How NFL Teams Get the Draft Wrong
Ask about Knowledge at Wharton
Episode Description
ABOUT THE EPISODE Why do NFL teams keep making the same draft mistakes? Richard Thaler, Nobel Laureate and Professor at the University of Chicago, and Benjamin Robinson, Founder and CEO of Grinding the Mocks, join the Wharton Moneyball team to break down how behavioral biases and flawed valuation models influence NFL draft strategy. From overvaluing certain picks to misreading data, they explore how decision-making errors persist and how emerging analytics are changing the way teams approach the draft.
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.
#nfldraft #nfl #sportsanalytics #behavioraleconomics ----- 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
- Welcome to Wharton Money Ball. Welcome to a full hour of sports analytics here on the Wharton Podcast Network.
- This is Cade Massey hosting this week with the whole crew. We have Shane Jensen in here.
- We have Audie Wynder in here. Eric Bradlow, who was not scheduled to be, but slid in here pleasantly at the last
- minute. So we're gonna be fully staffed for this show. We've got a double guest show this week.
- We usually run guests for half the show and open lines the other half, but we wanted to hit the NFL draft with two
- barrels this week, and we had the chance to get two great guests, and so we jumped on it.
- First up, Richard Thaler. Richard Thaler is a professor at the University of Chicago.
- He also is a Nobel Prize winner. Was it 2017, 2018? I should know that, Dick, and I don't.
- 2017 or 18? He- 17. So we've been doing this show for 12 years, guys, and we finally have a Nobel
- laureate on the show. This is the first. Dick, thank you for making time. We'd talk with you anytime.
- We're happy to have time with you. That doesn't mean it's the first time you invited me.
- Well, that's true. That's true, and I don't know why that is, honestly. We would take you on a lot of
- different topics, but we decided to grab you on the NFL draft. We've just come through the draft.
- We're gonna have two guests. We can kind of debrief the draft, but we thought we'd chat with you first, Dick,
- and I'm gonna take the perspective, I think the easiest way to think about it, I'm
- gonna take the perspective of your wife, the honorable professor, former professor, Franz LeClaire, who has
- not really loved the project we've been working on all these years. She doesn't really understand why we waste our
- time with the NFL, and so I wanna ask you that question, Dick. Why is a Nobel laureate messing around studying
- the NFL draft? I mean, why is that worthy of your time? Yeah, well, as you know, I have one
- criterion for anything I choose to do, which is it has to be fun, and so studying sports is fun.
- Now, of course, that is not a satisfactory answer to Madame LeClaire, but my serious answer
- to that question is that we don't have that many opportunities to study business decision-makers
- making decisions, so there's a big merger going on, and there's competing bidders, and we don't
- know what they're thinking and how they came up with the numbers, how the process evolved.
- In sports, we get to watch, so some team continues to punt on fourth down, fourth
- and one everywhere on the field. We get to see that, and most of my life, we would just say, what are
- they doing? But many years ago, you and I got the idea that the NFL draft would be
- an interesting institution to study because it involves so many different things, so I mean, the
- institution is interesting. Well, why do they have a draft at all? You could ask, like in European soccer, there's
- nothing like a draft, but they have it, and there's this odd chart, and our history of this went back to learning about
- this chart and actually calling up the guy who made it, and that's an interesting story.
- The guy was a partner of Jerry Jones at the Dallas Cowboys, an engineer, and the
- coach at the time, Jimmy Johnson, asked him a question. He said, look, we get offers to trade
- the fourth pick for some other picks, and we have no idea what the right price should be.
- Can you figure it out? And this guy's an engineer, so he's a numbers guy, and he got all the trades
- that had happened up to that point, and my impression, maybe your recollection of this conversation
- is better than mine, but my recollection is he said, kind of wrote them all down,
- and then kind of drew a line on a piece of graph paper, and he made the first pick worth 3,000, and the
- last pick worth one, and there's a smooth curve you can draw through that, and so
- for a while, that was proprietary with the Cowboys, but then assistant coaches move around.
- Now, Google NFL draft chart, and it will pop up. It hasn't changed in 30 years, and teams
- use that as like a price list. I think of it as like the blue book for used cars.
- You can look up what your 10-year -old Honda Civic should sell for, and so the paper we wrote a long time ago
- that my wife mocks us about asks the question of is that price list the right one?
- So you can trade the first pick for the eighth and ninth picks, according to that
- chart, or for half a dozen second-round picks, and that was the origin of our research paper, and what we found is that
- price list is very wrong. It's much too steep at the beginning that you'd much rather have the eighth and ninth
- picks than the first pick, and we've now had a re-dive into the idea of this asking, since we published a paper about
- this 15 years ago, has anything changed? And the short answer is no, so it's been nice talking to you, Cade.
- Well, let's take it one step further, and then Adi wants to jump in with a question, but I would say what did you
- learn? What do you think we have learned? We came back to it. God knows, I hope we learned something from
- coming back to this. It was bad enough that we spent so many years working on it the first time,
- and now we're gonna do it again. What have we learned this second time around, Dick?
- In what way do you think about this differently because of what we've done the last
- couple years? Well, so I think there were two big questions we started with this time.
- One is, do they still use this Jimmy Johnson chart? And the second is, have they gotten any
- better at picking players? Now, the reason why that chart is wrong is that, you know, I said the first
- pick is worth the same as eighth and ninth. That suggests the first player has to be
- twice as good, excuse me, or twice as valuable, or more precisely, how much he contributes
- minus what you have to pay him. The surplus has to be twice as much for the first one as to any pair
- you could trade for. And for that to be true, you'd have to be really good at predicting.
- Mm-hmm. And our favorite stat from the original paper was one we called the better than the
- next guy stat, which is take any two players who play the same position who are drafted one after the other.
- So the third running back compared to the fourth and so forth. And we say, what's the probability that the
- earlier player is better than the next one? And in the original paper, the answer to
- that question was 52%. Now, if they're perfect, it's 100. If they're random, it's 50.
- It's a little better. And it's a little higher in the first round. And so that was the first question is,
- well, now we've got stats and the combine and AI and who knows, maybe they've gotten
- a lot better at this. And the answer is they haven't. That number is now 53%, first round 58%.
- So they've not all of a sudden mastered the art of predicting who's gonna be good.
- And the second one is, do they still use that chart? And the answer is yes, kind of ridiculously
- so. So there was one trade early in the draft, the Browns traded down from six to
- nine or 10. And on the chart, that was worth 1 ,600 points. And they paid 1,602.
- And so- Okay, those prices are like floors. So they're the asking price. If the buyer's really keen, they'll sometimes pay
- more than that. But we don't see too many sales. Well, and that finding, I think was-
- I can't hear you. Let me know if you can't hear this, but the observation is that we thought they
- might learn away from that curve. We thought the prices might change. I mean, they've had more than a decade
- of learning. And in fact, it went the other way. The curve is the price. And it's just unbelievable that that curve was
- created based on 1980s trades. It's been in existence for 35 years now, and it hasn't changed.
- Free agency has come, salary cap has come, and the curve hasn't changed. In fact, if you look at the coherence
- to the curve, it's gotten tighter over time. So it's one of the most surprising things.
- Let's take a couple of questions. Adi was first and then Shane. Yeah, so I've actually had the opportunity to
- dig into this a bit and even published a paper on it, which I think Cade is certainly aware of.
- And I was really surprised when he told me, Cade told me that it hadn't changed at all.
- Because we know a lot of people in sports and they're pretty smart. I mean, not everyone is smart.
- I mean, I'm gonna be perfectly honest. But there's a lot of smart people, including
- some of our former students and very close friends now are pretty high up in lots
- of places. And when I explained what you guys were finding that it hasn't changed, the remark that
- came back to me was, well, we don't maximize expected value with our trades, which is
- where your wrongness comes in. The utility is expected value. We actually look for tail distribution probabilities.
- And so I went back and I looked at that and I found that the trade curve actually does match a utility function that
- weights with much, much higher probability superstars. So they've cut off, it's hard to measure
- superstars. It's a second contract. So, and taking a lead from Cade, what we did was with size of the payroll,
- essentially the cap that you're devoting to one player. So if you call superstar something like 15
- % of the cap, then, and you wanna make a trade that balances the expected number
- of superstars, then it actually does match. And that seemed to be a very highly,
- at least it puts a gloss on it. I mean, for many trades are really dumb because they aren't trading.
- They think they're doing that, they're not. They really should be trading expected value.
- But for some teams, I think the smarter ones, when they do trade up for a higher pick, they have that in mind.
- So we've done the following analysis to look at that. We've looked at every two for one trade
- that has happened. And then we've also looked at every feasible two for one trade.
- So I mentioned you could trade the first pick for eight and nine. You could trade it for seven and 12
- and six and so forth. And what we find is when you trade down, you get the same number, you get
- twice as many games started from the two players and the same number of pro ball
- appearances. So you don't, if we're settling for all -star appearances, you don't do better.
- You break even on that and you get - That's not enough. All-star period, that doesn't quite do it.
- You need superstar status, not just all-star. Yeah, well, if you heard of, there's a
- guy, kind of a tall guy, Tom Brady, I think 199. I rest my case. Hold on, I'm gonna- Not an anecdote
- on the one extreme? For real? Adi, what you need- Well, maybe I can jump in with- No, no, let
- me jump in. Shane, Shane, let me jump in first on this point. And we should continue the conversation in other
- places, Adi, but what you need for that curve to be correct is that the team places zero value on any level of performance
- other than extreme superstar. That's right. You need zero value on anything other than
- superstar. And there's no evidence from any labor market in professional sports that suggests that that's the
- value they place on performance less than superstar. I don't doubt, Dick and I said in
- presentations in the 2010s, in the 2000s, we said, look at this, the Pro Bowl curve
- comes pretty close to the Johnson curve. The Pro Bowl curve, I think you're right.
- You said this thing a minute ago, you said, this is what they have in mind when they're making the trade.
- Yes, it's something they have in mind. It's the psychology of the trade up, but it is far from a rational psychology.
- They care about getting starts, even out of first round picks. They pay a great deal for mere starters
- in every free agent market. The psychology might be a heavyweight on superstars,
- but there's no way that that rationalizes the trade curve, given anything else that we observe.
- Two quick comments. First, you don't have to put zero. You can have an S-shaped curve and
- you'll get the same thing. Secondly. I'm gonna have to understand that better. I don't accept that as a rationalization.
- You can have a logistic function that approximates it. Oh my God, this is gonna be the
- most extreme indicator function. Are you kidding me, Artie? Come on. It doesn't go that quickly.
- The second thing is, and I think this is sort of relevant, when you talk about...
- Well, you know what? I'm just gonna pass. All right, so I'll just say the following.
- In the formal analysis of the paper, which I think is gonna be out anytime this decade, the way we value players is by
- what percentage of the salary cap they get in their first free agent contract. Yes.
- That's what I do. That'll include superstars. And we'll have this conversation again when the
- paper is done, and otherwise we're gonna just go into the weeds. Shane. Yeah, and I hope I'm not just asking
- the same kind of question from a different direction, but earlier when you were talking about
- what they could have learned and how you kind of expect things to be updated, you
- mostly talked about that you would have gotten at least over time a more accurate evaluation
- of the actual talent. But that's only, I think... I just wanna clarify that that's only kind
- of half of the story in terms of like whether a one is worth a seven and an eight or whatever.
- It's not just that accurate, like the accurate evaluation of their various talent.
- It's actually also a big part of the story is how the talent distribution, or I guess how the value distribution falls off away
- from one, right? I mean, if you had sort of, I guess, an extreme enough distribution of actual value,
- the number one pick could be worth seven or eight. It's just in actual football, it doesn't seem
- to drop off like that. I guess it's maybe the empirical part of this. You know, if you look at the first
- round picks over the last couple of decades, there are a lot of busts. Just look at the quarterbacks drafted by the
- Chicago Bears until the recent one. And they're, you know, it's not the case that they hit very often.
- And, you know, as Mr. Brady shows, or for God's sakes, Mr. Irrelevant, Brock Purdy, it's
- not the case that the best players are all in the top five. They're not. It is also the case that there are
- just more busts at the top than we remember. We think of those guys, because they are
- the highest likelihood of superstar, we think that the likelihood of superstar is higher than it
- actually is. Eric was gonna jump in with a question. Yeah, so Professor Thiel, I wanted to ask
- you, so you also studied learning, right? And so I would think, I'm a Bayesian, but that doesn't mean just in Bayesian computation.
- I do think about how people update beliefs, and I do a lot of work on that. So if the chart is mispriced in some
- way, why aren't teams learning this over this long period of time, especially given, in my
- view, and your guys' empirical study, the overwhelming evidence that they do? Yeah, so I think most, so I think
- almost every team has somebody on it who has read our paper most have somebody who's
- been able to update it. There are many other updates around the professional sports world.
- So, and I think a lot of teams have another chart they think reflects actual value.
- But my impression from talking to people in the league is that the Jimmy Johnson chart
- is the list price. And if you're selling, that's your asking price. And you'll get a lot of nos.
- You'll hear from teams, yeah, we wanted to trade out, God knows what the Rams were
- thinking, but you'll hear, oh, they wanted to trade out, okay, then what price were they
- willing to take? And we don't see very many trades below the chart. So smart teams would have to be willing
- to sell at a discount. Right. So, and there's an old story, well, remember, right when the first version of our paper
- came out, the 49ers had the first pick and they needed a quarterback. And for months, they were debating between two
- hot prospects and they couldn't make up their mind. And a reporter, a local reporter called me
- and said, what advice would you give to the Niners? And I said, I'd announce a sale on
- the first pick, 20% off. And that's not what they did. And they used the first pick to draft
- Alex Smith. And there's another guy who was playing on the other side of the bay, who had
- a pretty good career and he went 24th. So, no one seems to, my advice is so worthless.
- No team has yet adopted the strategy of not knowing what to do with the first pick and announcing a sale, because it makes
- them look bad. Well, this is the thing that I want to emphasize because it's a very fair question.
- It seems to be an off equilibrium place we're at. The price that is so rigorously adhered to
- doesn't seem to match the value of the assets that come from those picks. And so the question is, why does it
- persist? Every team in the league has a chart and this is what happens guys. The trades come up and they've got Jimmy
- Johnson chart and they've got their own charts. And they compare it on all these charts
- and they know that the price is basically gonna be Jimmy Johnson. And then they decide how badly they want
- it by looking at their own charts. And so I asked a guy just last week, we were texting just before the draft.
- This is a guy who has run the trades for multiple teams. And I said, why would you say some
- teams are not willing to accept less than Jimmy Johnson, basically fail or sale, even if
- their other charts really like the trade at a lower price? So this is key. Why does Jimmy Johnson pricing persist even when
- lots of teams have sharper charts that diverge from Jimmy? Yeah, so this is the guys I got
- permission, I'm not gonna tell you who it is or use any team names, but he gave me permission to use his text.
- Why is he selling down? He says, this is the answer. He gives me the answer. He says, perception.
- This is quote from our text last week to my question. He says, perception, how another team was looked
- at, for example, when they got way too little in that general manager's first trade.
- You never wanna look like the sucker. And the safest way not to is to not deviate from what has gotten that deal
- done before, which is Jimmy Johnson. And that's the story. Taylor talked about the interview with the guy
- who created the chart, Mike McCoy. He told us that in 2006. He says, this way, no one gets skinned.
- He says, it's CYA, CYA, because these guys, they don't wanna look bad. They don't wanna look bad in the press.
- They don't wanna look bad to their peers. And that's enough to keep them. Now, one last wrinkle's needed.
- The guys who wanna buy at lower prices can't access it. They can't short. There's no shorting in this market.
- So the only guys that could move it are the guys that would sell at lower prices.
- But there are these basically punitive measures if you sell below Jimmy Johnson.
- It's hard to buy that they won't do that, that you're essentially saying that they're letting
- something that's been known for 20 years get in the way of, I mean, listen, look
- at fourth down decision-making. It took forever, but now they all do it. It's done, right?
- It took forever. They don't all do it. So they get it like half right. We didn't know how this was gonna come
- out. What I'm telling you is it doesn't look like they've learned anything. Well, let me ask you another question.
- How do you know how to build a Super Bowl-winning football team? What's the secret sauce?
- Maybe there's something, I mean, why would we in the outside know how to do that?
- Because implicitly you're saying that you're not doing it right because we know how to build
- the team and you don't. How do we know that? No, I think, look, no one would hire
- anybody that's on this call to do such a job, but look, right? I'm sitting in the office of a money
- management firm that has my name on it. And we don't think we know how to run companies, but we think that we can
- put together a portfolio of stocks that outperforms its benchmark. And that's by being right 53, 54%
- of the time, right? So I think teams, the smarter teams are doing better, but they basically are, that's because
- they trade down more often and never trade up. But they don't wanna look bad and they
- run a big risk of getting fired. Look, 10 of the 32 coaches got fired last year.
- I don't know the data on GMs, but you can't afford to look dumb. I actually look forward to hearing about that
- study that shows that the teams that don't ever trade up and don't trade or trade
- down more often, they do well. I don't know if it's false or not, but is that true?
- It's something we looked at. We used to call this the golden, what the Holy Grail.
- The Holy Grail, yeah, right. There's so much else. Some teams are better at fourth downs.
- Do they do better? The Steelers were terrible at fourth downs and somehow still managed to win.
- So there's blocking and tackling and the stuff we geeks care about. Well, Adi, the last thing we do in
- our first paper is to run an analysis of the accumulation of draft capital in one
- year and the downstream consequences on the playing field. So it's wins. What we care about is wins, right?
- And what we find is the accumulation of draft capital doing the trade down, managing their
- draft capital in the way we would recommend is positively correlated with downstream wins.
- Now, is it direct or does it just reflect the kind of sophisticated management? We don't know, but we find a reliable
- relationship between exactly what you just said. It's the last test in the published paper.
- Trading down, accumulating draft capital is positively related with downstream wins.
- Eric has a question. Yeah, I want to ask Professor Thaler another question. Here's another way to rationalize, and I wonder
- if you guys have looked at this. Suppose, since as you know, utility functions also
- sometimes have uncertainty bought into them. In other words, we take the mean utility,
- we might add an uncertainty term. Is it also possible that the whole mechanism is due to the top-end picks are
- seen as much more certain with lower standard error? Later picks is seen as much more variable.
- Therefore, a rational person would like, even it seems overvalued if you're maximizing expected utility, but
- if it's a risk-adjusted expected utility, it's not. Because that could also rationalize it.
- People overestimate the certainty with which the top picks. So it's a variance story, not just a
- mean. Well, trading down even past the first round works like two-thirds of the time.
- And it's just not the case. After the first pick, it's better than the next guy's 52%.
- And you can't predict the tails. So I can't think of any rational basis by which you wouldn't prefer twice as many
- starts and the same number of pro bowls. And to be specific about the variance, the
- variance actually goes down later in the draft. It doesn't go up, especially if you're looking
- at surplus. Surplus, which is after you've paid them, you've got this very high volatile left tail at
- the top of the draft. And later in the draft, yes, you've got outliers, but mostly you've got a bunch of
- low performances and standard deviations are coming down. Last comment. The big variance comes from Tom Brady and
- Brock Purdy. Guys, we're going to need to cut Professor Thaler loose. Any last questions for him?
- What questions remain for you, Dick? Give us one last question you have about the draft.
- Now that you've looked at it, come back to it, looked at it some more. You've got decades now of looking at this.
- What questions remain for you about it? Well, I think the, within the team dynamics and the role of the owners, I
- think would be really interesting to study. And let's just say they don't always help.
- Okay, why don't we wrap it there? Dick, thanks so much for making time for us. Appreciate it.
- Good luck with everything you have rolling out there. Just off the line with Professor Richard Thaler
- dialing in from California. He spends most of his time, half his time out there talking NFL draft.
- We are gonna spend the next half hour with a one-on-one fight between me and Adi about our papers.
- We'll suspend Ben for next week. We have to work some things out, okay? Is that really happening?
- No. Alternatively, we'll fight- Shane and I could moderate that discussion if you'd like.
- Okay, I'm ready. I am ready, but we're gonna- I've been to an econ seminar, and those are
- pretty bloody, so. I've never been to one, and I've started to wade into that territory and realizing it.
- Yeah, Adi decided to turn the first half of the show into an econ seminar, so we'll take it up.
- We'll take it up elsewhere. We've got instead Ben Robinson. We are an all-in-one show.
- All-NFL draft show this week. We're gonna return to previously scheduled topics like
- baseball. Adi has some primary research, some new primary research he wants to share.
- We'll talk about it. But in the meantime, we still are processing what happened over the weekend.
- NFL draft had its thang. Ben Robinson, for those of you who don't know, created and runs Grinding the Mox.
- Ben has been on our show many times over the years. He was, and I think maybe it's safe
- to say, the first aggregator of Mox out there. He was an early, at least an early
- aggregator of Mox. And he has, I would say he's grinded over the years and built a little bit
- of an empire, an influencer. He's got a partnership now with the NFL. He consults to teams.
- More and more, the conversation around the draft is around things like this, like the consensus
- and big boards and how teams perform versus those measures. And so in many, in a very real
- way, Ben's work has become central to the conversation around the NFL draft. It's been fun to see.
- We felt like we caught him pretty early. He's making, I think he's really making the
- conversation smarter out there. Even the analytics of it, like how do you aggregate these things is become super interesting.
- And so the methodologies are interesting. So Ben, welcome back. Thanks for making time for us.
- Glad to have you. Oh, it's an honor to be here. And in terms of being an early Mox
- draft aggregator, I was not the first. I initially thought I was the first. Okay.
- But would you guess Brian Burke was the first? Of course he was. Of course he was.
- Tough to be first on anything ahead of Brian Burke, turns out. But you've certainly invested more over the years.
- Give us a sense. Give us an update of state of the union of Grinding the Mox through the 2026
- draft. What's important? What are the high points? How should we think about your enterprise at
- this stage? Well, yeah. Like I said, it's an honor to be on. So yeah, no, Grinding the Mox started off
- as kind of a humble side project and now we're a data and insights company. And so, you know, this year I have
- two team clients that I've worked with and I, this year added five more. So we're working with about a quarter of
- the league is using our data in some shape or form to inform their draft preparation
- and draft planning and draft strategy as well. All the internal tools that teams have at
- their disposal to think about the draft and how they want to maneuver and plan to
- get either the players they want or the kind of positions they wanted to target in
- the draft. But overall, we're trying to, you know, do what we've always done, which is, you know,
- using mock drafts, the wisdom of the crowds and machine learning to predict the draft.
- And so we have a really consistent hit rate of getting about 80 to 85% of the top 100 players matched.
- This year was a little different, but then in the first round, we're usually getting between
- 28 and 29 of the 32 in our top 32. And so that type of consistency is really hard to repeat by most experts, because as
- we know, in the wisdom of crowds, these things go up and down and up and down. So the data can help a lot in
- setting those expectations accurately. And like you said, making people smarter consumers
- and fans of the draft. So two clarifying questions, just to stay with this for a moment, more background.
- You said you have seven NFL class. Now, how do they use your data? So there's a lot of firewalls that exist
- between me and the teams that I work with. So I don't necessarily get to go into
- the draft rooms and hear, this is what we did, this is how we used your stuff. My general sense of just the state of
- the art and knowing what I know about the industry is that lately teams have been
- developing internal simulators. So the types of simulators you see online, the teams are developing that internally with a
- combination of public data, their internal grades, maybe you might pay for some betting data too
- and throw that in there. And the decision-makers are using that in the lead up to the draft.
- You know, when we talked a lot in the past about fourth down decision-making, we
- were often saying, you know who's really good at making fourth down decisions? People who get a lot of experience with
- that. In game, it just doesn't happen very often. And so we were saying, hey, they should
- play more Madden. That is like more realistic. You'll get more of these scenarios where you'll
- get to say, okay, this is the risk, this is the reward. And it's the same for the draft.
- Having these like internal tools, these simulators that can give more GMs more bites at the
- apple. You only get a certain number of picks every year, a very small number. And so being able to thread the needle,
- think about your strategy. If this thing happens, how does my strategy change? Having more bites at the apple is a
- big piece of internal draft planning. And so my data feeds into some of those products.
- I think some teams are also real time in the draft using it to better understand
- what's likely to happen in the next few picks. So for example, if a team gets a call and is offered a trade and they
- have to ask, if I trade back, what's the chance my guy is still gonna be there? And they need a good sim to answer
- that question. And of course it has to be a sim that's updated continually as people make picks.
- And so I can imagine your data being super helpful for that process. Last clarifying question before we go to Eric,
- how have you gotten better over time? I'm guessing that these are more accurate than
- they used to be. What is the frontier for you as you improve your model over time?
- I think the main thing is about data innovation. Methodologically speaking, teams are largely using models that
- have been around for a while. You'll hear about like the, my model is the model stuff that I sell to teams
- is like a Bayesian hierarchical model. It's not actually super complicated. Those types of models have been around for
- a while. You'll hear about like other types of methodologies as well. But for me, it's a lot about data.
- It's less about necessarily, hey, let's try to like methodologically fix this issue.
- Because there's a lot of uncertainty in the draft. And so to me, finding the ways to
- inject new data sources is what's gonna make your model, I think, better. The better data you'll usually weigh out over
- the better method in this case. There is like some amount of evidence to be made that if you wanna get a
- very accurate simulation in some ways projection, though a lot of the mock drafts are kind
- of copying each other in very like mimicky ways. And so you still, but you still need
- to be able to find those needles in the haystack that matter. And so you need to have access to
- as much of the data as you possibly can so that you can find the people that are giving you the appropriate signal that
- have the track record of being accurate. And then don't just mimic what everyone else
- is saying. It actually has something that brings value to the conversation. So I think you still need data and
- that's the thing. So I've integrated big boards as well, just like straight up player rankings into my data
- as well over the years. And that's been a meaningful addition to help provide priors.
- But to me, it's more about just the data you can bring. Like I mentioned before, the mock drafts, the
- big boards, some type of, whether it's like gambling odds or maybe even prediction market odds,
- depending on those, that can add a lot to a value and not necessarily the modeling
- approach. That can be fairly straightforward. Eric. I don't know if this has ever happened
- on our 12 years of Moneyball where the guest has literally answered every single point of
- my question. I was gonna ask him, would you rather have better data or better models?
- I was gonna ask him how he deals with the correlated nature of the drafts. Like I just copy what Cade Massey puts
- up there and now you got seemingly two observations that are convergent, but I've actually added
- no information. But I will ask one question then. But first of all, thank you for the
- clarification. In a world where everyone seems to be using neural nets and more complicated nonlinear types
- of models today, despite I'm a Bayesian and I'm Mr. Bayesian hierarchical model, I love fitting
- them. Why not jam a massive neural net against your data where you can find all possible
- combinations of things that help you predict outcomes or like why in some sense be more
- statistician and probabilistic-like and why not jam the old black box computer science model at
- it? I think a lot of it has to do with the uncertainty in the data that you're having.
- So the mock draft world of data is a censored world of data. Most mock drafts are only 32 picks long.
- The first round is what gets the most eyes, the most attention. And so you're dealing with data that just
- has a bunch of built-in issues I think that statistical models do a better job of addressing.
- I think you can throw the machine learning methodology at it too and you probably would
- get pretty decent results. But I do think that the uncertainty, like being able to account for the uncertainty and
- actually understanding the generation, like how your data is generated, what it represents, you wanna try
- to match that to the problem. But then I think the other thing too is just- You score yourself, Ben, beyond
- just I'll call it the pure hit rate. Like in other words, being off by one is very different than being off by 15.
- And in some sense, you could imagine a black box model always in general underestimates uncertainty.
- And so you wanna- And you feel better about that in that way. Yeah. So I mean, you could come up with
- a model that could be 95% right, but then the confidence intervals are huge and it just has no meaningful interpretation.
- But to me, there's just a lot of uncertainty with the draft. And so you just kind of have to
- be, get used to being wrong a lot. And so, and then there's also these correlated
- errors you mentioned. So in the first round of the draft this year, my mean absolute error, just like
- very basic metric of prediction accuracy was very average. The first round, there might've been a couple
- big picks, but for the most part, the players went right around where we thought they
- would go. Especially if we value higher picks more than later picks. But then if you move into the top
- 100 and that's the first three rounds, this year's draft, in terms of like my average
- mean absolute error, just without adjusting for any pick value, it was seven picks higher on
- average. So that means in the top 100 this year, there was just a lot more uncertainty.
- And that's just with my static pre-draft model. I don't have like a live model that
- I use. The teams obviously have these tools to be able to make those assessments in draft.
- But yeah, there was a lot more uncertainty. And I think some of that was because
- of these correlated errors that you could pick up. So one of the things that was the
- story of round two and three of the draft this year was the rise of the, what I would call the overpriced blocking tight
- end. The teams, a lot of NFL teams are following the money around what is innovation in
- the NFL. And the Rams ran a lot of three tight end sets where you need a blocking
- tight end to have the threat of the run game and teams want to be able to run those.
- And so they were like, well, if that's the new scheme innovation, we want to be able to get that.
- It bumped up the price in the draft of the blocking tight ends. Seth Walter from ESPN asked a bunch of
- team sources about a bunch of players who went earlier than expected and asked them, is
- that really your reach? And for the blocking tight ends, he said, actually, I think they were just a lot
- very undervalued by the draft community this year and overvalued or maybe more valued correctly, right?
- By the teams. So these, you could pick these up in a model, but if you just jammed everything
- into a machine learning model, would miss that all the time. Real quick point about that dynamic.
- This is a classic case of non-stationarity. The world is changing. So we don't know who's got the valuation
- correct, but we know if everyone's playing offense or a lot of teams are playing offense
- differently now, however we valued positions in the past will be less relevant. And so it's an open question on where
- the tight end value is going, but it's just this perennial challenge of modeling when the
- world is moving. Adi. So actually we did a little impromptu seminar on draft valuations, but we weren't thinking at
- all about kind of what you're thinking about. Just to, so I want to clarify, you're
- trying to predict where the players will actually go in the actual draft. Yes. As opposed to figuring out, well, who's the
- good player that's going to go lower or higher? And you're trying to model future value, which
- is what we were talking in the first half hour of our segment. So my, so the question is, is that
- something that teams want to know from you? Or, I mean, are they just interested in
- knowing the person they want, where is it likely to go and how do they have to trade or position their order so that
- they make sure they get their people? So to answer your first question, so I'm, yeah, I'm trying to predict where players will
- get drafted. So my friend Arif Hasan creates every year since like 2015, what he calls a consensus
- big board, where he gets hundreds of player rankings and aggregates them together.
- Recently, he did a little analysis on his sub stack, which is called Wide Left. And he analyzed how good different publicly available
- metrics were at predicting where players are going to get drafted. And also how good some of these metrics
- were at predicting future value based on the pro football reference, approximate value stat, which is
- probably one of the best stats we have, you know, for looking at value. And my data predicts the draft the best.
- I'm very proud of that. I think I do a nice job. It's good to hear. My data predicts actual NFL success of those
- publicly available metrics, the second best. The thing that actually predicted NFL valued the
- best was actually his board of player rankings. But by me being closer to the actual
- draft in the top 100, you end up getting a lot more of those right than wrong. I still think he needs to do like
- an additional analysis where he compares it to the actual draft. But by being closer to where these top
- 100 players should be selected, we do a better job of predicting the draft than even
- some of like the other publicly available aggregated boards out there. So sometimes you do want to try to
- be as close to the market as possible. And I know that Timo Rieske from, I don't know if he's still at pro football
- focus or not, but had done this analysis in the past about, you know, do you wanna be closer or further away from the
- consensus? Where does it matter? When does it matter? And so it does have some impact, whether
- they do the teams wanna use my data to predict where players are gonna be drafted
- or where, yes. Did they wanna use it to predict how good players are gonna be? Player evaluation, no.
- And so I think the vanguard of player evaluation is gonna be trying to beat the draft, not trying to approximate the draft.
- But I do think it's something you should consider. It might make your model a little smarter
- pre-draft. But yeah, I think teams are doing a lot more different types of things on player
- evaluation on the data side. And I'm still not a hundred percent sure that it's making its way into the decision
- makers. Like, I don't know how much it's infiltrating decision-making. There's a lot of alignment that has to
- happen organizationally for insights to get from the draft, from the analyst room into the draft
- room. I remember reading Future Value, which is a great book about, you know, prospecting in baseball.
- And there were teams that were drafting entirely from models. I don't think we're anywhere near that in
- the NFL yet. Scouts still dominate the conversation. And I think we've also learned in other
- sports that scouts have really important data that they bring to the table about players.
- But yeah, teams are using this to try to kind of figure out how the draft might play so that they can find those
- specific players that they want. And so what I find is that the teams that are reaching the most in my
- data for players are usually just, they just don't consider that. They just have thought before the draft.
- And we might use the analytics department to make sure we're in a position to get
- the guy we want in that area that we want. But we're not using them to figure out
- which players to pick. We're super scouts. We're the smartest guys in the room. We're gonna pick the guys we want based
- on the coaches we have to maximize the opportunity to still have our jobs next year.
- Can I ask a quick follow-up question? Just like a real, because if you imagine, if you look back at the draft years
- later and you rank the players by quality and then you rank them by their draft,
- how often does the top five players in the draft by quality get drafted in the top five?
- Is that? Yeah, just by the way, as an extension to the question I was gonna ask you,
- Ben, by the way, was a related one to Adi's, which is how much of the variation does, if someone just did pure, forget
- what position I need, forget any of that. I just rank according to perception of quality
- and I correlate that with grinding the mops. Adi wanted the top five. I wanted, in general, for the top hundred.
- How much variation is sucked up just by the pure quality rankings versus, let's call it,
- the context of I need a tight end, I need a running back, et cetera? So we have a student who did this
- for the NBA. I know the number. I'm curious to know what the number is for the NFL.
- This was exactly, I have another question for Ben, but this is my question. So the number for the NBA is one
- and a half. So of the top five players in a draft class in the NBA, one and a half are drafted in the top five.
- Interesting. I wish you hadn't given that number yet. I was gonna ask everybody to guess.
- Yeah. Well, I would have guessed the answer to my question was about a quarter to a
- third of the variation in where people are drafted, it would have been based on pure
- quality. Like if all Ben did was use a consensus not of where people are gonna be
- drafted, but of the rank order of quality, that that might suck up 25% to 30% at most of where people are
- actually drafted. But Ben, maybe you know the answer to this. I don't really know the answer.
- I think it's my data is very correlated to the draft. So for the most part, my data might
- be saying, hey, like this guy is expected to go here and he goes there. There's like differentials between where players were expected
- to go and where they go. That's kind of what we know is that what Timo Riske has brought into this conversation,
- which is that draft reaches are real and steals don't necessarily exist. My bet is that it would probably be
- more than one and a half. So there's the classic, this is the classic draft curve literature and actually our friend, Ron
- Yurko at Carnegie Mellon, they did a little bit of an update on the draft curve.
- How do you, do you optimize your draft curve for just value? Do you optimize it for superstar?
- And they were doing it with Madden rating data. And so that was a new analysis that
- they released in recently that probably would do a better job of telling you those sorts
- of things. I don't necessarily think we have the public data to tell us necessarily who is that
- or not. I think the Madden data might be some of the best data we actually have publicly
- to do that across positions. But I would like to think, I think it would be higher than one and a
- half, but I don't know if I would, I would say probably maybe two. I don't know if I would say anything
- more than that. I can answer the question this evening if we just consider the second contract value, which
- is I think a reasonable way to go at it. And I'd be curious to get everybody's number
- out of five on average, how many of the top five contracts, once they reach free
- agency, how many of those top five in a draft class were drafted in the top five?
- I'm going under the one and a half number that Adi just said. Under one and a half.
- That's strong. That is strong. It's an average. So you can pick a non-integer. 2.5. 1.2. Yeah.
- Under. I'm going under also. Ben, let me ask you. Not around 1.2. I like your number,
- Eric. You anchored me. God damn it. Let me ask you a question. My feeling is that usually in a draft
- class, you usually have a couple of quarterbacks that go really high. That's the thing.
- The quarterbacks are gonna get you. This year's class might be biasing you to think that this is not an average quarterback
- class, but I do think that, yeah, like in a more normal quarterback class, there will
- be two quarterbacks that go in the top five and then they have a decent chance of being in those numbers or an edge
- rusher or a tackle or, you know. So like when you get to drafts and you see like are these non-premium positions
- going in the top five a lot? Sometimes. This year it did. And so, but yeah, overall, I think it's
- more than one and a half, but yeah, maybe I'm just over-indexing on my mental model.
- I'm gonna go under. I'm gonna go under as well. What is your thought? I mean, I can't even tell you the
- dismay of my son, one of my three sons who does a lot of work in sports analytics, trained by Adi.
- When Jeremiah Love was drafted third, he was apoplectic. He just could not believe it.
- So did you have him going that high in the draft? Forget whether he's the right, it's the right
- choice. No, I understand that's not what you look at. Did you have Jeremiah Love going like a
- third to the Cardinals and was that what the consensus said? So in the grindingthemocs.com public expected draft
- position board, he was my fourth ranked player. In the Bayesian model, he actually was ranked
- third. So, but the idea, I think the vast, that was kind of his ceiling was gonna
- be pick three. And he went there. It definitely changed the complexion of the top
- five of the draft. I think it did a lot of other teams some really strong favors because it definitely
- shook up. I think that if Jeremiah Love hadn't been drafted third overall, he would have gone fourth
- overall. I think that would have been a big disservice to Cam Ward who the Titans drafted
- first overall last year to draft a running back that doesn't really help him at all
- or change the offense in a really meaningful way for him and like the evaluation of
- him as a quarterback. But yeah, I had him going that high. It's very regularly known that we'll have these
- conversations around positional value of players especially at the top of the player rankings.
- They just tend to go for the most part at the top of the player rankings. The one position that we see year in
- year out where if a player is ranked high, and we saw that this year with Ohio State safety, Caleb Downs, safeties get depressed.
- Their value, they always are going a good amount later than we thought in the draft.
- So teams are willing to draft running backs really high. They're willing to draft linebackers really high.
- Safeties for whatever reason, they get pushed down relative to expectation. So Caleb Downs was one of my three
- biggest fallers in the draft. He was my seventh ranked player and he ended up going 11th.
- And so that's sort of the one thing that is the exception to the rule of highly ranked players go high in the draft
- as safeties. And I don't have a great answer for it, but we saw this year the Seahawks
- win the Super Bowl, not drafting a safety in the first round, although I did have
- Nick M. Unwary from South Carolina in last year's draft as a late first round value in my
- numbers. But that safety position, if you can find that perfect blend of someone who can play
- the run in the box and then cover tight ends and receivers up the seam, that's become a lot more valuable.
- But for some reason we're spending high draft picks on blocking tight ends and not as
- much on safeties that can make a difference in the defense. Ben, you mentioned that Downs was one of
- your big three fallers. Can you give us the others and then can you give us those players that went
- most outperformed? They did better than expected. Who were the surprises for you in that
- direction? So the three biggest surprises on day one of the draft, number one was Alabama quarterback
- Ty Simpson. So that's like a whole podcast in and of itself around what the implications are of
- drafting a quarterback in the first round who I did not think would go in the first round at pick 13 overall.
- But yeah, so he went at pick 13. There were very few people who projected him. I had this metric weighted mock drafts where
- basically not every mock draft is created equal. Some mock drafts are created closer to the
- draft. Some mock drafts are made by smarter predictors and he only showed up in 4%
- of mocks to the Rams that were in his weighted mocks. There was a lot of noise of him
- going to be traded up for by the Arizona Cardinals, but that didn't happen. So that was the biggest shock of the
- first round was Ty Simpson. But the Minnesota Vikings, they drafted defensive tackle
- Caleb Banks from Florida who I also thought was a second round selection. Coming out of the senior bowl, he was
- one of my biggest risers in my data, had a really lovely senior bowl. He's a big man and he has been
- dealing with some injuries his entire career at Florida. And he re-aggravated an injury at the
- Combine and he was already kind of a inconsistent player, but you can definitely watch a
- couple of games and see like there are some real flashes of brilliance there. In terms of some of the biggest fallers
- in my data, players that fell the most kind of in the draft, at least on night one, one of the biggest values in
- the draft this year was Ohio State linebacker slash edge rusher, Arvel Reese. He was in the running to be the
- second overall pick. And I think most people believe that if he wasn't the second overall pick to the
- Jets, that he would be the third overall pick. Didn't he go five? He went fifth.
- And so that was really surprising. I think that that was one of the biggest surprises of the draft.
- And I think that if Jeremiah Love, like I said, like that was, I think that the Giants were probably thinking about some of
- the other guys that would be available and were completely taken aback. I wouldn't have surprised me if they had
- maybe had Reese number one on their board at the beginning of the draft. So they got a ton of value.
- I think they had a great draft. They also drafted a tackle at pick 10 and Francis Maui Noah from Miami.
- So I thought that was really great value. I also thought the Eagles got some good
- value from this Makai Lemon draft selection. I thought that was like a little surprise
- that he fell a little bit. Ruben Bain from Miami. He fell a little bit in the draft.
- So the draft is full of these like little surprises. So far I'm liking this. It's the Eagles and the Buccaneers.
- I'm liking this so far. Don't interrupt him, Cade. Let Ben keep talking. As a Pats fan, I'm also happy the
- Eagles selected a wide receiver. Yeah, yeah. But yeah, there's all these great stories.
- The other guy that was the surprise really was Omar Cooper Jr., the wide receiver from
- Indiana. Before the combine, he was thought of as a second round pick and then people really
- looked into him more and liked his profile a lot. And the Jets saved him the indignity of
- having to be a day two pick and took him via a trade up at the end of round one.
- So that was a big surprise. I had him as my 20th ranked player and he ended up going 30.
- Ben, that reminds me of a feature of your work that is notable, especially before the
- draft. You plot the trajectory of these players' projections over time. So you show guys not rising or falling
- on draft day, but like in the lead up to the draft, in the weeks up to the draft, guys move up and down
- the mock board. What have you learned over time about characteristics that lead guys to go up and down
- ahead of time? I think a lot of the time, in the preseason, there's a lot of variance.
- Over the cycle this year, there's usually a lot more fallers than risers. So the risers usually just come out of
- nowhere, like you didn't know about them or they were very low on your radar preseason.
- This year, there were at least three quarterbacks who I thought were gonna potentially have the
- profile of being players that would be undrafted. Penn State's quarterback, Drew Aller, Clemson's Cade Klubnick,
- and LSU's Garrett Nesmeyer. Coming into the preseason, all were considered potential
- top quarterback prospects. And along with some other prospects deciding not to enter the draft this year, that made
- the quarterback class really kind of unexciting. But Fernando Mendoza, you also, it takes time
- to kind of, for people to update their priors. So Fernando Mendoza has a couple of good
- games. You didn't think much of him coming into the season, although I think some people did
- think that he was potentially on the radar for prospects. But as he starts to really rack up
- these things and you get the sense that it's not just Indiana beating down on Ball
- State or whoever, and they start to play well against Penn State and Oregon, you kind
- of begin to update your priors and then you kind of, it takes a while for those things to change in a meaningful way
- for the crowd. So the players that do come into the consciousness, they come into it in a big
- way. So the big risers this year were Arvel Reiss and Fernando Mendoza. But Arvel Reiss is probably one of the
- biggest risers in the draft this year, just because in the preseason, they thought he was
- just like a linebacker who was gonna be playing linebacker. No one really cared about him.
- But then he started showing up in all these games and you begin to kind of consider him as potentially the top prospect in
- the draft, somebody who you would pick at second overall, despite not really having played Ed
- Drescher at all as a full-time position. All right, we're gonna have to cut you loose, but Eric's got a last question for
- you. It's gonna be a yes, no answer. Are there team effects in here? Indiana wins- I was gonna ask, oh,
- that team effect, okay. No, no, team effects, like Indiana wins- The college team, college team.
- And now everybody, it relates to Cade's question about risers. Like Indiana wins the national title, then all
- of a sudden wide receivers in Indiana are going higher than expected and all. How does that work?
- I think a little bit. I think you'd like to think that that's correlated, right?
- Like if Indiana's winning the title, they should have a lot of draftable players.
- There's so much attribution there. Like coaching makes a big difference. Obviously playing, you're playing at different positions means
- different things. Luck can mean different things. But yeah, to me, yeah, there's definitely a
- lot more of a microscope on players on good teams. And coming into the playoffs, I thought Fernando
- Mendoza had done everything he needed to do to prove that he was the number one overall pick.
- I thought he had done everything he needed to do. But then there were other players who even
- produced well and people were still saying that they needed to do more. And so for example, that was kind of
- like Ty Simpson. Coming into the playoffs with Alabama, he had been pretty up and down.
- He had gotten hurt. He had played well against some like low level competition, struggled against some others.
- And to be honest, I really didn't think he did a lot more. I was surprised that he went as high
- as he did. But yeah, in terms of team effects on the college side, I think it's kind of
- just baked in in the revealed preferences. On the team side, you were mentioning like,
- are there team effects in my model? My model is team agnostic. I do think that there's not enough historical
- track record for most general managers for us to be able to say anything meaningful.
- And so I choose to believe that you want to inject less certainty into your model.
- And so I don't mean to throw any shade, but ESPN has this draft predictor or draft day predictor model.
- And one of the things, and I'm going to write about it this summer. I've been collecting data from the draft day
- predictor for a long time, like ever since it's been live. And if you do, I'm going to do
- some type of like Breyer score analysis of looking at what your prediction said around the
- time when the player was going to go and then the actual, you can make that comparison.
- And my guess is that when you do tend to inject, like the teams are only considering these players based on scouts, inks, needs,
- or the teams are only going to consider this player at this pick because of that.
- You're just putting your finger on the scale when you need to be letting it ride,
- letting the randomness, the flatness of some of these priors instead of trying to over-priorize
- your model. Super interesting. You're full on flirting with us when you're talking Breyer scores.
- I mean, come on. Just directly flirting. Why not log loss? Yeah, yeah, exactly. I think we'll do both.
- We'll do both. Just an example. Do it both ways. Exactly, do it both ways. All right, Ben, we should let you go,
- man. Thank you for making time. Congrats on making it through another draft season.
- Good job. And we'll look forward to seeing what you do next year. Yeah, thank you guys so much.
- And like you mentioned, my data is everywhere, but this past year, grindingthemocs.com, my grinding
- themocs substack, I'm writing there pretty regularly. And then I also had my data in
- NFL IQ, which is NFL Next Gen Stats off-season dashboard, which was pretty cool and
- exciting. And yeah, no, this year was great. Getting to work with about a quarter of
- the teams in the league has been something that I've been working towards for many years.
- So it's a long time coming. It's great to be back on with you. Great fun, good work.
- You've been grinding. You have literally been grinding and it's good to see a payoff.
- That has been Robinson, founder and CEO of Grinding the Mocs. And that has been a full NFL Draft
- Show. We seem to be themed these days. About two weeks ago, we did a full golf show.
- That one was unintentional. This one was planned. We'll get back to a little more open
- -minded, broad-minded, including baseball. And let me point out to some of you guys, they're playing playoff hockey right now.
- It's the best thing going. It's great. Enjoy it. For the whole team here, for Eric Bradlow,
- Shane Jensen, Adi Weiner. For our guests, Ben Robinson and Richard Thaler. For our production team, Dion Simpkins, making it
- always happen. Marissa Patel, I mean, Marissa Reyna. And Dee Patel, the boss lady, Dee Patel.
- Thank you guys for everything. Thank you guys for listening. Come back and join us next time.
- Between now and then, enjoy your sports.
Brief Summary
Richard Thaler and Ben Robinson discuss NFL draft analytics, team decision-making, and the ongoing relevance of outdated valuation methods.
Key points
- Nobel Laureate on NFL Draft. Richard Thaler, a Nobel Prize winner, discusses his surprising interest in NFL draft analytics.
- The NFL Draft Trade Chart. Thaler reveals that the NFL draft trade price list is outdated and inaccurate.
- Perception in Trades. Teams avoid deviating from the Jimmy Johnson chart to prevent looking foolish.
- The Risk of Looking Dumb. Coaches are afraid to make bold moves due to the risk of being fired.
- Draft Capital and Wins. Accumulating draft capital through trading down correlates with more wins.
- Ben Robinson's Influence. Ben Robinson has become central to the NFL draft conversation with his data insights.
Episode highlights
- Nobel Laureate on NFL DraftRichard Thaler, a Nobel Prize winner, discusses his surprising interest in NFL draft analytics.“Why is a Nobel laureate messing around studying the NFL draft?”1:48Apr 29, 2026
- The NFL Draft Trade ChartThaler reveals that the NFL draft trade price list is outdated and inaccurate.“The price list is very wrong.”6:44Apr 29, 2026
- Perception in TradesTeams avoid deviating from the Jimmy Johnson chart to prevent looking foolish.“You never wanna look like the sucker.”24:30Apr 29, 2026
- The Risk of Looking DumbCoaches are afraid to make bold moves due to the risk of being fired.“They don’t wanna look bad and they run a big risk of getting fired.”27:11Apr 29, 2026
- Draft Capital and WinsAccumulating draft capital through trading down correlates with more wins.“Trading down, accumulating draft capital is positively related with downstream wins.”28:58Apr 29, 2026
- Ben Robinson's InfluenceBen Robinson has become central to the NFL draft conversation with his data insights.“Ben's work has become central to the conversation around the NFL draft.”34:16Apr 29, 2026
- Predicting the DraftA discussion on how to accurately predict player draft positions and future value.“I'm trying to predict where players will get drafted.”46:15Apr 29, 2026
- The Value of DataExploring how data analytics impacts player evaluation and draft predictions.“My data predicts the draft the best. I'm very proud of that.”46:53Apr 29, 2026





