is helpful to help us get put up in the billionaire algorithms which we do not control. And science and science education and news is so

1:05:43important especially in this day and age which is it is progressing faster than ever and it is could not be more it could not be uh less understood in an era where it's growing so quickly and so any support that you could give us is super appreciated. All of our shows are on fabod.com. You can listen to us on all of the podcast networks. Our clips are on social. If you have found yourself stumbling upon this episode after watching some of our World Cup content in the past couple of weeks, welcome. This is what the show is like when we're not talking about football. We will talk about a little bit at the end of the episode, but what you just experienced in the first half is really

1:06:25what the show is about. It's great for a car ride, a commute to work. It's great if you want, you have some kids getting to the teenage years. You want to get them into some content that's about science, but everything else is boring. We have Drake memes. We understand what's going on on the interwebs, what the kids are doing these days. There's something about 67. Who knows? But it's a great way to get everyone started and engaged with science that's fun, entertaining, but still detailed about the fundamentals and the facts. And with that, what we are going to do, which has been a long time since we have done it,

1:07:07is bring back one of my favorite segments, which is are you smarter than

Are You Smarter Than a Scientist?

1:07:14a scientist? Our favorite game show. Oh, okay. With our resident PhD Krishna Chowdery. As we know the World Cup is about to end. Everyone is going to have to find something to entertain themselves. And to end the World Cup journey, we have our question this week is to name the 10 most common injuries in professional football. And I will just note that this is based on UEFA elite club UEFA elites uh clubs injury study which is a long-running European

1:07:55pro club uh data study that was first published in 2009 with newer findings from the 2021 and 2022 seasons including their latest ECIS update in the British Journal of Sports Medicine in 2023. So, this is not just us making up what it is. This is published data. I'm trying to give you time to think because now you are on the spot for the most common. >> Can I Can I Can I just name like the the specific body parts? >> You you I you're going to need to name a body part >> and potentially the type of

1:08:37>> thing that's happening with the body part because I initially made this as just body part only. But that was too easy. >> Okay. >> So, >> um All right. Uh what about like the uh Achilles tendon? >> Achilles tendon? >> Like the Achilles, you know, the >> your Achilles heel, >> the heel. Achilles heel. >> So, the question is, is Achilles heel an answer on the list? And unfortunately, >> no. >> Oh my god, I'm already That is going to be your first strike early in the game. early in the game. I don't This is the first time I don't even have one on there. [laughter] >> Okay. One actually. Okay. I'm pretty

1:09:19sure even though this is this is European football >> and American soccer, there should probably be concussion on here. No. >> Like people are getting hit with yellow cards like like Yeah, I would say. Okay, let's do concussion. Is concussion on the list? I really hate to say it. No, >> but it's not in the top 10 muscles. >> 10. Okay. >> In the most common. >> You were more directly correct with your first >> It's It's got to do with legs, huh? All right. >> I'm just I'm just going to say that

1:10:01this is >> Yeah. You know, hamstring. >> Okay. So, we're saying hamstring. >> Yeah. And we're going to be on the board with a hamstring strainer. Okay, I'll give you that. >> Okay, >> so very common. >> Very common. Yeah, I I think you've even mentioned it. >> Yes. At 24%. Uh it's it's kind of the meme. The celebrations have been made over people grabbing their hamstring. Yeah. Like after they score a goal, pretending they have a hamstring injury. >> Okay. >> So, we're on the board. >> Okay. So, we're cooking. You got You'll get some You'll get some here. >> Okay. Okay. Um, I'm going to need some lifelines though cuz I'm already on two. Okay. Um, okay. So, so you're going to have

1:10:41>> So, what what I would what I would say that what I would say here is think about, >> you know, football is a is a sport where we we run long distances, >> right? And there's a part of the body over those long distances that is very active in in being a in being a uh try not to be too specific here and being a facilitator of other parts of the lower body moving over long distances. >> Um no I'm thinking of the heart. >> No no no >> but that's that doesn't make sense. The lower Okay. the think about the the different component parts and something

1:11:23has to be a translator between >> Oh, like the knee. I don't know. I'm just saying. >> Okay. So, um, >> wait, what's in the knee? Oh, the ACLU. >> No, that's the [laughter] that's the that's the uh >> No, but you know what I mean. >> The thingy. The There are three letters there that are correct. >> The ACL. The ACL. >> Yeah, ACL tear. >> Okay, there we go. is our number 10 answer. >> And and a little bit lower than many would expect at only 2%. >> Okay. >> Um >> but like that's the knee. >> That's the knee. >> Okay. That's the knee. We're getting a tear in the knee. >> All right. We got hamstring, which is I think thighs. Now we got knees. No, there's got to be something in the foot.

1:12:04Like the guy the guy from Canada like broke his Didn't he break something because of the Qatari um red card? I I plead the fifth. That's got to be something. >> I plead the fifth. I would give I'll give you another lifeline which is I would stay in muscles more than bones >> more than bones. Okay. >> Only Christian Pulysick gets a micro fracture. [laughter] >> Well, that's actually not true. And so let me not that's not it may be in our answers. So that that's not true. But I think being more muscle oriented would be the correct approach here. I know there's people who are watching at home

1:12:44and they're like bing bing bong boom bing bing and [laughter] they're just going through this list. I've gotten actually probably six to eight of these injuries myself. >> Oh, really? >> Yeah. >> This is muscle um let's see, no hamstring is done. The cal calves calf strain. We will go ahead and put calf strain at number seven with 4%. We are on the board again. You're starting to >> Okay. >> You're starting to wind it. It's been a while since we've done it. I will note Christian is a biophysicist. Physicist cuz biohysicists people get annoyed about it. >> Yeah. But mostly like at the molecular level, not uh [laughter]

1:13:25>> molecular and cellular level, not at a full organism. Um no, calves, hamstrings, hamstring. Um >> you're on the board. You're on the board. >> ACLU is the There's got to be something with the foot. How about um like a No, I already said Achilles. That didn't work. Um again, think about soccer being a sport where we run a lot. >> Yeah. Okay. So, calves. What else? What is this thing called? The thigh. No, thigh muscle. The quad. Quads. Quad strain. We got it on the board. Number six. Quad strain with 5%. >> Okay. Now he's got Okay, I How many

1:14:06muscles [laughter] can I think of that are in the lower body? Um, [gasps] >> we're cooking. We're missing number two, three, four, five, eight, and nine. There's a lot of items on the board. Well, strange. Damn, I really thought concussion was going to be on here. Um, thankfully, we don't make contact with our heads that often. CTE is not a problem in the game of football. How about uh he No, I don't. Now I'm just blanking. I literally have no idea. Like I'm going to say heel. >> The heel. >> The the heel. >> The heel muscle. >> The heel muscle. Is it on the board for

1:14:48our last strike? Unfortunately, >> that is not on the board. >> Yeah, that's fair. >> This was a tough one. >> Yeah, I just shows you how little I know about the human body. was really trying to mostly rage bait the audience who are screaming the answers at home right now. Um, I kind of put you in a tough spot, but we're going to go through with the correct answers. And our number two spot >> was a contusion >> at 17% which you can get bruising, bone contusion. >> Okay, >> very common. Uh, number three is an injury I currently have after going to shout out to Ben Ben Fong's birthday

1:15:29party. Uh where I got a groin strain. >> Groin >> 9%. >> I could have I should have done groin. Yeah, >> very common. >> I know you said Achilles. It was close but not quite the ankle sprain. >> Oh, >> which was the answer we were looking for at 7%. >> Probably an Achilles in there. I >> I I don't think that's correct. >> No, >> I'm shocked we didn't get the the ankle sprain. Now, we talked about the ACLU u tear. >> Yes. [laughter] But we did not talk about the MCL sprain, which is another type of

1:16:10abbreviation sprain. I'm not a doctor. >> I know there's ACL, MCL. >> Everyone talks about it. It's so common. our number eight uh which is more in the direction of the former captain of the US men's national team Christian Pulysick with a fracture. >> Oh yeah. Okay. >> At 4%. And our last option I was trying to talk about the connector is a meniscus tear. >> Oh, I could have gotten that. >> Which which was I could have gotten that which was at 3%. Uh we did not get a [screaming] >> this episode. No, I didn't deserve that. >> It It was not It was not uh that Well,

1:16:51but we really do appreciate all of our listeners joining for another great installment of Are You Smarter Than a Scientist? I did put Krishna on the spot today and the stream deck worked perfectly with no interruptions. So, we are not rusty at all as we are now pushing an hour and 15 minutes for what was supposed to be a short episode. But as you know us, we like to yap on this show. And so we're going to go now into the last of our two stories which are going to be fun. It's going to be relaxing. We're going to wrap this up pretty quick. Here we are going into our England versus Norway World Cup goal

1:17:34ball camera cable. Hey Jude Bellingham story. Yes. So Norway versus England in the quarterfinals.

Norway–England Skycam controversy

1:17:45Um, the goalie of Norway kicks a ball and I think we got a video of it on Fox Sports. This is video 14. Yeah. Um, this on Fox Sports, right? He he he kicks the ball. Apparently, it hits a wire, one of the wires that um holds up the >> sky cam. >> Sky cam. And then it lands directly on an English player who then dribbles, gives it to Bellingham somewhere, and then Bellingham scores. I think that's Anthony Gordon, $80 million Barcelona player, eat your heart out, Marcus Rashford. And then the Norway team's like, "It definitely hit." And Fox, the Fox Sports, there's Erling Holland saying it definitely hit. And Fox Sports

1:18:26on their Instagram says it hit a wire. >> So they went they went on the record to say it hit a wire. Right now, let's go to the next Let's go to the next video. Um, this is from TSN, which is, I think, the Canadian ESPN. Um, and they show

Did the ball hit the cable?

1:18:43from the side. >> It hits the spider cam and then it goes down, changes the trajectory. >> Looks a little infantino-ish. A little sketch. >> It looks a little sketchy. It looks like I mean, they're doing this whole like, you know, zoom in and things like that, right? The next video we've got is just a clean video that I found of the ball going up and then it definitely like falls down. >> Bruh, >> it definitely falls down like >> you know um the camera itself is moving which is why I can't really get a like a good handle >> right on what's what's going on >> on what's going on. But like it seems like it goes up and then and then it

1:19:23falls faster than it should. >> This is not a rainbow where there's gold at the end of the rainbow. No. Right. And it it seems more than air resistance because I do I do understand that like you know as it's not going to be a perfect parabola. The parabola is going to get squished as you keep going. Just doesn't look quite like air resistance. >> A little sketch. Right now the ball has a sensor. That's the story which we talked about. >> Yes. We talked about the ball has a sensor and um it's taking data at 500 hertz. It's got this inner um this IMU, inertial mass unit, right under the surface of the ball. Um it goes off and

1:20:04it's sending data at 500 times a second. Right? This thing is measuring acceleration in three directions. So FIFA puts out this video of the ball. There's the data from the IMU. In the bottom left corner, we see it like what looks like the if you watch uh ER or the pit or any hospital shows, the little thing that shows your heartbeat. Yeah. This this is flatlined. >> It's flatlined except for like when the when the when the goalie kicked it. >> That's it. That's it. >> It's flatlining. There's a little bit of undulation. You can actually tell if you look closely. There's a tiny bit of undulation. >> It's not totally flat. It's not it's not a line. It's not a line, which tells me cuz there's a theory that's going out

1:20:44there saying that like, oh, perhaps the ball um cut off like the it was far enough away from the receivers to where the data packet wasn't going through. >> I don't see that because I'm seeing >> it's still it's continues to it continues >> like it's not flatlining. There's still

The ball-sensor data

1:21:01there's still like salient data going through. >> It's clearly a consistent signal that is not it's not just whatever uh uh y equals zero. >> Exactly. Yeah. So, so now this brings me to my next point, right? Which is, and I I'd like you to stay on this, right? >> Okay. Well, we're gonna keep this up. >> Yeah. So, it brings me to my next point, which is the ball is now traveling through the air. >> Yeah. >> Right. It's it's experiencing freef fall. It's also spinning >> because with as with any ball, and I think you can see in the in the in the video, the ball is actually spinning, but this thing is flatlined or almost. There is a tiny undulation which I think is caused by the spin. But the

Post-processing and missing labels

1:21:41what I think is happening is that there is post-processing algorithms that are taking the raw data because what we're seeing right now is just a 1D trace. First of all, it it's infuriating to me that there's no axis labels. >> Okay? My PhD adviser would [laughter] would have my neck if I presented him any data that looked like that. And for good reason. Okay, you got to have labels on your axis. Tell me like a seconds uh and and what the y-axis even means. Is it a meters/s? Am I looking at acceleration? Am I looking at angular acceleration? What am I even looking at? Is it jerk? Is it the derivative of

1:22:21acceleration? So on and so forth, right? So anyways, there's got to be there's like a sensitivity threshold, right, for when things get triggered, >> right? >> And I think there's got to be some post-processing of the data that is happening, >> right? >> Which which is true of of a lot of sensor platforms where you have the level of volume of data. There's just so much noise. You have to what are you what are you looking for? >> Yeah. And in any case, the IMU is measuring three units of acceleration. It's an X, Y, and Z in acceleration. If something is spinning, you're going to get a acceleration in X, Y, and Z, and all of those are going to cycle through which is going to cause the spinning. So, the fact that I'm seeing this like

1:23:02flatlining means that there is some post-processing going on >> fundamentally, >> right? Fundamentally, we're not we're not even seeing the like cuz it would be anyway, we would see a three-dimensional thing. >> Yeah, I would see three traces at least, but I'm seeing only one now. Perhaps they're signaling out one, but in any case, I'm not seeing the whole thing. Okay, we can all agree that this is not raw data that I'm seeing >> 100%. >> Okay, with no labels. >> Yeah, with with with no labels. Okay, now the second thing is now if the ball is spinning and it's it's subject to this constant acceleration, right? There's got to be an algorithm that's looking for a discrete touch like the one that the goalie had, like the one when he when he fell. >> Um >> or the one where it hit the hair of the

1:23:43one player on the >> and actually that's the next one. If you if you go to if you go to the next one, we've got Croatia versus Portugal. There's a little bump of it hitting the the guy his hair and then all of a sudden um Croatia's goal is disallowed. Right

Could an algorithm filter out the impact?

1:23:58>> now, >> there's not a lot of information >> about the post-processing, >> yes, >> of the data. There's a lot of information out there on the FIFA website about the 500 hertz, about the IMU, but at the end of the day, fundamentally, what is the data that I'm seeing on my TV that they're presenting as evidence? I don't know. Now, what the worst case scenario would be that there is an AI algorithm that is behind this, right? that has been trained on player touches, that has been trained on what to look for when I kick a ball, when a a hand

1:24:41feels a ball, when it hits my head, right? These are things that the AI algorithm would have been trained for. And perhaps the signature >> per chance >> per per chance per chance the signature of the ball hitting a wire >> is a very different signature in that IMU compared to hitting a human body because there's a springiness in us compared to like the wire. But I guess the wire also has a springiness. But like maybe maybe the timing is like small enough where like the algorithm thought it was just noise. It's like some type of thing, right? I'm just saying I I don't think this is a good way. >> It's not in the training data set. >> It Yeah, perhaps it's not in the

1:25:21training data set. And so now >> it gets filtered out. >> It gets filtered out because it thinks it's noise. >> Look, >> I don't know. >> I think so if you we we do know after our momentum match momentum video that there are folks who've worked on some of FIFA's uh graphics and and digital analytics tools that are displayed in game. We would love to actually understand yeah this question. It's it's in the weeds.

Why didn’t the IMU detect it?

1:25:50>> It's in the weeds, but like it's also like fairly clear from the video that it hit something, right? How do we explain? >> It's like it's like it's I don't know. Like it's it's fairly clear from the video that there's a change in momentum. How is it that the IMU on the ball did not pick it up? Both things can be true. >> Right. Right. Right. Right. Like you can it can be true in saying well we have these ball sensors and >> and I don't Yeah like nothing got triggered and so the referee didn't see it. Okay. >> However, we have visuals from multiple angles that don't align with that description of events. >> And so this is a great learning opportunity. >> Yeah, >> we'd be happy to have you on and to

1:26:31discuss is there a filter? Is there sort of sort of an AI algorithm that helps to distill down these three axes? Yeah. >> Into what we see in a 2D plane with no labels. And I get it. This it's sports. We people don't know want to read. >> And fair play to the Norway team. Like they they said like let's not make this a big deal. >> Let's play let's play. Let's play. However, it is I Yeah. I I I'd like to know, right? And and I also think like it's also just a bad way to do if you are in an experimental lab and you are only relying on a single sensor that has a heavy amount of data manipulation.

1:27:12That's just a bad way to do things. You should always have coincident data. And so like and I'm pretty sure FIFA has a bunch of cameras as we were talking about, right? So there's got to be other perhaps that modality of data sensing is not in real time and and so as such >> and and and as such you can't stop the and perhaps only the sensor is the thing that is like affecting real-time decisions and if there's a filter there at least at a minimum we can say hey yes we need to expand the filter a little bit to account for this incidental event type that was not in our initial problem set when we defined find how to define the filters. And that's okay. This is

1:27:54just like this is fun. Like this is >> I just I don't know that like we don't know. >> Uh I'm literally just speculating, but like given what I've seen. >> It doesn't it doesn't add up. >> Doesn't the math is not mathing. Yeah. >> And so we would appreciate an answer. FIFA. We are here to give you a platform to explain if you would like to. And if you would not like to, we will still get the answer one way or another. Yeah. And so I'll figure it out. >> Um it is just uh >> these are the challenges that happen when you put new technology into production and it happens in hardware, it happens in software. We now it was the US. We have these crazy sky cams,

1:28:37>> blah blah blah. They're moving quickly.

Is FIFA favoring Argentina?

1:28:40Players can kick the ball higher now. All this stuff. We're going to move to our second one which is another interesting >> Yes. aspect of the World Cup. >> Yes. I'm sure you've seen the entire world turn on Argentina. >> It ha It has happened. >> It has happened, right? Uh this is the BBC classic BBC classic BBC >> ahead of the uh Argentina England match. Are Argentina being treated favorably at the World Cup? Um is Messi and Argentina being favored by FIFA? Right. >> So um there's a lot of just like rhetoric out there. Um, a lot of anecdotal evidence, a lot of um, people watching the Egypt game, >> witness testimony, one would say,

1:29:22>> "What's going on?" People watching the Swiss game saying, "What's going on?" Um, there is a guy from Northeastern from the Northeastern Sports Statistics, uh, NetSi Sports Research Group, um, Brennan Klein. So, he compiled some data. >> We love data. >> We love data. Okay. So, here's a chart that's created by him. Um, what he's

World Cup VAR outcomes

1:29:46showing is this year's World Cup, all of the teams that have seen the most favorable V outcomes, that's in blue, and those that have seen the least favorable V outcomes using a baseline of per 100 fouls. So, he's >> he's normalizing for the number of fouls. And [clears throat] he's saying per foul, how many V decisions were turned in favor of me and how many VR decisions were turned in the favor of >> per 100 fouls? >> Yes. >> Per 100 fouls. >> Per Yeah. Yeah. Per 100 fouls is just a way of saying I'm normalizing by the number like I'm just dividing by the number of fouls, right? So it's like a per foul type

1:30:26>> type thing, right? But this becomes a percentage because now you have 100 in the denominator. I got you. >> Um, so, okay, Mexico actually tops that list. >> And then it's Argentina. >> Okay. >> Okay. And then it's a bunch of other countries. Um, England is somewhere in the middle. France is somewhere in the middle, but France is also net positive. There's no net negatives against them. >> Yeah. >> Um, there's another plot that that's that's that's there. >> I just want to note that the US is net negative. >> Yeah. >> We V does not support the US. No, not at all. >> Or Canada for that matter. Uh or Belgium. So, I can't complain about the Belgium game. >> Yeah. [laughter] And and so the next the

1:31:07next um plot is something that I found on Reddit

Fouls, yellow cards, and referee tolerance

1:31:13here. The Yaxis is the opposition fouls per yellow card. Okay. So, fouls per yellow card that you're getting, meaning um how many fouls your opponent commits until they receive a yellow. Okay. what's the like the threshold? >> Mhm. >> And on the um the lower the the lower on the y- axis the the more favorable the team is. >> Okay. The x- axis is the delta between penalties won and penalties conceded. So there that's how you get 3 2 1. Okay. Argentina the bubble size by the way the the the size of the bubble shows how many fouls the team commits until they are shown a yellow card. So the bigger

1:31:55the size of the bubble, the more tolerant the referees are for that team. Argentina's on the bottom right corner, [laughter] right? And so

Small samples and confirmation bias

1:32:08>> with a pretty big bubble >> with a pretty big bubble, right? Which means the bigger the bubble, the more tolerant the referees are, so to speak, in this >> in this >> in a very crude crude example, right? >> And >> it naively >> Yeah. would suggest, >> yeah, >> that the referees are being biased. >> Argentina won most penalties while its opponents receive a yellow card after five fouls committed. They get a yellow after 19 fouls committed. >> Yeah. Yeah. So, Argentina needs 19 fouls to commit to get a yellow. Their opponents need about five. >> I don't know. >> Yeah. Right. And that's what that's what I would say, too. I don't know. >> Because I want to believe.

1:32:48>> Look, >> I want to believe. But I'm I'm I'm going to also have to put on my like data scientist hat on. >> Okay. >> Okay. >> Okay. >> And just do some good hygiene. >> Okay. >> Okay. >> Okay. >> First of all, >> yes, >> it's an incredibly small sample size. >> Okay. >> All right. If you go back to that, right, >> this is 22. >> Yeah. Uh if you go back to that, okay, the x-axis is disc is discretized by negative 1 0 1 2 and three. But they've made the the plot fatter on the x- axis. So, it seems like the x-axis is the more salient dimension. >> I see that. >> But like there's nothing there's no way

1:33:29to have a 2.5. >> Yeah. Right. Right. I I see what you're saying. >> Okay. So, so already there's there's a bit of like movie magic happening. >> There's a little bit of manipulation of the visual representation. >> Now, the other thing is they're treating every foul equally, right? That's not completely true. Like if you've got a team like Argentina that presses all the time, you've got someone like Messi who's like always in your space then like if you commit a foul against them, there's a higher chance of like that resulting in a penalty. That's not truly accounted for here. >> Okay. >> Right. So style of play perhaps is not something that is that is accounted for. Like Kate Verde for example, like those

1:34:10dudes had a great defense for for quite a while. >> So So yeah. So part of what you're say like okay I think style of play I I'm I think that example may be contested but generally speaking um this idea of if you play haram ball which is what Arsenal plays >> and you just sit and it's 11 of you in your own box and then you just wait to counterattack like necessarily you don't have a lot of opportunities as much theoretically as many opportunities to be fouled. >> Exactly. Egypt certainly did that >> because you're not on the ball. >> Yeah. >> Right. So, so that's not something that has been taken into account. Also, the types of fouls like Argentina

1:34:51may like the fact that they require 19 fouls to get a yellow. Perhaps it's because they're doing smaller technical fouls. Now, anecdotally, I would say that's not [laughter] true. >> The comments are going to go crazy. Um, but there's all sorts of like I I think with all of this stuff and I think the author of the North Northeastern paper, right, or the Northeastern study actually said this, like there's like this kind of exercise given the small data size. >> Sure. >> And like the heat of what we're doing like the exercise is fraught with confirmation bias, right? But at the end of the day,

1:35:32it's the World Cup and I personally [laughter] want FIFA to be biased to Argentina. So, so at the end of the day, all I'm going to say is like, look, is this is this truly convincing data that Argentina, sorry, is this truly convincing data that there is a bias for Argentina by FIFA? No. But do I care? Also, no. All right. So, yeah. # #rigged # stop the steel >> uh better callinfantino. I just want everyone to know that I am not an Argentina supporter and I do not condone any rigging. However, I will say

1:36:12the comment thread is going to be full of the shot of Messi giving a red card foul raise on the calf, one of the main injuries we saw from our Are You Smarter Than a Scientist game show segment. uh and say, "Oh, how is this not a foul never gets called?" Like, that's I'm just saying it because I already know. >> Yeah. There's >> it's going to be the image that is all in the clips of this. Yeah. And let's put this at the end of the clip. >> Yeah. >> So that people who comment it before reaching the end of the video, which apparently no one watches >> 90 seconds of content Yeah. anymore >> before just ra >> I just I just want you guys to know you you are your brain is rotting because

1:36:53this is a 90second clip and you saw Christian say something about Argentina that was favorable and then we came in here and said yes we know about the Messi foul with the thingy and his plate in the calf and you're going to put the image of it and you're going to look stupid. I just want you to know that [laughter] that's I just got to I had to get that off my chest. >> Yeah. And and yeah, the point the point basically being like >> look, I I don't know how much of me thinking that there's an Argentina bias, it's just because of all the stuff that I'm seeing on social media. But for the purposes of me enjoying this England versus Argentina game, I am believing 100% of it

1:37:35>> and I want Argentina to go down. And that is rare coming from a of a from an Indian >> 100% [snorts] of Indian descent. memes with the map where it's like the two countries that care. It's like Argentina and India that want India to win and the rest of the world is England. >> No, like I'm I want I want England to win. Rule Britannia as they say. >> Again, we'll put this at the end of my part of the clip >> so people also No, but this it'll be fun. It's great. >> The World Cup is great. you know, the US is out, which we're not gonna comment on, but yeah. Um, I as a Chelsea fan of the Chelse is I want to see Tuke be successful because my club unceremoniously

1:38:16sacked him after he brought us a Champions League with nobody's >> like like he he he brought us a Champions League with >> like like with with anyway I'm not can't think of analogy but like with we did not have the best players on paper. >> Yeah. >> Right. And then there's some >> and you got to the won the Champions League >> and we won. >> That's quite good. >> Captain America Christian Pulik has a Champions League thanks to Thomas Tukul, the current England >> gaffer. So I know he's good. >> I know he can do a lot with a little. There's talent on the England squad. >> I I'm so conflicted about this game.

1:38:57>> Yeah. because I have I have disdain for both for different reasons >> and it's like it's what's the lesser what's the lesser of two evils? I will say as a fan of the Prem >> I will likely have to say I would prefer >> to see England win this one and then lose in the final. It can't come home right now. We can't have us lo Let me not I'm not going to talk about >> Yeah. at America's 250th. Yeah. Like like we can't this is supposed to be our celebration and then they win it. >> Yeah. >> We just can't have our >> Especially after we we didn't just lose, we like got embarrassed >> and then the Europeans rub it in our

1:39:39face. >> Yeah. And and like we deserve it and we deserve it cuz we we had the whole red card bal like it got reversed. It was just like it's just like it was like the worst way to go out, dude. It's bad management. Okay. >> Yeah. It was so bad. >> None of us asked for anything to be rescended. All right. >> [laughter] >> I know Quansa had the twoame suspension which doesn't make sense and Balagan got it removed. This is getting in the weeds. Yeah, this is not a sports podcast. However, so the last thing that I want to say, okay, is like cuz I was like going in a deep dive about like what could it be, right? >> Right. >> I have a theory about the Argentina bias. >> Okay. >> That I think is new. >> Okay. >> Oo, new Argentina bias.

1:40:20>> So, it could be >> that there is a bias for Argentina and FIFA has nothing to do with it. >> Okay. Okay. It could be that they have a homecourt advantage at all of these games because I don't know if you've seen in all of these Argentina games, the stands are flooded with Argentina fans. >> It's the Messi effect. >> Yeah, it's the Everybody has an Argentina jersey, mostly a Messi jersey. >> Ah, I see what you're saying. >> So, there there is stud there are studies. So, this is in 2007. um Boco um and other researchers what they did was try to try to quantify the home

The science of home-field advantage

1:40:58field advantage >> which is a well doumented phenomenon right like there is a home field advantage for like the players they get riled up there's a home field advantage for refereeing there is a referee bias that contributes to homefield advantage in the English Premier League and it has to do with just like psychology right like the referee doesn't want to give a bad um a bad decision to the thousands and tens of thousands of people that are watching him. >> Now, it didn't quite work with the Balagan >> right situation >> situation, but it did work when Argentina was in Miami >> where Messi plays in the MLS in Miami,

1:41:38>> which is home field effectively. >> Exactly. And like even the Kansas City game, right, against Switzerland, packed with Argentinian fans, >> every single one of these Argentina games just has so many fans, right? And so what these guys did was, you know, it it's hard to disentangle um the the performance of the team with the performance of the referees. But what they did was separate out referees and look at individual referees whether they were judging a home game versus another another game and look at individual referees. So now you've washed out the effect of whether the team is at home or not. Okay. And in the next figure, we'll show you this is a figure from the from from the um

1:42:21from from the paper on the x-axis all the individual referees and this is the mean goal differential that they could have contributed to based on their yellow cards and and all and all things like that. All of it is positive. Meaning like the referee bias is always positive for the home team. >> That's very even then. Okay. >> You could be like, it's really hard to conflate the home field advantage from the referee and the home field advantage from the team. Like the team is more inspired. Maybe they're more violent. >> The 12th man, the fan, like all these things, it's they they play there every whatever. >> Exactly. So, what we really need to do as a control would be to remove the 12th man.

1:43:01>> Right. Right. And lo and behold, the COVID pandemic happens.

COVID created a natural experiment

1:43:09Oh, wait. This is perfect. It's the only time you'd be ever be able to get this data. >> Yes. It is the only It's the only time where you're playing football and there's no one in the stands. >> There's nobody there, right? >> Wait, this is great. >> So, now you've got to control. I That's why I wanted to bring this up because I thought it was so cool that we've got a 20 We've got a 2007 study that is showing that there is a home field advantage for referee bias, right? referees are biased for the home team but you could always be like it's so hard to conflict% right okay um the next paper >> yes >> is the natural experiment by Arendelle >> um in the frontiers of behavioral economics

1:43:50>> and they looked at the same pattern but now dur for the games during the co >> 19 good because you can't there is no there is effectively at least you could say it's the pitch still but really it's about the fans and the energy and the ambiance.

What happens when the crowd disappears

1:44:07>> No, zero. >> And they confirmed that referee bias. They looked at 7,000 matches across Europe and they showed that when the crowd goes silent, there's no crowd. The referee bias for the home team is eliminated. That's incredible. >> Yeah. >> So, and again, this anecdotally everyone's believed this to be true. Yeah. Forever. It's like this is why homefield advantage is even a concept. >> Yeah. Right. It's like a homefield advantage not only in the psyche of the players, but now they're showing it's in the psyche of the referee himself or herself. >> There's bias in the >> I thought this was so cool. >> No, that's really good because again you we will we will probably never really ever get

1:44:48>> No, we're never going to get that experiment with these. >> I mean, God willing, right? Like let's hope so, >> right? Right. Especially with the SK because you could argue like, oh well, one team can be banned from having fans in the stands for some period of time. But we're saying this was everybody. >> Yeah. for a whole year, >> right? And the amount of data that they've used, it's like 5,000 games, 7,000 games. Now you're getting a statistically significant effect. >> So >> people are so glad. >> Yeah. I thought it was good cuz they also built on the other one. It was like, well, let's let's double and and so so the point is it was positive in the study where the fans were there >> and then you took away the fans >> when the fans were not there, which means understandably the rest want to get home safe. >> Yeah. And so they like, you know, might

1:45:30be a little and so >> and just like psychologically, right? Like there's something in your psyche when like there's there's like 10 like 50,000 people booing you. You don't you don't want it, >> right? Even though you're supposed to be like neutral and whatever, there's going to be part of you that wants to be liked by the 70,000 people, right? That you want to hear cheers for when you do this and you say no penalty [laughter] or penalty or whatever, right? So >> this is going to do numbers for the Argentina bias community. >> Yeah. So that's my theory which I think and I think it's a new theory and so this is an out for FIFA because FIFA doesn't have anything to do with this. This is just like the home field

1:46:11advantage that Argentina has. Again, this is terrible data science because clearly it didn't work for America, >> right? However, one would argue that our fans don't instill the same fear in referees that footballing nation fans do. >> Fair. >> The English fans, the Argentinian fans, like these Americans, they're they can't even USA. Who's scared of that? >> You know what I mean? >> Yeah. And Yeah. [laughter] And then we stop caring. >> Yeah. Right. Immediately they're not going to do anything. They're going to watch the TV. Um,

Final thoughts

1:46:44>> fascinating. This has been a blockbuster rundown. Our longest rundown by literally 2x. >> Yeah. >> Because we just love yapping and we love talking to you guys and there's so many good things and exciting things happening. Uh we covered uh black holes digital heart twin, the uh ultrasound CT scanner from MidJourney. Two wonderful World Cup stories. first about the controversy about the camera which again FIFA if you'd like to come on on and explain yourself you have a platform to do so and where does it wasn't about the

1:47:24Argentina Argentina bias specifically but homefield advantage is it statistically significant referee bias being actually the thing that creates the Argentina bias I think is a very clever hypothesis >> we will see what the algorithm and the comments on the internet say, "My name is Lester Nar, joined as always by my co-host and our resident PhD and Argentina biased hypothesis creator Krishna Chowdery. We really appreciate you all as always for joining us meandering through our friendship in a journey of curiosity and science. We

1:48:06will see [music] you all next week.

1:48:17>> [music]