 This is Covering the Spread. Here are your hosts, Jim Saunas and Dr. Ed Fang. What is going on everybody? Welcome on into Covering the Spread. That's right here on the FanDual Podcast Network and NumberFire.com where today we are talking about the return of the EPL and soccer analytics with Ted Knutson of StatsBomb.com breaking down his past and his son of Pinnacle and letting you know how you can use soccer analytics to bet on the EPL. My name is Jim Saunas. I am a senior writer and analyst for NumberFire.com. Joined here as always by Ed Fang. You can find his work at thepowerrank.com. Ed, the EPL, back today. We have the Belmont Stakes coming up this weekend too. It's finally a good time for sports once again. How you doing? I'm doing good. Yeah. I'm excited about Premier League because, I don't know, none of my friends would come watch the Bundesliga with me last Saturday and I feel like because they're all into Premier League and I feel like when I text them again this week, they'll probably be up for watching some soccer so that is a good thing and yeah, looking forward to talking to Ted. Yeah. And we have, I think Ted's going to be a tremendous guest. You can find Ted on Twitter at mixedknuts, mixedknuts, really enjoying his last name. He is the co-founder of StatsBomb.com. We're going to talk about the services StatsBomb provides, what makes their expected goals metric different than some other expected goals metrics out there. We'll talk about the return of EPL, we're talking soccer analytics, his time with Pinnacle. So a wide-ranging interview here, but Ed, you've met Ted previously at Sloan, correct? Yeah. I usually see Bill Connolly and kind of follow him around a little bit at Sloan because he's the man and yeah, we ended up talking to Ted once in the hall. That's really the great part of going to Sloan, just meeting people and chatting people up in the hall. So yeah, and his reputation precedes him with his background on Pinnacle and it definitely knows the stuff. It was at Sloan where I first learned that Leicester City is pronounced the way it is because I am an ignorant American and had no idea, I was like, what's that word this person just said? And I realized it was, oh, it was Leicester from context and I was like, wow, I'm really behind on all this stuff. Well, I mean, you know, it's just, and there's funny pronunciation, so. And that's from someone who used to live it in Massachusetts. So like Worcester is the same thing, like who knows why that's pronounced the way it is. I have no idea, but it was definitely a good time. And probably two thirds of British folk probably didn't know how to pronounce it until they won the Premier League in 2015. That's also very true. We also, the Belmont Stakes coming up this weekend. Are you going to take a peek at that or will be mostly EPL stuff for you this weekend? I'll take a peek. Yeah, turn it on. Yeah, I can find it at the race. But yeah, I mean, I'm not much of a horse racing person. So I'll stick with soccer. If you are looking for some horse racing insights, we had John Sheeran, a Fando sportsbook on, I guess it would be a couple of months ago now. I mean, he talked about analytics you can use in betting horse race inside. I recommend you search for Cover in the Spread wherever you listen to your podcast. Search for the episode with John Sheeran and listen to that. He had some great resources you can use when trying to bet on horse racing. So make sure you check that out just by finding the Cover in the Spread podcast page wherever you listen to your podcast. And while you're there, leave a rating interview and subscribe as well. Speaking of the Belmont, the Belmont Stakes are coming up on Saturday and you can stream and bet all the races on the FanDuel Racing app. The FanDuel Racing app is specifically designed for sports fans who are new to horse racing and they've made it simple for anyone to learn the ropes. Best of all, right now, FanDuel Racing is giving new users 20 to one odds on any horse to win the Belmont Stakes. Just sign up and deposit on Friday or Saturday before the race. Then bet five dollars on any horse and get one hundred dollars if they win. That is the FanDuel Racing app, age and restrictions, residency restrictions apply, site credit expires after 30 days. Gambling problem called 1-800-522-4700 to get you set for the Belmont Stakes this weekend. Now, before we dive into our discussion with Ted, we're going to take a look back at last week as we both have Cover in the Past to discuss, mine on NASCAR, Ed's on the Bundesliga. Ed's went a little bit better than mine did. Covering the past. All right, so last week here on Cover in the Spread, we both had recommendations that we had. Mine was for the NASCAR Cup Series race in Martinsville. And I wanted Kevin Harvick at 12 to 1 to win Martinsville. The thought process was the current form is elite. The track history, not great, but also not bad enough to avoid him. And Harvick ran well for most of the night. He had a seventh place average running position. He was fifth at the end of the second stage and was generally running towards the front. But they completely whiffed on the setup in the final run. And the final run was a very long one. Harvick actually went a lap down, which is not great. But so that one did not go well. He was 12 to 1 to win that race. Seventh place average running position, you know, it's whatever. I don't feel great about that one, but I think the process was fine. You had, though, had a good one. You had Munchen Gladbach. Hopefully I'm saying that right against Bayern Munich. And at the time we talked, your numbers had Munchen Gladbach at 17 percent to win, and they were plus 700 to win outright. And you did get some good closing line value on that number. And it looked like it was going really well for a long time in that match. It was tied at one goal apiece. Then Bayern Munich got a goal with four minutes left to take the lead and win. So a super close match and you've got it at seven to one odds. It didn't work out. But I think you have to feel pretty good about the process behind that. Yeah, I feel good about the process. I mean, part of it was Bayern did not look nearly explosive without two of their best offensive players, Robert Lewandowski, Thomas Mueller. And when you look at the expected goals, it was 1.5 for Bayern to 0.8 for Munchen Gladbach. Munchen Gladbach is a good team. They're fifth in the table. They're definitely a solid Bundesliga team. It just it just looked like the Bayern Oswa were just way too high for missing two of their top guys. So yeah, no, I mean, I do feel good about it. It was interesting watching the game to see Bayern's defense like allow some pretty decent opportunities amongst amongst that. Expected goals. And I was kind of wondering if that kind of buried itself in the numbers. When I look at expected goals adjusted for who you played, you know, were they closer to the middle than you would expect the top team? No, they were still second, you know, very, very not too far behind wherever the top team was into Bundesliga. So it's always interesting for me to kind of compare, you know, eyes versus, you know, what the numbers say. Oh, and also like Bayern, I mean, they let some awful goals in against another match earlier where they didn't look like they can mark anybody on a set piece. So but they're still a very good team by the numbers. Yeah, absolutely. But a good bet there for sure. It didn't work out, but I think the process was good. And we're going to talk a little bit more about that match is we're going to talk some soccer now with Ted Knutz and follow him on Twitter at mixed nuts, M.I.X.K.E. or M.I.X.E.D. K.N.U.T.S. He's a co-founder, co-founder of Statsbom and we're going to talk about the return of the Premier League. We're going to talk about soccer analytics who sounded pinnacle, really fun interview coming up with Ted Knutson right now. Covering the present. Let's bring Ted Knutson into cover in the spread to talk a little bit about the English Premier League. Ted soccer is back. I guess football is back. How are you doing today? We can call it soccer. That's fine. I'm from that country originally, even if I haven't been there in about 13 years. Yeah. Yeah, I'm good. It is kind of like a big day because we finally have some EPL action back. So I appreciate you taking the time to talk to us when I feel like you're probably just like sitting there like bubbling with anticipation to actually see some real matches in the EPL again. Yeah, I'm stoked. This should have been Euro's time and and Euro's and the World Cup or like the the smorgasbord of betting. So I'm a little sad about that, but it's nice to have like any sort of football or soccer back. Yeah, absolutely. It's nice to have just just distractions back in our lives and being able to watch some sports because it's been it's been a long couple of months, that is for sure. Now, Ted, I think the interesting thing about having you on here is that you've worked in a couple of industries that are hyper relevant for our podcasts because we love numbers and we love betting and you've worked in both industries. You worked with Pinnacle back in the day. You now run a sports analytics company. So I think that you'd be a good person to ask about the overlap between the two between sports analytics and betting. How much of an overlap is there between those two industries? Pretty significant. So the reality of what both of them are looking for is better predictions, right? You want to be able to predict something better and it's something about the sports. And now, whether that is a particular outcome or whether it's a player and their valuation and their potential and all of that involved or even like a head coach, you know, you're you're hiring a head coach. What is the actual impact of that head coach based on their history, the talent they had access to and that the competition that they were in or the opposition they're playing against? Like all of these things come into play and you use data as one part of it. But most of the the syndicates that I've sort of been around in the past also have like a human layer on top that are making those final judgments because the humans often have external information that the models might not. Yeah, absolutely. I completely agree with that and how you need to add the human element. How has that evolved over time? I mean, you know, was the human element kind of more important back in the day when we had less data and now, you know, with all the stuff that you're doing at Statsbomb, is it getting less or have you seen that evolve? I think that there are some groups that have more confidence in their numbers and have probably diminished the human element. Some of them have also, like they realize that the models get weaker in some weeks and so they have to have more human oversight in those spaces. It really depends on like what your framework is for gambling and some some places I know the lines are much tighter than they used to be. Right. When I first started batting back in like 2005, 2006, they were soft and they were soft all over the place. And then I went into Pinnacle and we closed a lot of the holes, which I should apologize to everybody who came after us. That is at least partly my fault and I've taken money off of your plate. But you've got better information these days, too, and a lot more broad spread info than you would have had back in the day. Yeah, absolutely. I guess like while we're on the subject of Pinnacle, what were some of those gaps you were trying to close? You know, what did they task you with in order to make those lines more efficient, at least on their end? Oh, God. So when we started, probably came in and first things we fixed were like baseball team totals and first halves and things that were derivative lines that were tied to the main lines. So that was one part of it. Realizing that the one of the old sort of directors at Pinnacle kept putting up wrong money lines in the NBA versus the spread. Like he was inside too far towards the favorite. And we were just like, there's value in that if we just bet the plus number on the dog. And so like over the course of a year, my partner just picked off all the plus lines. He would keep an eye on it while that guy was trading because he wouldn't believe it. And we're like, no, no, this is the sheet. This is the whole sheet of like drive stuff. Like this is the money line for these spreads in these totals, like, ah, whatever, you know, the market's not telling me that. So like he just went in and he bracked of like two million in profits in his account just from picking off lines. We were telling them it was mathematically incorrect. And finally, they stopped hanging those lines. And so yeah, that was that was another thing. But what we did there was basically we took like, I don't know how much people know the background of Pentagon and stuff like that. But on January 9th, 2007, they cut off their entire US business in a single day. 70% of their business one day, they cashed them all out and they didn't take any more for quite a while. And I'm not gonna go into those stories because that was other people and they got punished for that. But they were probably the biggest US bookie at the time. And this is land-based or internet-based and they cut off all their US stuff. And we had to retool the entire product line to be international. And so we went in and we rebuilt entire sports and soccer was suddenly super important because everywhere that's not the United States, it's the biggest sport. Live-geek betting was something that we spent a lot of time and a lot of arguing about, whether that's on American sports or on soccer or what am I thinking, Eurobasket or stuff like that. Tennis, tennis was a huge one. So spent almost an entire year like turning around the tennis department and all sorts of other stuff. And I wasn't the only one, like we had lots of good people there. Many of them were still there got 12 years after the fact. So yeah, there's a lot of longevity at that place. Absolutely, that's really interesting and to hear everything that has to be done to try to bridge those gaps. But there are still some gaps needed to be bridged. And we can still find edges via sports analytics. So when you are looking at areas where markets may not fully account yet for analytics, which markets are the most advantageous? In other words, like, are there markets where analytics are still ahead of sportsbooks potentially giving us a lingering edge despite the improvements they've made there? Folks are absolutely, and this is true even in the biggest of markets. The fact of the matter is like the lines are imperfect, usually right up until about an hour before they close. And even then, plenty of times, they never quite get perfect because there's pushback and people who wanted a piece of the action got enough of it or they bled most of the value out of it, not all of it. So yeah, some things are ones where you have expertise that the boogie does not. So we used to have accounts that would come in at that Japanese baseball and Japanese soccer. And they just had this expertise about the Japanese league that we were never gonna have and couldn't read the language. So you're looking for translated info. And that's true throughout almost all of the different leagues and the smaller leagues. So if you're a D league expert or something like that, that's a place that the boogies are probably not gonna pay that much attention. But there might be enough value there that you're happy to make that at least your hobby, if nothing else. Yeah, absolutely. So Ted, tell us a little bit about Statsbom. You guys do soccer, analytics there. What services do you provide for betters and other people and what makes it unique? We're a data company. And our goal is to become like the world's best sports data company. And I know that that might be controversial to some people, but we're a startup and you gotta have big goals. And right now we're probably, and in fact we are like the best event data company in the space. So we produce event data, which is like 3,400 events a game for 70 competitions in 2020, which is a pretty massive amount of data. We're also tooling up to collect our second sport starting next month. And that'll, we'll probably talk a lot more about that as we get toward the end of the year, but we're pretty excited about it. And what we realized along the way, having been in both the gambling space and worked for some soccer teams is that the need for better data was actually pretty important. There were things that were being collected, there were relevant, there were things that weren't being collected that actually really relevant. And that's true in gambling and it's true in the sports space as well. So you walk into a coach's office and the coach is like, well you don't know these things because it's not in the data and they actually matter. And you're like, yeah, okay, you're probably right. And what we did was we just started listening to people. And this was true for gamblers as well. We have a number of professional gambling friends around the way and we talked to them as we design better data specs to say, hey, what would move the needle for you from like a betting perspective or to make it easier for you to talk to your coaches. And so in the soccer space, what we did was we added a bunch of things that are sort of unique selling points. Like in terms of our expected goals, we have the location of the goalkeeper as well as the location of all the defenders and the attackers around the ball for every shot. And that was like what we showed up with in 2018 when we launched the product. But we also continued to add things. So like we have past footedness. So we know which foot the pass came with. Shot footedness obviously we've got that. But we added a thing called shot impact height. So we know now the height of the ball when it was struck by the player either with their head or their feet. And there's a ripple effect that says that that's somewhat important. You used to have proxies, but there's a big difference between a header that happens at like six feet where you're able to get your head on the ball or something that seems off the top of your head at like eight feet tall. You made an amazing leap, but you really can't redirect it. And so these little sort of bits of better data then bring in the outliers. And the more you're able to model for outliers in like a gambling sense, the better and more accurate your prediction should be. So then the more likely you are to have customers that really like using your data. We also pay a lot of attention to quality, which is something that gamblers care about and sports teams care about, like you don't want to have errors inside of it because if your model sits on data, like data is the most important part of that even if you've got modeling approaches that try to account for maybe some variation inside of it. Yeah, and you guys supply the expected goals for football reference. Is that correct? We do have expected goals for FB ref, yeah. Yeah, no, that's excellent. Just take us through a little bit. I really love kind of the granular level that you guys are doing that. That's gonna be different from what I'm seeing on these Bundo Ziga broadcasts on like FS1, right? Yes, so that one is done by the Bundo Ziga themselves. And then I think I'll have to still collect the event data there. And there's some sort of pairing with AWS, which may be a marketing pairing. I don't know how much PFF use AWS either, but you know, it's there. We've seen it before. So yeah, like our stuff is still a bit different. And I've actually done some comparisons with the video that they launched to explain it. And I think they don't incorporate the same stuff that we do, or at least not at the same level. So I think ours is still a bit better, but I haven't seen a broad sample of that and that fits on that model. Well, it sounds like you've done a lot of work specifically with player tracking data. And we've seen player tracking data in other sports too, like baseball has done a lot with that in defensive metrics. They've been using a lot of player tracking data and we've seen next gen stats with NFL doing player tracking data too. So do you think that your company has an edge given your experience using player tracking data with soccer? I'm not saying you're going to expand to baseball or football, but do you feel like you have a built-in edge given the knowledge you have of all that player tracking data and what is relevant and what is not? So player tracking is one thing that has just started to be incorporated into a lot of the models. So like we actually are an event company that has, I would say that we're like a stepping stone in between the full tracking, but if you had to choose between the two, you would choose events because events is much richer and it gives you the context around things. Whereas having the spatial elements of everybody is very useful, especially if you're playing the sort of the game, but without knowing that the shots are there and the passes are there and all of these other things, they're missing. So we're in the process of sort of building hybrids between those using computer vision and technology along with sort of more traditional methods and a lot of human collectors to watch over the machines to make sure that we produce really good information there. In gambling right now, in some cases like I don't know how useful player tracking is in the NFL, right? Like I'm not sure about that because the context is super valuable and building a lot of sort of metrics and ratings inside of it. Maybe in soccer, it's a little more useful. Maybe in basketball, it's a little more useful. But yeah, like we're sort of, we're on this trajectory, we're starting very low and very dumb and I'm not saying that as like, we're ignorant, we don't know. And so as you bring in more stuff, you still have to be careful to make sure that it matches up to the things that were there historically and that your model calibration makes sense and you're still reflecting the actual goals in games and not just what you think they should be based on a lot of machine learning and potentially some black box approaches. So Ted, I mean, when you're talking about these event driven data, you're talking about real humans tracking these types of events, right? Yeah, so we do probably 80, 20 right now. It's humans that are collecting the events and then about 20% of the machine and we're gradually shifting over to the machine and having the humans sort of make sure that the machine doesn't screw up. And I think that that's the way that our sport's gonna go and I also think that that's the way that some of the other American sports that have not been that way will go. But there are so many closet cases where you actually need humans to be able to resolve that problem. And like say American football, there's just so much overlap, right? Like there are a lot of bodies that are on top of each other around each other all the time. And if you're just doing like an obstacle tracking or a fully automation, like there are plenty of cases that are just gonna be noise to a machine. It's really hard to pull that information out. Which I think is one reason why the NFL initially went with a chip based approach for their tracking because they needed to keep the body separate and obstacle was gonna have issues with it. Yeah, that's all. Thank you, Ted. That's exactly kind of where I was going. Like what do humans add that machines never could, right? So you're just saying there's some events that you could just never pull out of the data no matter what the granularity of it. Sure. And also just like the models and the machines are not good enough yet. And maybe that is like a video quality or not. But there are plenty of dirty secret tracking data companies that add like hundreds of humans to try and clean up their data and make sure that it's usable for everybody else. Yeah. No, wait, was I not supposed to say that? Sorry. You can say whatever you want. We're good here. Now, one of the services that Statsbomb provides is player level data that can translate across leaks. And I assume all the different events certainly helps inform that player level data. Do you utilize that individual player level data when you are trying to crank out a betting model as well? And if so, do you think a better position Statsbomb to react when there are player level changes within a team or a league given you have that individualized data? Sure. So that's like, you know, the value of the replacement in gambling model is significant. Like Lee and Almesi goes out. What is the impact of Almesi versus his backup? Now, Almesi is like, you know, a huge example. But I think when I was in the gambling space, we kind of realized that no one could really be worth more than a third of a goal. It's like a team impact. And that's like a lot of study. And, you know, there may be outliers, but in general, that's kind of like the maximum. And this was during Almesi's prime and Ronaldo's prime and stuff like that when we were working on that. So there are other elements that, you know, some places and some groups may incorporate, particularly with regard to goalkeeper quality. Like if your goalie is not very good and you've got a pretty good idea that he's not very good, like that's gonna degrade that model pretty significantly. Detaching that as a skill element from, you know, the team XG is a big deal. And there's, you know, there's some art in that as well as the science. Finishing ability is probably a little bit too overrated. You know, your ability to put the ball away, like outside of Messi who's totally broken. And we have all of Messi's data available for free. We neglected the lean on Messi data biography, released it to everybody for free. And he's just not a human. He's a bit like Steph. Like Steph hits shots that other people cannot do. And Messi does that same thing for like, you know, 16 years and it shouldn't be possible, but he is. But everybody else in the world, yes, usually you fit within a certain set of bounds and you might be on the top end of that. But it's really not a massive impact on, especially the gambling lines, because you only get, you know, say three shots a game at most of the elite levels. And so like it's not as big as people sort of estimate, but the goalkeeper impact I think is probably pretty big. Yeah, that's interesting. I mean, so if one player is only worth at most a third of a goal, something like Robert Lewandowski is going to be one of the best goal scorers in the world. Both he and Thomas Mueller were out. Is there like collective effects there? Having both those guys out, because they didn't look explosive against Munch and Glabak at all this past weekend. Well, Glabak's really good though. I mean, don't sleep on those guys. Not least because like one of my early scouts is the assistant head coach there. Oh. But they're the cascade effects. So if you take out like two or three players, especially that are pretty important, then like that team is very different. And that is something that we ran into year after year and the gaming side back when that used to be my career. So yeah, like one player missing, but the rest of the starting lineup there, pretty confident that you have the right value estimate if there are multiples missing and it feels like a big drop off, like that line can move up ton. Yeah, no, it was interesting. And I mean, Bayern was still a huge favorite in that match. There's probably a lot of other effects playing there because they did win and clinched the Bundesliga. But I thought it was interesting that, according to some numbers I had, it didn't look like there was an adjustment for an absence of probably the two best attacking players. Yeah, that's not a huge surprise. Like sometimes these adjustments happen really late. Also, right now, and I know this is on your agenda that you want to talk about, but like the home field thing is scary. Yeah. And people are not sure how to deal with this. And if you're a big better, you usually want real confidence that you understand like what's happening. And so putting down full of ounces is terrifying at the moment. But we don't know what's going on, right? Like, you know, I looked over at 538 and they're reducing home field by 10%. That feels too low. That feels way too low. I mean, just without the data, and we know that there's been a negative home field, I mean, a negative, road teams have scored more goals since the Bundesliga came back. I don't think there's an away team advantage, but. I think many people would like it if 538 ran a book. And that's in multiple sports. That's really funny. So let's talk about that, Ted. Not specifically about the 538, but let's talk about the EPL resuming because they will have no fans. And we have some data. It's not robust data from the Bundesliga and now LaLiga being back too, but how much are you thinking here as far as baking in an advantage to the home team, if any? How much of an adjustment would you be making if you were trying to handicap these things now with the way that things stand? Given that we've got a reasonable amount of Bundesliga data, it's possible that home is just zeroed out. Just nothing, right? And that's, it might be worse than that, right? Cause like you tend to play a little more aggressive at home and teams can't do that anymore. And so like maybe it's even messed up some stylistic elements. But yeah, I, you know, I would be scared of putting large numbers on this, but I would think that, you know, instead of reducing it by 10%, you're looking at anywhere between 50 and 80% at a minimum and that's enormous. Yeah, for sure. I mean, in some of the models I'm doing, I'm just assuming it's zero, just as kind of a first guess. I mean- And this is the fight, right? So there was this huge fight in the Premier League from all the teams that are relegation threatened and they knew that they weren't gonna be able to have fans, but they still wanted to have home advantage. And there's like reasonable conclusions to suggest at least that, that's irrelevant. Like you guys shouldn't have thought about this. Just one more thing on the home advantage, you know, we know from research that, you know, the ref's subconscious bias plays a little bit of a role in that I feel like there is potentially still a little bit of that. I think back to when there probably should have been a penalty called against Dortmund at home against Bayern in the very last minute. It wasn't called. Bayern still ended up winning that game. I would bet that's probably that 20% remaining home advantage that you think is still kind of there, right? Just still there. Possibly, but you know, not, you hear so much more too, right? I think for the refs, like this is the thing that's underestimated. Like when you have a big open stadium, you hear everything. And it used to be that the crowds, especially in the Bundesliga are so loud that like you don't hear a lot of stuff. Like, you don't necessarily hear a crunching tackle as much. You see it. Like you don't hear the moment a ball touches or a foot touches a ball or anything like that. Now, like it's all there. And also like that you've seen some ref head snap around like, oh, I heard the coach tell him, you know, he said the C word or something like that, that, that. But yeah, it's like, well, you know, it's like bullderm, right? Did he call it a C word? Yeah, but yeah, it's a little, I don't know, it's really difficult to estimate. And I think that even so, like I've done a lot of research I've read on this internally, externally from different syndicates even about home advantage and why they think it exists. And you know, there was still this, the huge element of uncertainty as to why, but we knew it was an impact. And we knew it varies across leagues. Like there is a crowd factor to some extent. There's also like a travel factor. You know, the United States research was always a big deal. Like the time zone factors, altitude is always going to be a thing. So like altitude is obvious, but like there aren't that many places that have significant altitude. Now, like you just have to say, yes, the crowd has like a really quite large impact and it has to be mostly on the ref and their ability to make effective evaluations of things that happen in a split second. It'll be really interesting to get that data. Once we have travel effects in place for larger countries and stuff too. So I am looking forward to that data from this fall, assuming that we do get American football and college football. So we are recording this Ted right before the Aston Villa Sheffield United match. So let's look ahead to Friday and Saturday and let you look at some games. Any markets you're seeing is being advantageous based on the numbers you see for this weekend's game, weekend's matches in the English Premier League right now. So we're going to go with Ed's estimation that there's very little home advantage. Is that what we're feeling? It works for me. Okay. I mean, Ted, that's just an assumption I'm making these numbers that I posted on my site. It is a, it is a, what shall we call it? It is a prior guess? Yes. That's fine. You know, I'm going to roll with this. Like basically I'm giving any good things that go wrong here. They're yours. That sounds good. Yeah, so let's see. One of the sort of sleeper teams in recent times has been Southampton. Southampton are considerably better than their position in the table. And their numbers make them look like they should basically be like seven or eight at least in our stuff. And so like anytime you get a chance to bet on Southampton, like it looks pretty sexy, Norwich, like, you know, I'm not sure that I'm excited by that number though. So like, you know, maybe it's a good estimate. I came into this wanting to bet on Southampton. And then I looked at the lines and I got scared. But the other estimate that you have to be a little concerned about is basically how much of these teams changed, right? Like it's been three months since they played. Plenty of these, you know, could have gone through like another full preseason. A lot of time with new managers, like in Arsenal, where that Brighton line is juicy right now. I think that I would definitely want to bet on Brighton if we had not had a three month break. And so if you think that, you know, Arsenal are not gonna overcome and they're not gonna be that much better than they were, which I think there's a chance that they probably can be. They had such a long layoff and they were tactically a bit of a mess, so our time has more time. But Brighton, even so, like I'm kind of excited about the Brighton line, I think that one's pretty good. Beyond that, you know, again, you're getting into the situation of how much is home taking over or not here? How much do you underrate Liverpool? Because they've been outperforming their numbers by so much for the entire year. And, you know, some of these are personal things. Like there's a certain point at which it's a judgment call and you're off the model map. You know, anytime I get a chance to bet against Newcastle, like I'm gonna be excited, they've been way worse than their league position. And, you know, I do not believe in the fighting Steve Bruce's. I, that's just, that's me, you know. Yeah, I don't know, do you have an opinion? Is there anything that's good here? Yeah, so I'm just looking at a couple of my numbers, which Ted, are your numbers really, because I take the data from football reference. I adjust with Lee Square, so I have some schedule adjustments in there. And I assume no home field and just run a plus on model on it. So I do have Brighton at about 36% to win over Arsenal. So that does certainly suggest value in that game. And then Newcastle, I have about 23% to win. And they're plus 240. Sheffield United, I think. So that would suggest some value on Sheffield. So Sheffield is 51% and they're plus 130. So. Yeah, definitely to Sheffield United when it looks pretty juicy. So see, look at that. I don't know if it's because we have the same data source or not. Yeah, no, my model is super simple. I would never bet just based on it. And so rely on the experts like you that watch all these matches and I'm gonna be getting back into it now as well too. Cool. That should be awesome. Ted, we're gonna let you go so you can go off and watch some soccer. But we appreciate you taking the time today to swing by, talk through all this and talk through your interesting backstory. And I appreciate the time. Enjoy watching some soccer once again and hopefully we can talk to you again soon. Great, thanks a lot guys. Enjoy your summer. Thank you. Covering the future. Big thank you once again to Ted Knudson for swinging by and breaking down the EPL and soccer analytics. Make sure you follow Ted on Twitter at mixed nuts, KNUTS, and go to statsbomb.com to check out his website and add some smack talk from Ted in there. I love it. Let's get some rivalries going here on Covering the Spread. Yeah, for sure. 538 starting a sports book is a pretty good line. Definitely enjoyed that one. Hey, I mean, I respect what those guys do, right? I definitely consult their site and I'm kind of the beginning user of some of their soccer predictions as well but I will continue to follow that. But what I did mention with reducing home field by 10% seems unideal to me. Yeah, I have no idea what the number would be if I were to put it in there but it'd be larger than 10%. I am a devout listener of the 538 Politics Podcast so again, overall respect them but I am all in favor of a smack talk. No objections to me personally so definitely a fun interview with Ted but overall I think it's fun to hear the differences and expected goal calculations and he had a Twitter thread about that earlier this week and the logic behind it in trying to calculate the outliers and expected goals I thought was really interesting. So I think, you know, just find it his Twitter account because there's some good thoughts there available for free and stuff. So it was fun. It was fun to talk to him about it. Yeah, absolutely. Let's move on here to covering the future for this week and Ed, you were talking a little bit about your Premier League numbers in our discussion with Ted. You wanna talk about those for covering the future as well? I talked to him about it a little bit in the interview with Ted. It's not very complicated. I take Ted's data, which anyone can get in a nice formatted CSV file from footballreference.com. I plug them. I adjust or shrink the schedule with a least squares algorithm. That gives me ratings for offenses and defenses. And then based on that you can figure out expected goals in a match between two teams. And honestly, I gotta thank Austin cast the other week for letting me know that that data was available because it completely eliminated the need to scrape any data. So when I do this, the schedule adjustments and the expected goals let me rank teams both offense, defense and as a team. And we all know Liverpool has a huge lead in the table. Very unlikely that they're not gonna win the Premier League but they're really not the best team. Man City is the top team when I look by adjusted goals, sorry, expected goals. And it's not just that they're the best team it's the margin. They're almost a half goal better than Liverpool when you look at it. So this isn't much of a surprise. Austin cast told us that a couple of weeks ago. It's kind of how they excel that I thought was kind of interesting. Like they really excel on offense. They have an offensive rating of 2.32 goals. So that means they would be expected. They would have an expected goals of 2.32 against an average defense in the Premier League. And that's just enormous compared to an average expected goals of about 1.35. And you really see like how much better the best teams are at offense. That's really where they excel. And just to put that in perspective like the worst team has an offensive rating of about 0.92 goals. So it really shows how good Man City is. And we see the same thing in the Bundesliga as well. When you look at Bayern Munich expected goals after adjustments there at 2.43. And so these top teams are really good on the offensive side of the ball. Talked about how Liverpool might be a little overrated. This is something that we see in the predictions as well. My numbers have them at a 54% chance to beat Everton which is less than the minus 165. They are a fan dual sports book when I checked this morning. The other team that I thought was interesting is Lester City. So this was the team that made all the headlines a few years back in 2016 when they got promoted to the Premier League and then won at some huge underdog to do that. They are third in the table so they're still a pretty good side. But the numbers suggest that they're overrated there at third in the table. When I look at expected goals adjusted for who you played they are down to sixth. That suggests that there would be a little bit of value in maybe going against Lester City but that's actually not what the predictions say. I'll have them about 54% to beat Watford. And that actually suggests value on Lester City because there are plus 115 in the markets. There may be something else going on there. I'm not exactly sure what it is. So I'm a little bit new at making these soccer predictions. You can check them all out at thepowerrank.com slash predictions. I have them for Bundesliga and Premier League. It's fun to hear about the numbers right when you start up a model because I'm always kind of nervous like giving my stuff out to the public when I don't have a lot of confidence in it but it's nice that the transparency, hearing about your process and going through it and the inputs that you had for it. So it'll be interesting to see how things play out because it's just nice to have new data to look at I think. Yeah, for sure. And it definitely gives me a lot of confidence that the teams that Ted liked, the numbers did seem to support. Yeah, absolutely. If you can have someone as smart as Ted siding with you, I feel like you're off to a good start there for sure. For my cover in the future, I wanna talk about NASCAR once again because there's not a whole lot else going on. I mean, outside of, I mean, as far as sports that I know very well because I don't know a lot of horse racing, not super up to speed on soccer yet. We're getting there, slowly increasing the knowledge but I wanna talk about NASCAR this week in Talladega and it is a high speed pack racing track and those tracks are generally super hard to predict. So for each race, I track the correlation between each driver's finish and their mark in my model entering the race. And if you look back to last year for all 36 races, the two Talladega races ranked 32nd and 33rd respectively in terms of the finishes and the correlation between the driver rating and where they finished in the race. The only race is worse than those Talladega races were in Indianapolis and the two in Daytona. So pack racing tracks terrible in terms of predictability for my model. So it does better here than Daytona but there is a lot of randomness in there. As such, I know I can't have a ton of confidence in the favorites and if I want to bet on Talladega, I gotta swipe at some long shots and the three I like the most this week are Eric Almerola, Clint Boyer and Matt DiBenedetto. Both Almerola and Boyer are 26 to one. They drive for Stuart Haas racing and teammates matter a lot at this track because teammates work together, manufacturers work together at tracks like Talladega and Daytona and Stuart Haas racing is among the best teams at these tracks. They basically throttled the entire field here back in the fall of 2018. Almerola won that race, Boyer finished second and Almerola also has a win at Daytona back when he was driving for Petty Motorsports back when he had much worse equipment than he has now. Almerola, seven straight top tens in Talladega. Four of those have been top fives that is borderline impossible to do with all the randomness here. So of the three drivers I like Almerola is my favorite bet this week at 26 to one. Boyer generally takes a weird strategy and runs toward the back of the pack for most of the race. He does so intentionally because it's gonna look like a bad bet for 90% of the race but he keeps himself out of trouble and ensures he is live for that final lap. But that's so that he has a legit car it doesn't crash before the end. He does have only one top 10 in six races in Talladega with Stuart Haas racing but he has the equipment he has the teammates to get the job done. So the Almerola one I feel a lot better about but also in a Boyer at 26 to one D'Benedetto a bit longer at 40 to one but again, going with the good teammate narrative here because he is effectively teammates with Ryan Blaney, Joey Logano and Brad keselowski and all three of those guys are 11 to one or shorter to win this race and D'Benedetto is 40 to one. He has proven his skills on pack racing tracks in the past. He's been good in Daytona. It's not a Talladega but his equipment now will be the best he has ever had and he finally has some teammates he can work with as well. So D'Benedetto is good if you want someone a bit longer at 40 to one. Looking at my model, it does have Chase Elliott at the top. It also in theory likes Ricky Sanas Jr. and Tyler Redick at their current odds but the reason I'm not gonna back them is that they're all Chevy's and the Chevy teams struggle to link bumpers in Daytona and I'm worried that could carry over to this weekend as well where bump drafting is a pretty major thing. So broadly not super into the Chevy's but the forwards of Almerola, Boyer and D'Benedetto are my favorite options here. And then I think this is one of the situations where the eye test does kind of matter and it forces me to lower where I'm at relative to my model on the Chevy's because the eye test at Daytona like they did well in the first age but they didn't give me a lot of confidence. So here is the situation where I really do want to put some weight in the eye test. Yeah, for sure. And I think, you know, the best of the best like use both numbers and the eye test. And very necessary for a weird race like this. And if you want to like get a weird taste in NASCAR and watch a really exciting race just in general, I would recommend Talladega. It is nuts. It is absurd. Like they're just free wide the entire race at times. So I would recommend checking it out. So you watch Belmont Saturday. Watch the Premier League on Friday, Saturday. Watch some NASCAR on Sunday. Not a bad little weekend for you right there. Yeah, for sure. What time do I tune in? Where do I tune in for NASCAR Sunday? Three o'clock on Fox, I believe. I think they have an Xfinity series race on Saturday too. So if you want like a little teaser before the Belmont you could do that as well. Ed, you had an interesting show over at the football analytics show this week. What was that? I appreciate that. It is titled, What is the probability of football in 2020? I think it is a question that everyone kind of wants an answer to. Yeah, I had a data scientist from the University of Michigan on who looks at population dynamics and we talked it through. And yeah, you can find the answers. We actually do put some numbers behind it. You know, a little bit of a spoiler. The chances are better in the NFL than in college. Not going to be surprised to anyone that follows this. But you know, I try to do different things on that show. This was very different from just talking sports predictions with someone who I know is an expert at the field. This was more, you know, it was interesting. Like there's no one that models like the probability of a second outbreak. Like it's not something that people do or at least not something that I could find. So, you know, there's a lot of predictions out there that are kind of beyond the scope of even the models. And, you know, we try to try to talk through those things and figure out how to put some odds behind it. Yeah, it's hard to have a second wave when the first wave never really ended. I think that's that's like, well, but the thing is like, you just don't want the number of cases to start ramping up in a desperate, you know, like in a dramatic way again. And, you know, we're kind of seeing that in certain parts of the country. We actually talk a lot about that heterogeneity and how it can affect football this fall, which I think was, you know, one of the more interesting parts of the conversation. So, yeah, it's a reasonable short. It was only like about 35 minutes. So definitely suggest checking it out on the football analytics show. You can get it wherever you get your podcasts. All right. Anything else going on for you over at the power bank for this week? Nothing much. You can always make sure you get all notifications by signing up for my free email newsletter. You can do that at thepowerrank.com slash predictions and sign up for their free email newsletter. And then you can check out all the soccer predictions for Bundesliga and Premier League there as well. Yeah, get those numbers at powerrank.com. Make sure you follow Ed on Twitter as well at the Power Rank. I am at Jim Sonnis, J-I-M-S-A-N-N-E-S. You can also follow the Fanduel Podcast Network at Fanduel Podcast. A lot of DFS stuff this week. We have, we already had an EPL podcast with Austin Cast. We had him back on for the DFS side of things to talk about how you play EPL DFS, because I had no idea. So I talked to Austin about that. He went through the way lineups work and all that. Make sure you check that out. We had a PGA podcast with Brandon Gedula who did really well, almost did really well last week for PGA because he was on Xander Shafley and Colin Mora Kawa. They both had bad puts at the end to lose but they were in contention the entire time. So that is available on the Number Fire Daily Fantasy podcast feed talking to Austin Swain later today about UFC and a NASCAR podcast on Friday. So find that on the Number Fire Daily Fantasy podcast feed. Big thank you once again to Ted Knudsen for swinging by and talking about everything that he has done in his past, talking about statsbomb.com. Make sure you follow Ted on Twitter at mixed nuts. Once again, nuts with a K there. Thank you to Calvin Theobald, our video producer for running the video side of things here today as always. Thank you, Cal. And thank you to everyone for tuning in. Enjoy the EPL, enjoy the Belmont Stakes, enjoy La Liga, enjoy NASCAR, whatever it may be, golf. We hope you have a good time and hopefully you are safe and healthy out there. We'll talk to you once again in the near future. This has been covering the spread right here on the FanDuel Podcast Network.