Why the Texans and Seahawks Stand Out in 2026

Knowledge at Wharton

Ask about Knowledge at Wharton

Episode Description

ABOUT THE EPISODE Which NFL teams have the strongest case to exceed expectations this season?

Aaron Schatz, Editor-in-Chief of Football Outsiders, explains why his projections are high on the Houston Texans and Seattle Seahawks, while examining the Los Angeles Rams, Buffalo Bills, Philadelphia Eagles, Cincinnati Bengals, and more. He also breaks down how quarterback play, coaching changes, injuries, schedule strength, and DVOA shape an NFL forecast.

ABOUT THE PODCAST Sports is a game of numbers. Wharton experts Eric Bradlow, Shane Jensen, Cade Massey, and Adi Wyner team up to tackle the world of sports, from current events to longstanding issues such as: What sports streaks are the most impressive? How do you rank the best players? Can athletes be compared across sports? Moneyball Highlights explains how decision-makers in the game can avoid the common mistakes and embrace the data. Episodes are recorded at the Wharton School.

----- The Wharton School of the University of Pennsylvania is committed to sharing its intellectual capital through Knowledge at Wharton, the school’s online business analysis journal.

Launched in 1999, Knowledge at Wharton offers free access to:

- Articles, podcasts, and videos highlighting Wharton faculty research and analysis of current business trends - Interviews with book authors from Wharton School Press - In-depth series of curated content like Wharton Executive Education’s Nano Tools for Leaders - A searchable database of more than 10,000 articles covering all aspects of business - A weekly newsletter that delivers Knowledge at Wharton insights directly to your inbox

Visit the Knowledge at Wharton homepage: https://whr.tn/3qUpSBE Subscribe to the Knowledge at Wharton YouTube channel: https://whr.tn/3DEAgAU

Read the full transcriptHide the full transcript
  1. Welcome. Welcome to Wharton Moneyball. Welcome to another full hour of sports analytics
  2. here on the Wharton Podcast Network. This is Cade Massey hosting this week, along
  3. with, so far, two of my longtime collaborators, Shane Jensen and Eric Bradlow are both here.
  4. Audie Weiner, our fourth collaborator, will be floating in here momentarily. We're going to have our typical format this
  5. week. We're going to have a guest in the first half of the show. For any of you watching this on YouTube,
  6. you know who that guest is. We'll get to him in one minute. We'll have open lines, open topics, second half
  7. of the show. We're recording on Tuesday afternoon, as we typically do. The show will go up on Wednesday.
  8. Gentlemen, it's a very, very happy day in Wharton Moneyball land. We have the first football guest of the
  9. year. We're not going to quite go 100% football, 100% of the time at this beginning now.
  10. We're going to work in some baseball next week. We're coming up on a stretch where we'll
  11. go all football, but we want to start early. Why not start early? They're playing games.
  12. Everyone's in camp. Pros in camp. Colleges are in camp. To kick us off, we have asked Aaron
  13. Schatz to join us. Happily, he was able to before I introduce him. Aaron, good to see you.
  14. Thanks for making time for us. Hey, it's always good to be on. Yes, we made it through another offseason.
  15. Quarterback battles are being raged. Guys are getting injured. It's all the news that you love.
  16. Contract extensions, everything you love about the preseason this season. That's right.
  17. Well, for those of you who don't know Aaron, let me give him just a little bit of background.
  18. Aaron was the creator of Football Outsiders back in the early 2000s. Along with that, he introduced a new stat,
  19. DVOA, which we should know right offhand, but Defense Adjusted Value Over Average, continues to publish,
  20. update, refine, find new uses for DVOA. He is now with FTN, and he also is with ESPN.
  21. He does a fair bit of writing on ESPN. He's got a podcast out there, Schatz and Tenier podcast, Schatz and Tenier NFL podcast.
  22. You can catch him a couple times a week talking NFL analytics with his protege, Mike
  23. Tenier. Aaron's also famous in my world for having hired and developed two of the best football
  24. writers in the universe, analytics-informed, analytics-savvy football writers in Bill Barnwell and Bill Connolly,
  25. maybe the leading two analytics guys writing in NFL and college, respectively. Also, Brian Fremo, among others.
  26. Aaron really, really advanced the conversation around analytics and continues to.
  27. Aaron, you should know that you're one of our favorite to start the NFL season with.
  28. We get you on these preview shows. We walk away every time going, oh man, that was good.
  29. So thanks for making time for us. Appreciate it. How are you doing? I'm honored.
  30. I'm good. It's been a good off-season. We did our preseason annual. It came out a little earlier this year
  31. than usual. This is what it looks like. It's the FTN Football Almanac 2026. In print, hard copy.
  32. You can buy it at Amazon, but also FTN fantasy.com slash almanac to get the PDF version.
  33. And that's my sales pitch. My sales pitch is basically me babbling for the next half hour, but that's, you find
  34. all of our stuff and all of my writers in the book and I'm ready to go, man. It's been a good, it's been a good
  35. year. I have a new hobby. My new hobby is old school football cards. I have had so much fun writing about
  36. the old play-by-play that I went out and bought old football cards with all the players that I've been writing about so
  37. that I can like see what they did in their off-season jobs in the seventies and whatnot.
  38. I was going to say, what counts as old? How old are we talking about? My oldest set so far is 78.
  39. Okay. That was peak, man. I think putting your 10, 11, that's like peak for sports imprinting in your brain.
  40. And that's how those, those late seventies years, by the way, that was 78 was I think Campbell's, Earl Campbell's first year in the
  41. NFL, number one draft pick. Yeah. So I believe he's a, yeah, 78 was his first year.
  42. So I have just a comment, Aaron, on that. I always like to comment on, I, I was, guys know my story, but first let
  43. me ask a question to you, Aaron. If I was to take that, what looks like very thick and, but good FTN document
  44. you just put up on the screen. And I was to put it into an AI engine and ask it, what are the
  45. top three things that this Almanac says for the upcoming season? What would it say?
  46. But what would be the big highlights? Do you think of the FTN? You know, if I asked it to summarize,
  47. I don't know, 400 pages, whatever it is into three big bullet points, what would it
  48. say? As a concept, you could also have a human summarize it. Yeah. Well, I'm asking Aaron to summarize that too.
  49. Just that, that the Seahawks and Rams are awesome. That Houston is our favorite in the AFC.
  50. What? That's our surprise. Oh my goodness. And I'm trying to think of like, what's
  51. the like analytical trope that we repeat over and over again. Maybe the quarterbacks have more to do with
  52. the sacks than the offensive line does, but the offensive line has more to do with
  53. rushing than the running back does. That's nice. That's nice. Wow. Very good. All right.
  54. Well, let's take some of those. I just, yeah. Cause since you're talking about the importance of
  55. quarterbacks, are you, I assume you're expecting a big jump back from CJ Stroud for that
  56. Houston prediction is tantalizing. I got to say, you're expecting a lot more from CJ Stroud this year.
  57. Those playoff games were so bad. He was 11th, 11th in past DVOA during the regular season.
  58. And the Texans as a team were 15 because a couple of Davis Mills games, even though they won those games, drags them down
  59. a little bit. All they need to do if they have a better running game and a better offensive
  60. line and tank Dell back is have an above average offense. And if their defense does what we think
  61. it can do, you know, caveats about defense being harder to predict than offense, but that's
  62. the best team in the AFC if their offense is above average. Wow. So you're, you're jumping them above some of
  63. the faves, long-time faves, and they seem to be favorites in most markets outside of
  64. yours. So clearly, obviously we're talking about the Bills, Ravens, Patriots, the Patriots, maybe too much into
  65. quarterback performance in the playoffs specifically. And if you take away quarterback performance in
  66. the playoffs, specifically, the Patriots have one of the best quarterbacks from last year.
  67. It's not that you ignore those games. It's that for Stroud, for example, that's only
  68. two games, he played what like 16 games last year and 14 of them, he was slightly above average.
  69. And then two of them, he was horrible. Aaron, talk to us a little bit about the quarterback is such an important part of
  70. the forecast. What is it, what is your model for quarterback performance? And are you still, is Stroud still improving?
  71. Is the arc that they improve in those early years? Do you, does he still get a boost
  72. because he was a high draft pick? Like how are you modeling his performance in 2026?
  73. No, I think by the time you get to year four, it doesn't do that anymore. It doesn't look at your draft status anymore.
  74. So it's just looking at his performance over the last three years. And what I try to do is you
  75. can't do a projection that looks at the quarterback completely separated from team variables.
  76. But I try to do that as much as possible. And then that projection goes into the team
  77. projection. So that way, when a team does change quarterbacks, you know, you know that Minnesota has
  78. a better quarterback this year, you know that when Denver went from Tim Tebow to Peyton
  79. Manning in 2014, or whatever it was, 2012, 2012, it was a huge jump, right? So you account for those things by trying
  80. to do the quarterback separate and put the quarterback into the projection. For the most part, my projections are more
  81. top down than they are bottom up. I don't like project every player and then add them together.
  82. I'm projecting the team. The only player who gets projected separately is the quarterback.
  83. Yeah, which is reasonable, right? Because we can't model all the interactions. And we know that it's not just some
  84. additive thing. Shane has a question for you. Well, so just since you're talking about it,
  85. can you talk about a little bit like how you approach a problem of projecting, say,
  86. Patrick Mahomes this upcoming season, where you've got obviously a multitude of good performance, but he's
  87. also coming off a really severe injury? How do you treat kind of cases like that?
  88. Yeah, I don't end up with enough guys who struggle after injuries to have an injury
  89. variable that means that you struggle after an injury. Now, obviously, different injuries have different amounts of
  90. struggle, right? Like Achilles, we have two examples in recent years of old quarterbacks coming off Achilles injuries
  91. and having no mobility at all, which is Aaron Rodgers and Kirk Cousins. But you try to throw old quarterbacks coming
  92. off an Achilles injury into a projection system and you have a sample size of two,
  93. it's not going to give you a variable that means anything. But we know, like when we look at
  94. Daniel Jones, we go, oh, he may have a problem with his mobility, although all the
  95. reports out of Colts camp, including from national reporters, not just from Colts reporters, are that
  96. he looks fine. All right, what about, here's another modeling challenge for you, what about new coaches?
  97. So our Baltimore Ravens are looking at a new coach, how do you fold that into your consideration?
  98. Yeah, it's tough because most coaches don't have a track record of improving teams or units,
  99. right? Like if you have a guy like Vic Fangio, he's been in like eight different places
  100. and everywhere he goes, the defense gets better. If you have, say, Mike McDaniel, you have
  101. a sample size of one. He's coached one offense. So I know we all believe that Mike
  102. McDaniel is a very good offensive coordinator and we can talk to film people and they
  103. will wax rhapsodic about his run designs and whatnot. But when it comes to what you can
  104. actually prove objectively, you can't prove, there's not enough sample to prove that Mike McDaniel means
  105. anything. So what you end up with is that in general, teams with new coordinators will take
  106. a little bit of a step back when they're learning new systems, all other things being
  107. equal. And you don't really account for specific coaches as much as you want to believe, like
  108. Jesse Minter was very good at running the defense with the Chargers. So you want to believe that he'll be
  109. very good at running the Ravens. But history is filled with very good coordinators
  110. who didn't turn around that side of the ball when they became head coaches. Like most guys who get hired to become
  111. head coaches were very good coordinators, right? It's not like Jesse Minter is the first
  112. very good coordinator to get a head coaching job. Yeah, they typically select from the good coordinator
  113. pool, not the bad coordinator pool when trying to hire a new coach. That's right.
  114. And some of them have failed. But besides Bill Belichick and the Cleveland Browns
  115. too, he failed. But I don't want to get to that right now. But let's talk about something else.
  116. You talked, that was just for Shane's benefit. You had talked about the top-down approach.
  117. Could you tell us as you get projections for these, let's call it higher level variables,
  118. how do you think about what weight to combine them to an overall team number? Is it based on out-of-sample predictive
  119. value? Or how do you even if you could predict offense grade, defense grade, maybe quarterback grade,
  120. whatever the major factors are, how much weight would you put on those as you roll
  121. it up? I mean, the weight is sort of based on the range of how things come out. I do a lot of regression analysis and
  122. offense comes out with a wider range than defense. I mean, that's how it works during the
  123. season, but that's definitely how- Just to be clear, the dependent variable is predicting the
  124. number of wins? Is that what you're trying to- Oh, no, no, no. What I do is I predict offense, defense,
  125. and special teams all separate. Yep. And then I add them together and I simulate the season with- Simple additive.
  126. I'm just saying there's no different weight to offense, defense. I'm asking how much weight do you put
  127. on offense, defense, and special teams when determining the game outcomes or the- Right.
  128. The weight is the range, right? The best offenses are always stronger than the best defenses.
  129. So, in a way, that's the weight. Special teams is much smaller range, right? So, especially with the projections, offenses will project
  130. from, I don't know, I'm making up these numbers, but like 15% to minus 15 % DVOA.
  131. Special teams will project from two to minus two. So, obviously, special teams aren't changing things very
  132. much. This is a very interesting. We've never, I don't think we've ever in our 12 plus years on the show, I
  133. don't know that we've ever had this discussion of what I'll call almost like self-norming
  134. weights in the sense of you're adding components together that have different degrees of variance and
  135. variability. And then almost by definition, something I do in quantitative marketing, the range defines the importance
  136. of a variable and something that's flat over, if I have different, let's call it, how
  137. much do you love TV sets? And if across the variation of TV sets, the range is low, then that's not an
  138. important factor. And it's that similar premise. It's interesting. I don't disagree with it.
  139. I'm just saying, I don't know that we've ever had a formal discussion of I'll call
  140. it inherent scaling being a way to weight things. Yeah. I talk about it in the book, but
  141. if you look at the range of my ratings, you end up with the idea that football, a football team is basically four parts
  142. offense, three parts defense, and one part specialty. And other people like Brian Burke have done
  143. completely separate research and ended up with the same four to three ratio, except they didn't
  144. do special teams, but the same four to three offense to defense ratio. Mm-hmm. Very interesting.
  145. Let's talk a little bit about some specific teams. I'd be curious what the boys are interested
  146. in hearing about. I want to start with the Rams. If you ask an AI to go pull current market probabilities, implied probabilities anyway, the Rams
  147. are head and shoulders above everybody else, 13 .5% or so to win the Super Bowl.
  148. Next to right neck and neck with each other are the Seahawks and the Bills, and then the Ravens right behind them.
  149. So a big two on each side, Rams and Seahawks on one side, Bills and Ravens on the other, but the Rams categorically different
  150. from the others. Are you buying into that level of optimism about the Rams? Can you break down the Rams for us
  151. a little bit? I feel like it's a little too high. The Seahawks were so good last year.
  152. DVOA including the playoffs, the Seahawks were the fourth best team I've ever measured.
  153. That's how good they were last year. So even if you come back a lot, I still ended up with them higher than
  154. the Rams. Although if Aaron Donald comes out of retirement, that's enough to tweak the personnel change variable
  155. so that the Rams would move ahead. But surprisingly, I still have the Seahawks slightly
  156. ahead and the Rams have a slightly harder schedule as far as who they play outside
  157. their division. If you ask me who's my personal Super Bowl favorite, I still like the Rams as
  158. the favorite, but not to the extent that the market does. Can you tell us a little bit more
  159. about the Seahawks doing so well by your DVOA analysis? Can you give us more insight into that
  160. and to what extent do you expect some of that to persist versus progress? Yeah, they were like eighth or something in
  161. offense and they were the best defense and they were number two on special teams.
  162. So special teams correlates from year to year much less than offense and defense.
  163. So that very likely comes back to the pack somewhat. But their offense should be pretty good again.
  164. Obviously, we don't have it projected that high because they lost their offensive coordinator.
  165. And their defense looks like one of the best in the league. We have it as our number one defensive
  166. projection. Now, I've been saying on some shows that in some ways, Seahawks remind me a little
  167. bit of 1991 Washington, which was a team that didn't have any particular superstars.
  168. I mean, no disrespect to Daryl Green and Art Monk, but didn't have a superstar quarterback,
  169. but was really good, well-rounded, historically good. And the next year, they did take a
  170. bit of a step back. And I wonder if the Seahawks will have the same kind of thing happen because they
  171. don't have the superstar quarterback and because they don't necessarily have certain players.
  172. Obviously, Smith and Jigba last year was amazing, but it was the first year of him
  173. being like that and everything. But as far as what my formulas say, they say, no, expect the Seahawks to be
  174. one of the two best teams in the league. All right. And just for checking my history, 91 Redskins,
  175. was that the wide right Super Bowl? No, they beat Buffalo by a couple of touchdowns.
  176. It was the first Buffalo Giants for the wide right. The wide right was the Giants.
  177. Yeah, it was the second. It's the second Buffalo lost 1991. Yeah, because the third and fourth were the
  178. Cowboys, right? Ugly Cowboy games. Yeah, yeah, yeah. Okay, very good. Just pinning down my memory.
  179. Yeah, so that was the first one. All right. What about here? Y'all get ready because here's my last
  180. one for a while. Bengals, as an NFC North guy, I'm legit respect for Burrow. He's been hurt, but I also have legit
  181. disrespect for the organization. Now, sometimes they get lucky with a bunch of great personnel.
  182. How does that all net out this year, do you think? The thing about the Bengals is that you
  183. will hear a lot about how their defense improved in the second half of last year.
  184. But in general, second half improvement does not indicate a carry over to the following year.
  185. And I know it drives people crazy when I talk about this, because it goes completely
  186. against common sense. And it goes completely against the narratives that we like as writers.
  187. But a good evidence of this is last year's Bengals, because the Bengals got better in
  188. the second half of 2024 on defense, and they came out in 2025 and were horrific.
  189. So their defense is not necessarily going to be that much better other than normal regression
  190. to the mean. But their offense should be fantastic with Burrow healthy. And they have, according to my numbers, the
  191. easiest projected schedule in the league. Ah, okay. Although the Ravens are like fifth easiest.
  192. They played the NFC South, which is- Yeah, right. Well, I don't think it's- I was
  193. going to say, the NFC, you usually don't think of the AFC North as an easy schedule kind of division, but I guess it
  194. just aligns this year. Yeah, they played the NFC South, which- Well, since you brought up the NFC South
  195. and my bucks, what the hell, I might as well ask. I could make an argument for all four
  196. teams in the NFC South this year, or I could say they all stink, and so someone has to win that putrid division.
  197. How do you see the NFC South this year? Yeah, our numbers like the bucks to rebound.
  198. I know some people are really high on the Saints. Our Saints projections came out really bad, like
  199. thinks their defense is going to regress back to what they were the year before, and
  200. doesn't really like Shuck. He wasn't really that good last year. You're talking about a second-round pedigree rather
  201. than a first-round pedigree. Carolina is what they are, and Atlanta kind of is what they are.
  202. So we like the bucks. The bucks come out as the favorites of the NFC South by our numbers.
  203. What's your take on Pennix? Is Pennix going to be a real quarterback? I know it's early still.
  204. Yeah, I mean, it's early still, but the injury-wise, it sounds like Tagovailoa is the
  205. starter. I don't know if Pennix is going to be ready for the start of the year, even if they wanted him to be the
  206. starter. So I projected them with Tagovailoa as the starter, but it's not like there's much of
  207. a difference between the two of them as far as expectation. I mean, Pennix, because he's younger, has a
  208. wider range of possibilities. Bills, give us some hope. Yeah, my numbers are surprisingly high on their
  209. defense. That defense has been good for a couple of years until last year. They've added some talent.
  210. Their offense should be great again. I like the Bills. I think we have the Bills second in
  211. the AFC behind Houston. We have them as the favorites in the AFC East ahead of the Patriots.
  212. Should be really good again. I don't know if their defense will really be top 10.
  213. I mean, they do have a new coordinator, so they got to learn some new stuff. But yeah, their defense came out surprisingly high
  214. in the projections. It has to do with what downs and distances they were good on last year.
  215. Okay. Let me ask a question. Which team has the max min? Which team has the highest war?
  216. The reason I was thinking about that was the Bills. If you told me the Bills win less
  217. than, I don't know, nine games, I'd be pretty surprised. Maybe Seattle and Rams are in that range
  218. too, but yeah. The answer is my projections come out with less of a range of ranges than you
  219. would think. I've got to, you know, it's one of those things I've always wanted to fiddle with
  220. a little bit to do a better job of making it so that certain teams have a wider range of possibilities because I'm the
  221. quarterback or whatever. So the best teams are the ones that really have the highest minimums too.
  222. Like the Rams and Seahawks were so good last year that it's very hard to see them collapsing because they were so good.
  223. Kneed would know this because of, you know, Mass EP, buddy, and all these other systems.
  224. But like every year we look at baseball, we believe that the, you know, distribution is
  225. always compressed. My guess is this is symptomatic of lots of forecasts where the aggregate distribution of forecasts,
  226. when one looks at it, one would say there's no obvious way to fix it, but there's a lot of compression that tends to
  227. happen. Okay. Yeah. I think the only thing that would kind of oppose that compression is the importance of
  228. the quarterback and quarterback injury. If that was something that was factored into
  229. the projections. Some of it is, look, we know that some team will go 14 and three, but
  230. we have no idea who it's going to be. So you don't end up with anybody, you know, USA Today or whatever does their projections,
  231. which are subjective. And somebody goes, oh, this team will be 14 and three. But in reality, even the best team, the
  232. chances that Seattle actually goes nine and eight are strong enough that we only end up
  233. with them with 11.4 average wins in our projections. Yeah. And I guess I was thinking more, less
  234. about like the expected kind of value, more about the, like, cause we were talking about
  235. the floor. I think the floor is the thing that really, I think would be influenced by like,
  236. you know, specifically quarterback health. Yeah. I don't end up with a bigger range
  237. of from the ceiling to the floor than I wish. I wish certain teams, it would make logical
  238. sense for certain teams to have a wider range. Aaron, what storylines can you point us to,
  239. especially that you'll be paying attention to early in the season? Like you're curious how this is going to
  240. work out. Do you think this might work out well, or this might work out poorly? Give us a few storylines around the league
  241. that you think are worth attention. Daniel Jones's Achilles, because the Colts were so
  242. good when he was healthy in the first half of last year. And then he fell off after he had
  243. a leg injury and then he hurt the Achilles and the, and the mobility is so important for him that that's going to really
  244. be something to watch. Am I right about Houston's offense is obviously a big one.
  245. Can the Dallas and Cincinnati defenses be good enough to make those real playoff contenders given
  246. that their offenses are strong? Can Travis Hunter play two ways? Still a question and an interesting story.
  247. And the Packers waiting for Micah Parsons will be an interesting story. The whole NFC North has terrible, terrible depth
  248. charts at edge rusher after the starters. Every one of those teams has bad depth,
  249. but the Packers are like really bad. Like until they get Parsons back and it's, Oh my God.
  250. Aaron, let me ask you a question. We've been talking for 20 something minutes about
  251. the NFL and we've yet to talk about the team that we live here in Philadelphia, the Superbowl champion from just two years ago.
  252. Is there any reason we're not talking about the Philadelphia Eagles and should we be talking
  253. more about the Philadelphia Eagles? I think it's sort of assumed that they're the favorite in their division.
  254. And yet it's also sort of assumed that because their offense is with a completely new
  255. scheme this year and traded away their best receiver, that their offense will probably not be
  256. good enough for them to be a real Superbowl contender. So I think they're in this sort of
  257. middle ground. That's not as interesting to talk about where they're not like one of the top teams,
  258. but they should win their division. So yeah, people aren't talking about them that
  259. much. If you talk to people who are film heads, they're very interested in the idea that
  260. they're going to put Jalen Hurts in an offense with more under center play action that
  261. should look more like what the Packers do, because that's where Sean Mannion, the offensive coordinator
  262. is coming from. So from a film watching and scheme perspective, it's very interesting.
  263. And the defense will just do its thing. Fangio's still there. The talented players are all still there pretty
  264. much. So, I mean, they re-signed a big extension for Jalen Carter and Mitchell and Cooper
  265. Dejean are still there and everything. But if I told you, Aaron, because of the, I'll call it the weakness in the
  266. NFC East, that let's imagine I tell you right now, the season's over and the Eagles
  267. end up as the, well, anybody with the number one seed has to be the favorite because they get the buy.
  268. But let's pretend that the Eagles are the number one seed, because it's not necessarily because
  269. they're better in the Shatz projections, but they ended up, they were in a weaker division.
  270. They went six and O in their division. The Giants potentially are no good. Commanders regressed.
  271. Cowboys are no good. So they end up, I'll make it up 13 and four and end up by de
  272. facto the number one seed because everybody in the NFC West beats up on each other.
  273. The NFC North beats up on each other. NFC South stinks. Aren't they then the favorite to win the
  274. Super Bowl? Yes. Okay. The number one seed, it would be very difficult. It would be very difficult given the current
  275. playoffs setup to go into the playoffs and have the number one seed not be the conference favorite, because even if the number two
  276. seed is much better than the number one seed, there's still whatever, 25, 30% chance
  277. they're going to lose in that first game. And the number one seed doesn't have to
  278. play a first game. Well, I've just come up with my theory about how the Eagles go in the Super
  279. Bowl, because they're in a weaker division. They have an opportunity of the teams that
  280. aren't the elite teams, potentially. They seem to me to be, well, I'm sure you guys do some sort of projection
  281. forward. Where do you have them finishing in the conference? I think like third or fourth.
  282. They have a tough out of division schedule. They play the NFC West and the AFC South, the AFC South, which is better than
  283. you think it is. Okay. Interesting. Well, Eric, if your story comes true, you'll
  284. be happy to know that Aaron thinks they're going to get the Texans in the Super Bowl.
  285. So then they'd be the favorite as well. So you're good to go. The D'Amico Ryans Yeah.
  286. We're going to let you go. But one last question. Give us one team that has been lately
  287. quite strong that you think will step back this year and one team other than the
  288. Texans that you think might really step up this year and exceed people's expectations.
  289. We surprisingly have the Minnesota defense regressing to the Mead. Now they lost some players, but I do
  290. think Brian Flores is such a good coordinator that he can prevent that. But as far as what the numbers say,
  291. even though their offense will be better, that probably is the answer for a team that
  292. we have dropping. And then getting better a team that people don't think of a lot.
  293. I mean, it's tough to do because a lot of teams last year were really good in DVOA, but not necessarily in win loss
  294. records like Detroit and Kansas City before the Mahomes injury. So the fact that we think Detroit will
  295. be good and Kansas City will be good, it's not like those are not teams that have been good in the past.
  296. They just weren't last year. Okay. Got it. All right. Well, this was exactly the roundup that we
  297. needed. This this primes the pump for the next couple of weeks. Thank you for making time for us, Aaron.
  298. We wish you the best with this early season run up. Good fun. You guys should track Aaron's work down.
  299. You can find him on FTN. You can find him on ESPN. You can find his Almanac, the FTN Almanac.
  300. You can find him on his podcast with Mike Tenier. A reminder, it's ftnfantasy.com slash almanac.
  301. And there is a sample chapter. So if you want to see what all the numbers look like for an average team
  302. and what our writing is like, you get to read about the Colts. And then if you like that, you can
  303. buy the book and get all the teams. Awesome. Great. Aaron, thank you for the time.
  304. Much appreciated. All right. You guys come back and join us after the break. Welcome back.
  305. Welcome back to Wharton Moneyball. Welcome to the second half of this week's show.
  306. Gabe Massey here hosting with the whole crew. Eric Bradlow is here. Shane Jensen is here.
  307. Audie Weiner is here. I can tell they're all in Philadelphia. Different places in Philadelphia, but they're all there.
  308. I'm away for the time being, just off the lines with Aaron Schatz, a longtime football
  309. analytics writer. Always fun to catch up with Aaron. He lives and breathes NFL, so he can
  310. talk about the league. And this time of year, that's what we need. We are on the Wharton Podcast Network.
  311. We're going to be up on Wednesday. We're going to roll into an open line segment now.
  312. Before we do that, on the way to that, let's fold Audie in. Audie, with some technical challenges, wasn't able to
  313. jump in the conversation, but he was doing research in parallel with our conversation.
  314. What did you find, Aud? I got three pieces of things that I was able to research while you guys were
  315. all talking. So what I did was I assembled a data set of the over underlines for win
  316. totals in the preseason, and then correlated that with the end of the season. How big a sample?
  317. How many years did you have to do that? 12 years. So we have 32 teams times 20. We got 639 usable values.
  318. 2011 Colts, couldn't find the actual, we know the actual value. We couldn't find the predicted value.
  319. But what's surprising is that the, well, maybe it's not so surprising. The correlation is 0.5, which is pretty
  320. bad. It's not, I mean, I mean, it's clear signal there. When you say it's bad. Yeah.
  321. It's not, it's just, it's a huge amount of noise. Like this isn't the physics, this is humanity.
  322. These are people. All right. I mean, like, like, like a social scientist would never call that bad.
  323. Yeah. 0.5 in social science is pretty good. We have, we have good teams and we have bad teams and we're actually pretty good.
  324. Actually, one of the things that's interesting about the correlation, I don't get into the lesson
  325. in statistics, but there was very little overlap between the teams we predict to be elite
  326. and the teams we predict to be bad. We get them easily split without, we're basically,
  327. that's an interesting, that'd be an interesting way to operationalize it instead of correlation.
  328. Just ask the following, Adi, like, what's the pairwise probability that the team you predicted to
  329. be better than the other team is actually better than the other team, right? Between any two teams, what's the probability?
  330. Right. So, so when you compare the number one team versus the 32nd, the odds of the
  331. 32nd has more wins than the first is really small, even though the overall... I mean, all pairs.
  332. All pairs. So I'm going to ask you a question. Let's say instead, you don't even know this,
  333. Adi. The first paper that I ever wrote was a paper with Shel Doxham, who was visiting
  334. Harvard at the time when I was a PhD student. It's called a correlation curve. So the idea is you don't actually compute
  335. one single correlation because it assumes a constant relationship along the range of X.
  336. You compute a curve that looks at a window and then computes a correlation across a
  337. window. So if you were to do that, like, would, like, this is the classic thing, like
  338. at the SAT, which you made, this was Don Rubin's favorite example. If you look at high achieving people, there's
  339. very little correlation between their SAT scores and their GPA. But overall, there's a strong correlation.
  340. But at the upper end of the distribution, it's weak. And there's all kinds of reasons for that.
  341. If we looked at your data, would there be a stronger correlation at different parts of
  342. the curve? Well, yeah. I mean, it's called range restriction is the first thing you do.
  343. But if you don't do range exacerbation, most people don't do that. So in other words, if I look only
  344. among the elite teams, there's very little correlation in their actual values in the wind because
  345. I've range restricted the elite. If I look only at the elite and the terrible, there's a huge correlation.
  346. It's gonna be like 0.9, 0.95, right? Or even higher, right? But when you look at every team together,
  347. which is typically what we do, then you get a correlation of a half. But let me get into my two other
  348. findings while I did with the background. One is there's huge bias at the top. So a correlation matrix should be unbiased.
  349. The regression curve should go through the vertical strips without showing any residual pattern, we would
  350. call that. But at the top, the regression curve misses the data pretty badly. There's about a half a win bias.
  351. We overestimate the top teams by about a half a win. At least the odds are doing that.
  352. These aren't models. These are betting odds that people have put together. So that's kind of interesting.
  353. And it relates to what Aaron was talking about. We're not having enough. We are too romantic about the top teams.
  354. And every year, there's these injuries and things happen, or they just don't do what we
  355. expect them to do. And you have this huge bias, bias you don't see in the other.
  356. You've introduced me to some nice words over the years and some statistical concepts.
  357. But this is a new one. The too romantic statistical concept is a new one for me.
  358. I just came up with that one. Is it more than just sort of like the market even reading some of that luck
  359. from the previous season into skill? Eric created a real narrative for the Eagles
  360. being the top team in the NFL in this upcoming season. And there was a lot of assumptions, heavy
  361. assumptions along the way. The third thing I just looked up was Aaron's very hot on Houston.
  362. Did he say he thinks they're the best team or what? Did he repeat that? He had them coming out of the AFC.
  363. Right. Okay. So I don't think the rest of the football world is quite as hot, although they
  364. do think they're a good team. So that's one we can throw around. But I remember years ago putting together the
  365. simplest regression forecast that looked at the last three wins and just did a weighted average,
  366. not equal weighted average. Last three seasons? Last three seasons win totals and found that
  367. works just about as good as the odds, the good fork. It's a little bit worse, but not much.
  368. And that's just back of the envelope, basically do no thinking. And you get almost all the signal by
  369. looking at your last three years and almost all that signal is last year. Yeah. Right.
  370. Okay. Good stuff. Thanks, man. Maybe we should have an RA sitting here running real-time research for all of our
  371. conversations. How do you get some good work in the first half of the show? Let's run around the greater world of sports
  372. and ask what caught your eye. Gentlemen, there's a lot going on, not just football.
  373. Soon enough, we'll be inundated in football. What else? What else you got? Well, I saw that.
  374. Oh. We're going to talk baseball, right? Right, Shane? Yes. Yeah, yeah. Well, Masters have just entered the top 10
  375. all-time in strikeouts, which is worth noting. And the kind of observation I have is
  376. I don't think that that's a top 10 locked in now for all time. There will not be anybody else joining that
  377. top 10. No. I agree. Okay. That's because we don't see the same innings pitch, but it led me to a question
  378. even about Scourger. It led me to a question about him. Isn't he already... Here's my question.
  379. Can you, for each of those top 10 guys, give us the strikeouts per batters faced?
  380. And presumably, Scourger will be relatively high since even though he's been in the league for
  381. a while, he's 42 years old, he has played half his career in an era where pitchers don't see that many batters.
  382. So aren't I even more impressed with him being in the top 10, even though he has been around for a long time?
  383. So, Baskett, how much variation is there among that top 10 for whatever the right statistic
  384. is for strikeouts per batter faced? Yeah. I mean, I think it is. Verlander and Scherzer have played long enough that
  385. they're still able to eke into the top 10, even though I think, I bet you if you looked at their innings pitch, it's
  386. probably gone down over the years, just the same as kind of the general starting pitcher.
  387. It's just that they had more than half their careers prior to that. Okay. I'm trying to understand the question.
  388. You're kind of like, they were able to do this in kind of our current pitching conditions.
  389. They did it in maybe like two-thirds of their career was not in these current pitching conditions.
  390. Well, it raised the general question of strikeouts per inning or strikeouts per batter, how much
  391. variation there is in that top 10? So basically, the pitchers more recently have been
  392. striking out more batters per night. And that's because they throw fewer innings and
  393. a lot fewer innings, but their canines are on average higher. In fact, among the top 10, Randy Johnson
  394. has the highest among the top 10. But Scherzer is essentially imperceptibly behind 10.61
  395. versus 10.55. The second highest out of the 10. So that was my intuition. Right.
  396. And the lowest on the list is Gaylord Perry. I mean, he just had so many damn innings that he managed to be in the
  397. top 10 all-time strikeouts. But, you know, Roger Clemens was 8.55, Norman Ryan 9.55. Brian Levin's in there
  398. too, I think. Yeah, Brian Levin is certainly. He's 6.7. So if you look at the guys today, they're striking out 11-12.
  399. Yeah. Okay. Okay. Now you've raised a name from the past, Gaylord Perry. I remember him pitching in his 40s, right?
  400. Did he end up throwing knuckleballs at the end? Am I making that up? Who? Perry?
  401. Yeah. He did. I mean, that's extraordinary. But he added that late in his career,
  402. did he not? Or was he always a knuckleballer? He wasn't an exclusive knuckleballer.
  403. I don't know the answer to that. I'll remember as a baseball card enthusiast, on
  404. the backs of the cards is where you'd find all the statistics. And Gaylord Perry had so much statistics, you'd
  405. need a magnifying glass to read it. He looked like an old man on the mound, is my memory.
  406. He was gray. Okay. All right. Shane, I'll cut you off after you mentioned Scherzer.
  407. Well, yeah. No, I mean, I think it was just an observation. I mean, I guess a little bit of
  408. a lament that we're kind of facing. We're kind of enjoy the era of great starting pitchers as it starts to kind of
  409. essentially phase out. So, Shane, let me ask you a question related to that. So, I think you've already been on air
  410. a couple of weeks ago. Obviously, none of us believe there'll be a 300-game winner again.
  411. Now you've said potentially no 3,000 strikeout or certainly, let's call it 3,500 to
  412. get to the top 10. Yeah. Chris Dale might make it to 3,000. Yeah. On the hitting side, do you guys, Adi,
  413. Shane, do you guys see hitters getting to 3,000 hits? Clearly, we see people getting to 500 home
  414. runs. There'll be a lot of people getting to 500 home runs. But is 3,000 hits within range anymore?
  415. I mean, does anybody hit anymore? Is anybody out there hitting non-home runs anymore?
  416. I don't know. Soraya is doing it. I mean, he's hitting for the Red Sox now. No, he's hitting for the Phillies right now.
  417. Yeah. I mean, with the DH, I kind of feel like the longevity, there's not quite the
  418. longevity kind of constraint that we're seeing among pitchers for hitters. But certainly, the way the game's changed where
  419. specifically getting hits is not necessarily the most prioritized thing out there.
  420. Can Bryce Harper get 1,100 more hits in his career? He could. He could. No. You think so?
  421. How old is he? He's judging 19. 33. Yeah, he's only 33. Judges are 35 years old, so he came
  422. up early. Harper averaged 150 hits a year for the next seven years. Sure he can, right?
  423. Can. I wouldn't bet on that. Tough. That's tough. What about Machado, too? He's pretty young.
  424. Anyway, it could happen. I think there's at least a chance of that, whereas we're basically saying this sport would
  425. have to change incredibly substantially back to what it used to be for there to be
  426. 300-game winners or 3,500-strike. I mean, it's almost terrible. I'm looking forward to that reversion, Shane.
  427. It would be glorious. We have an easy way to do it, which is just chop two or three pitchers
  428. off the roster. I love that fix. I love that fix. Adi, what you got? What caught your eye?
  429. I got a bunch of things. So I'll start with one that happened last night, which is an answer to a trivia
  430. question, which is how can you have four outs in an inning? Was this in the Red Sox-Blue Jays
  431. game? Well, you could have a strike out. No, it's not an out until it's thrown out.
  432. It's an amazing asset called a timing play. And what happened yesterday actually wasn't an actual
  433. throwout, but it's possible. So basically, what happened is the Red Sox had runners on first and third, and a
  434. fly ball was hit to center field or right field. And they thought there were two outs.
  435. And that's because the scoreboard said there were two outs, but it was an error.
  436. There's only one out. And so when the fielder for the Blue Jays caught the ball, the Red Sox runners
  437. ran on contact and the runner on third scores, and the runner on first is at second base.
  438. And the right fielder threw to first base, doubling off the runner on first, creating a
  439. third out. But then they had to do a replay to judge whether or not the runner scored
  440. before the out. Because it's called a timing play. So if he scored before they doubled him
  441. off of first, that run would count. But here's where the fourth out could come. If they had doubled off the guy at
  442. first creating the third out, then challenged the guy who ran from home from third base.
  443. If he ran early, they could get him out. And that would obviate, they would remove the
  444. run, and that would be four outs in an inning. It's amazing. Technically, it's three because the runner at second
  445. would not... If any of our listeners are confused by all this, by the way, I was watching
  446. this game live, and it took them about 25 minutes real time to figure this out. To figure out the policy in New York
  447. for the umpires of the game to figure this out. So... Yep. But the mistake that the Blue Jays did
  448. was they should have challenged the runner on third for going early, and they didn't do
  449. that. There's about 10 different mistakes that happen for this scenario to technically evolve.
  450. But that was very fun, and a highly unusual event. So, Adi, have you moved your baseball viewing
  451. from the Yankees to Red Sox Jays? Oh, no. Just John Boy Media and V. No, they're exciting teams right now.
  452. Yankees didn't play last night. Yeah. Hey, what about this news? I just saw that Apollo, our private equity
  453. firm Apollo, of sorts, just took a big piece of the Yankees. Surely, sometime a valuation's going to come out.
  454. You don't buy without a valuation. Certainly, Apollo's not going to buy without a
  455. valuation. So we're going to get a valuation on the Yankees at some point. I got to imagine 2.6 billion, which
  456. is the number I saw. I got to believe... I'm just making 30%? I thought it was capped.
  457. I thought equity was capped at like 15 % or something of a franchise. I thought there was some room.
  458. But different sports do have different caps, and I don't know what it is in baseball
  459. right now. I thought it was 15% baseball, but I'm not sure. I mean, the thing about the sport is
  460. only one owner can be the managing owner. By the way, according to Forbes... Hold on.
  461. Just work with Shane. If Shane's right and just say they got the full... That's a huge valuation.
  462. Multiply it out. Look, Forbes in 2026 just published the Yankees' estimated franchise value at 8.5 billion.
  463. And so the most recent one by Sportico, 9.4 billion. So I'm going to stay with my number
  464. that they bought 30% of the team. I'm going to go lower. I'm going to go lower. Now the question is, why did they do
  465. that? Now you have a 7.8, which is below those two numbers. And these things just go north, north, north,
  466. no? I mean, I guess kind of from the consumer side, why do we care? Occasionally we're a business school show.
  467. Every now and then. And these valuations, especially with John... I mean, actually, it's interesting that Apollo is
  468. in there. It's Apollo Sports Management, which is different than Harris-Blitzer Sports, obviously.
  469. They're related, so it's interesting how it works. All right, Eric, what you got?
  470. What cut you off? So a bunch of things. I keep going back to what's obviously the
  471. excitement in Philadelphia with a starting five of Tyrese Maxey, BJ Edgecombe, you know, Jalen Brown,
  472. LeBron James, and Joel Embiid. So it seems like an amazing team, except for a couple things concern me.
  473. Maxey is not a good defensive player. Jalen Brown is a turnover machine. And if you look at advanced stats, not
  474. a great defensive player. LeBron James is 42 and no longer a very good defensive player.
  475. And Joel Embiid is no longer a very good defensive player. So one concern I have is, are they
  476. going to give up 120 points a game? They may score a bunch, but they may give up a bunch.
  477. That was one concern. And second, I've always had the concern in the NBA, I call it the there's only
  478. one basketball hypothesis. And so someone's going to have to, maybe two of those four stars, let's assume BJ
  479. Edgecombe's not yet a star, although I love his game. Maxey, Brown, James, and Embiid all average 25
  480. plus a game, roughly. Well, all four of them ain't scoring, ain't even taking enough shots to score 25 a
  481. game. So which two of them are going to give up shots for the greatness of the team?
  482. And that's the part I'm intrigued to see. And I think it'll be LeBron because LeBron,
  483. most people have said throughout his career, he's much more of a willing passer than shooter.
  484. So I think he'd rather have 15 assists than score 25 at this point of his career.
  485. And I don't know who that other person is. But that's the part I'm intrigued to see.
  486. And part of it was because the NBA schedule just came out. And of course, the opening game of the
  487. season is Sixers at Knicks, which is, of course, a fascinating game to start out the
  488. season. But that's what caught my eye. Sixers at Knicks is a good start. That's great.
  489. Remind us when the NBA season starts. Mid-October? October 20th, I believe, is the first game
  490. of the season. And they always have the one game that night, and it's Sixers at Knicks.
  491. But yeah, October 20th is the first game of the season. Of course, it's peppered in other things.
  492. Of course, it's Sixers-Lakers on Christmas Day. So that, of course, is exciting and other
  493. things like that. I see. When does the hockey season start, Shane? It's just a little bit before basketball, I
  494. believe. I think it might be earlier. Is that the window where all sports are going?
  495. Yes, October is, I think, the peak. Actually, it's September 29th, Shane. It's the first game.
  496. I knew it was near the start of October. You get 10 days or so, one week to 10 days, where all four sports are
  497. going. Yeah, well, it'll be a full month, right? Baseball goes almost this far. Yeah, baseball goes forever now.
  498. Basketball starts October 20th, so... Oh, right, yeah. Yeah, it'll be whatever. However, October 20th to the end of the
  499. World Series. That's right. I'm going to go... We should call that a national holiday.
  500. Come on, guys. A week long. A week long. Too much to do. I can't go to work. Let me do a quick one, then we'll
  501. do one more round. I thought I wasn't interested in the FedEx Cup, the end of season playoff, but I
  502. think I am going to get pulled in. Maybe one of the reasons is because they've done away with this staggered thing for the
  503. final tournament. They have a points buildup for the guys that don't follow it. I don't follow it that closely.
  504. Points build up over the year for who qualifies. They just made their first cut after this
  505. last weekend, the top 70 on the points advanced to play next week. Points matter to get into the round of
  506. 70, and then they'll cut to the round of 50, and then that next tournament cuts to 30.
  507. It's this tournament level cut, which is really interesting. Interesting design.
  508. It's usually a two round cut. Of course, now it's a full tournament four round cut.
  509. The first one's been made. We have two more to go. When they get to that final 30 golfer
  510. field, they throw the points out. In the past, they started with head starts and stuff, and it's been weird and wonky.
  511. Up until this year, they retained still some of the points system, even in the final.
  512. You're trying to guess who's actually ahead based on... Someone could win the last tournament, but not
  513. win the FedEx Cup, which happened, by the way, like Tiger Woods won the FedEx Cup
  514. one year, but Justin Rose won this, or vice versa, that kind of crazy thing. It does lead to some weird probabilities right
  515. now. Scotty is by far more than twice as likely as Rory to win the whole thing, but that's because he's got so many points.
  516. There's no way he's not making the final field. You have to kind of model it at two stage.
  517. Are you going to make the final 30? Right. You've got a much more... Like the final round, you've got a much
  518. more uniform distribution where... That's right. As a matter of fact, the other thing
  519. I like about the design, by the way, Cade, is that there is a player, I forget who it is, that is injured.
  520. The first week is 69. They actually don't replace the person. As a matter of fact, some players, you
  521. can imagine, let's take Shuffler. He could say, I don't even want to play. Matter of fact, Shuffler could decide, I don't
  522. want to play the next two weeks. Just being around the 30, and you know what? He's in no worse position because they don't
  523. do the points thing anymore. Enjoy your boys. Enjoy the boys. You've won over the course of your season,
  524. basically. He could. That's going to be interesting to see over time whether people start doing that a little
  525. bit. All right. We have time for another quick round. A quick round, guys. Maybe we'll catch Eric first because he's about
  526. to slip out of your air. Yeah. So the other thing I was starting to think about, I made a prediction two and
  527. a half years ago that I thought Djokovic would never win another major. However, Sinner's now injured.
  528. And so he's not playing in the Cincinnati Open here at his ankle, which is the week before the US Open.
  529. We know Algaraz still hasn't played since April. So if I told you right now that
  530. Sinner and Algaraz are compromised going into the US Open, maybe Algaraz is not even going
  531. to play. We don't know. All of a sudden, the odds of Djokovic winning another major don't seem so far-fetched.
  532. It goes back to even what we talked to Aaron Schatz about. You know, everything seems so likely until injuries
  533. happen. And then all of a sudden, we're like, how is Djokovic going to beat Algaraz and
  534. Sinner? Well, maybe he's not. And so maybe he doesn't have to. And by the way, I'm not assuming that
  535. Djokovic is going to beat. There's lots of young guys coming up now. There's Feast.
  536. There's Jodar. There's all kinds of guys, Shelton, that could beat him. But I'm just commenting, if you told me
  537. that Algaraz and Sinner are compromised significantly, Djokovic has to be the favorite at the US
  538. Open. He'd have to. Even over like a Zverev? Really? I didn't know that. That's a good question.
  539. Great question. I would say the odds makers would probably have Zverev. Zverev still is the favorite.
  540. But I think if Djokovic played Zverev, I think almost everybody would probably pick Djokovic in
  541. that match. Eric, this is just one more reminder to me that I feel like I always want
  542. to know, Grand Slams, for sure, that's what matters. I get it, for sure. But I also want to know Grand Slams
  543. normed for level of competition. There should be a unit, a standard unit for like how many of the top competitors
  544. were actually in that tournament. Because it varies greatly in tennis for exactly
  545. this injury reason. And everybody's Grand Slams are not equal. Whenever I get it, I get it.
  546. But I still want that. I want that additional information on how weighty the tournament was.
  547. OK. I mean, I'll have to run in just a second. But all I'll say is Djokovic won 24
  548. Grand Slams in an era with Nadal and Federer. So that's all you really need to know.
  549. And I understand. And by the way, you can say, well, he's padded it the last couple of years.
  550. No, he hasn't won one the last couple of years. So most of them were against peak Federer
  551. and peak Nadal. So, you know, in some sense, it's almost like, you know, I don't need to tell
  552. you what the entire distribution of other players was because there was Federer and Nadal.
  553. And by the way, Wawrinka was no joke. Andy Murray was no joke. So I'm just saying he came up against
  554. some very, very stiff competition. Odd. What do you got? I got a lot because I just finished,
  555. you know, a week, you know, Moneyball projects always give you great ideas, even though they're
  556. done by high school kids. They find very interesting stuff. Just to make sure our listeners know, you've
  557. been teaching multiple weeks in the Moneyball Academy. And so you've been neck deep in all
  558. of this stuff. So I'll think of just a couple that are off the top of my head. One of the best groups did an interesting
  559. study of how soccer players change their behavior after a yellow card. Because once they got a yellow card, they
  560. don't want to get it ejected. So the style of their play changes. And you can measure that.
  561. There was a very interesting piece of work done by a group out in the San Francisco group that studied how golf players age,
  562. their aging curve broken down by the different component of their game. So you have your driving, you have your
  563. approach, you have there's a there's the fourth stage, third stage. I forget the name of that.
  564. And then putting and what's fascinating about that. I thought that the golfs would have a
  565. higher peak than other sports because it's not it's not like other sports in lots of
  566. ways. But now it peaks just the same place most sports do. And they go down steadily in beginning in
  567. there at 30 as they get older. I thought that was remarkable. So that's the high level story.
  568. What about the the facet level story? Were there aspects of their game that peak
  569. earlier? Yeah, actually, surprisingly, so putting doesn't change that much. In fact, there's hardly any knowledge that there's
  570. much inputting, as we talked about before. But they do their different parts to do
  571. change. I think that that third one, which I forget, maybe you can tell me what's that
  572. short game, but the short game, it's like that goes badly. I thought I thought it would be driving
  573. that would go down, but it's not that they it's not driving. They might make up for inaccuracy, what they
  574. lose in distance as possible. It could be, but they all go down. And I thought that was interesting.
  575. The other two things I saw, I'm I'm I'm I'm filling up the box here, but I went to JSM, which is a joint
  576. statistical meetings. And in fact, I was online at the hotel for the meeting last week, but I
  577. didn't get to talk about any of the talks that happened. There were lots of sports talks.
  578. I didn't go to even the majority, but there were two that were fascinating, one by
  579. Richard Smith, who's a very famous time series analyst and a pretty dedicated runner.
  580. And he talked about handicapping the Boston Marathon, not handicapping the way they work.
  581. That is, at every age and sex, there's a cutoff. And the goal is to set the cutoff
  582. so that if you're below it, you're in. And if you're above it, you're not in. But it never works that way because they
  583. have a fixed number of positions. And every year, vastly large number of people outperform the predetermined cutoff and they have to
  584. reject people even if they even though they made the cutoff. Our colleague, Paul Saban, has been a victim
  585. of this several times in the Boston Marathon, where the cutoff is three hours and he
  586. runs it in two hours and 56 minutes. And they have this buffer in order to, and he doesn't make it.
  587. So it's fascinating how he talked about how you would try to figure that out. And the problem is, is that it changes
  588. behavior. Once you publish that number, people start running faster. I mean, Debbie, if you haven't seen the
  589. work by my friend, collaborator, advisor, George Wu, he has a paper with Devin Pope that
  590. shows marathon times as a function. Just marathon times, period. Just look at marathon times, like tens of
  591. thousands of marathon times. And there are all these little spikes at the round numbers.
  592. Right. And so people push themselves to hit those numbers, their goals. And then there's very few people that are
  593. falling just beyond the goal. There's very few people that fall above the goal. Just before the goal.
  594. You don't even need the statistical test. You just look at the data. It's like all these spikes on the round
  595. numbers. And so when they publish what the cutoffs are, everyone just beats it. And every year it gets lower and lower.
  596. They keep beating it. Because the question would be, what is going to be the threshold?
  597. They can't beat it anymore. But they haven't found it. Why do they publish a threshold?
  598. Why don't they just say, the line will be somewhere in this envelope of this range.
  599. Based on historic history, the line is somewhere in this range. And we just cut it off.
  600. Certainly, if they didn't publish a specific threshold, you might have less of this.
  601. People would still be beating for that range. But it might just incentivize a little bit
  602. of that weird behavior right around. I wonder whether they are forced to say the year before what the threshold was.
  603. Maybe that's just what they're just publishing last year's threshold. That would be fun to have Paul Seidman
  604. on and talk about how he's trying to get in the marathon. There's obviously sparks every time you jump up
  605. an age group. When you're in the low end of the age group, you're clearly advantaged.
  606. All right. Shane, what have you got? Well, Jordan Alvarez is still leading BAL in
  607. the Triple Crown, all three of the Triple Crown categories. Nobody's talking about it.
  608. Or at least we haven't talked about it. Why aren't we excited about that? I asked you, Adi, are you not excited
  609. about it? Because you still don't think- I'm excited about it. I guess I'm not excited about it.
  610. It's not like- Yeah, I'll start to get excited about it towards the end, right? Because it's- I was going to say,
  611. it's not like you're not excited about Jordan Alvarez. It's that you still think there's a good
  612. chance it doesn't happen. Yeah. Batting average is so fickle in the end. RBI is- There's only a couple of
  613. units ahead on both RBIs and home runs as well. Right, those are changing. So it certainly could change.
  614. And it's fascinating when it comes down to the wire, because that's when it gets really
  615. good. Jordan Alvarez is a great player, right? I mean, the real issue is what the
  616. hell is going on with batting average across the league? And it's just tanking. Here's a stat that I saw.
  617. You know, the Yankees, typically a team that scores only two runs or fewer, they're-
  618. typically, they win fewer than 10% of their games. The Yankees have won half. And have had a lot of games where
  619. they've scored two runs or fewer. They are not hitting. Their pitching is amazing, but their hitting is
  620. horrifying. And you've got to wonder- But again, Norman, I think that is a league-
  621. you've probably watched the Yankees more than most, but that is a league-wide phenomenon.
  622. Phenomenon. It's ridiculous. Pitching is great. Hitting is not so great. I don't know how long it'll be before
  623. the league does something. And I think the obvious- Well, they keep trying to do- I mean, they
  624. keep altering it little bit by bit here and there, you know, to kind of try and change the course.
  625. I mean, I think it's inherent that, like, you know, pitching is proactive and hitting is
  626. reactive. And so, so much of the kind of, like, technological and analytical tools that bring to
  627. bear disproportionately help pitching. They have so far. Although, if you've seen, you know, YouTube videos
  628. of this technology they use, where you see a video of a picture and a machine throwing the ball exactly as they throw it,
  629. it's remarkable what you could practice in the, in the, in the batting cage. It's, it's, it's just crazy.
  630. I mean, we had those clips on our show three, four years ago in the beginning, when we first heard about them.
  631. When we saw when it first came out, we had Travis Chalchuk last week, who was trying to talk about injury and prediction.
  632. But to me, the only answer is you just, the starters have to get in there more. There's got to be a budget of, of
  633. innings and they just can't be throwing. I mean, yeah, we need to make pitching a little worse by construction, by like not
  634. having like, you know, an inning each of like people throwing 110 miles an hour.
  635. Well, all you have to do is keep your roster larger. I mean, just, and this is a stat
  636. that I love to, to, to, to reiterate is, but back in the seventies, we talked about the 78.
  637. Teams carried eight pitchers. Now they carry 14, 13, 14. Maybe that's job creation, baby.
  638. All right. Let me jump in here with one last one that we can wrap up. I don't have anything that particular.
  639. I just want to drop a quick note on college football. Since we haven't talked about that, we'll jump
  640. into it in a couple of weeks now. But I think if I had to say, I had to try to bait you guys
  641. into the sport for a moment. I had to say what's most interesting. It's the, the number of teams who are
  642. seriously in contention for national championship ex-ante. And it's just, the sport just continues to
  643. broaden a little bit geographically across conferences. So for example, if a quick scan across
  644. the books would put, there's a tier of seven teams that are between 13 and 8%, 8 and 13%.
  645. So I could top 17, 17, seven with, with a pretty tight spread between eight and 13.
  646. So just running them down from top to bottom for you. Notre Dame, Texas, Ohio State, Oregon, Georgia, Indiana,
  647. Miami. Nobody would be surprised if any one of those teams won national championship.
  648. And that's just a different world than we used to be in. Yeah. I was going to say like five years
  649. ago, how many teams would kind of share that amount of percentage? Like one or two, three, typically like kind
  650. of. You know, the, the favorites were like 30, 35%, which is half of that stuff, I'd
  651. say. So, you know, two or three, by the time you get to the fourth, they're like,
  652. well, maybe somebody sneaks in there. Anyway, we'll talk in more detail the next
  653. couple of weeks, but it's a fun. Yeah, can I just ask one question though, just before you cut it off.
  654. Is that because we have less data because everybody's shuffling their teams with players or because
  655. there really are all equally valued? That's a great statistical question. And it is surely the case that the
  656. models are having a tougher time now because rosters are fluid. But I think the number one driver, Adi,
  657. is that the true differences in talent have been muted because it's, it's more spread out
  658. now. There was a while there where Clemson, Alabama and Georgia just stacked talent and, and, and
  659. just disproportionate to everybody else. And NIL and transfers have really spread it
  660. out. It's chaotic, but it is a force. It does spread out the, it does tend to, you know.
  661. That's right. That's right. All right, guys, why don't we wrap it up there for the team here, Adi Weiner
  662. and Shen Jensen, all the way to the end. Eric Bradlow, almost to the end, for our
  663. first half guest, Aaron Schatz for the team behind the scenes. Jacob Grondek on the board, our producer.
  664. And Dee Patel, boss lady, who made a cameo here in the top to say hello. We hadn't seen Dee in a little while,
  665. but she runs the show from behind the scenes, directing things. Many thanks to all those folks and to
  666. you for listening. Come back and join us next time. Between now and then, enjoy your sports.

Brief Summary

Aaron Schatz discusses NFL analytics, team projections, and player performances for the upcoming season.

Key points

  • First Football Guest of the Year. Excitement builds as the hosts welcome Aaron Schatz, the first football guest of the season.
  • Quarterback Projections Explained. Aaron discusses how he models quarterback performance, emphasizing team variables.
  • Seahawks vs. Rams Analysis. Aaron breaks down the Seahawks and Rams, highlighting the Seahawks' impressive last season.
  • NFC South Predictions. The NFC South is unpredictable this year, with all teams having potential to win.
  • Eagles' Super Bowl Chances. The Eagles are seen as division favorites but face questions about their offense.
  • The Evolution of Pitching. Discussion on how modern pitchers are striking out more batters but throwing fewer innings.

Episode highlights

Tension

45/ 100Some tension

Related episodes