357: Finding Meaning From the NFL Combine

00:00:02
Speaker 1: This is the Label podcast with your host JJ Zachersa, JJ zacher Resa.

00:00:11
Speaker 2: What's up everyone. It's JJ Zacharyese and the editor in chief at fandle and at Number five dot com. In This is episode three hundred and fifty seven of the Late Round podcast with many shows that are part of the FanDuel podcast Network. Thanks for tuning in. So over the last month or two, I've consistently been saying things like my model doesn't care all that much about the NFL combine and people tend to overrate the combine. That's all based off of what I've researched, but I haven't done the same type of research as today's guest has. That guest is Kevin Coole. You guys might know him from Twitter dot com as at Kevin COOLEPFF, but Kevin's a data scientist at Pro Football Focused, and he's one of the most thoughtful processed football analysts that you're gonna find. Recently, he put together an article on PFF called players with the best NFL Combine results for Fantasy Football Success, and in that article he did some really interesting analysis on how important the combine is for each position in fantasy football. So I wanted to talk to him about it, and that's what I did. So, Kevin, you wrote an article on Pro Football Focus recently. You talked about utilizing combine results to help in fantasy football, and in doing so, use the tree based modeling approach. Can you sort of explain that approach and how it works?

00:01:26
Speaker 1: Sure, yeah, I mean, not to get too technical, but what happens here and I illustrated in the piece, is that all the different metrics are fed into the model, and what it does is it splits the sample by different numbers. So, for instance, when we're looking at running backs, one of the first splits in a sample tree that I put on there says if whether or not a running back runs faster than four point four seconds on their forty yard dash, So it kind of splits there. And it does this splitting to try to find the most concentrated samples of what you're looking for and in this case in a fantasy football way.

00:02:01
Speaker 3: And then the.

00:02:02
Speaker 1: Different approaches is kind of do this over and over and over again, and then you get an idea of how these things work and the benefit of using this type of modeling approach versus a linear regression or something like that is that sometimes there are interactions between these variables. Sometimes, you know, a three code time for a very big running back could be more meaningful than it is for a smaller running back or something like that, where it's just not all a linear relationship. So that's why I use this modeling system to be a little bit different than some other approaches.

00:02:30
Speaker 2: Yeah, so looking at the combine results, obviously there's going to be different levels of meaning for different positions, whether you're looking at quarterback or running back or wide receiver. Based on your approach, which of the main fantasy football positions seem to show the tightest connection with combine results? And then where did these results seem to matter the most.

00:02:48
Speaker 1: I mean, it shouldn't come as a surprise that the running back position is where it seems to matter the most. I think most people in fantasy football kind of know this by now when they think about the replaceability of the position.

00:03:00
Speaker 3: And one of the factors there.

00:03:01
Speaker 1: Is that, you know, these guys are all very highly skilled and it's a very athletic sort of position. I mean, we saw instances during the season where running backs can just step in and then end up playing very well in a particular team in a system, and just because it's very much an athleticism instinctual based position. So I think for running backs, we were able to get scores and I was projecting in here at the percentage that a percentage chance that a player would be successful, meaning would get a top season, And for running backs it was a top eighteen season within their first three years. And for them, you know, we're able to get numbers into the seventies just looking into seventy something percentile for top guys just looking at their combine and not knowing anything else about them. It's also a lot more accurate in other studies that I've done in forecasting draft position. So it seems like athleticism matters most for that position.

00:03:53
Speaker 2: Yeah, you know that. So let's let's talk about the draft capital stuff, because I obviously, you know, I'm modeling this stuff as well from a running back and wide receiver perspective, and production is clearly a big piece of that. And whenever I'm looking at what's working within the model and what's giving it that higher R squared and whatnot, I've noticed that combine results aren't all that impactful for running backs and wide receivers. But draft capital is part of my model in some way because I don't want to just throw away probably the biggest and most important piece of information when I'm trying to project what's gonna happen in fantasy football, because I'm drafting after the NFL draft happens. So the narrative that I've sort of told myself as to why that's the case is because it impacts draft capital. The combine can impact draft capital, and if draft capital is sort of capturing that, then that's why the combine isn't mattering as much within those models. Does that make sense? Is that logical?

00:04:53
Speaker 3: Now? That that makes a lot of sense.

00:04:54
Speaker 1: And in fact, another study that I had done, which was looking at projecting draft position based upon each each of these variables separately and getting some coefficients and getting some measures around how impactful the different drills are, and then also projecting real value in our war metric for these different players, you can see how closely linked the two things are. But usually it is some of these things getting out of a higher impact on what goes on as in the draft than what.

00:05:22
Speaker 3: Their actual value is.

00:05:23
Speaker 1: So I think you're right to say that draft capital often trumps everything, But again, it kind of depends on the position as far as how important it actually is. And these positions have changed a lot over time too, which I think is another interesting takeaway that I had from this analysis is it's still pointing to things like big, fast running backs being important. It's still pointing to big receivers being important because that was really the prototypical way of getting scoring for most of the sample going back to two thousand and six till today, and things have changed quite a bit, So I think that's also.

00:05:53
Speaker 3: Something you have to keep in mind when you're looking at these models.

00:05:56
Speaker 2: So what measurements in particular, You mentioned that running backs seem to have the most impact, but what measurements at the combine seem to matter the most at the running back position.

00:06:04
Speaker 1: I mean, it's a smattering them, but I think that the forty yard dash is highly important, though getting less important and things The broad jump is also something that's kind of been steady as far as it's importance.

00:06:14
Speaker 3: When I'm looking at studies that I did a few years.

00:06:16
Speaker 1: Ago versus today, and you know, three code time is becoming much more important. That was something that popped a lot more in our war Metric for real player value because of the fact that we overvalue receiving so much more to rushing as far as it's the fact that it can't be replaced as easily. But even the fantasy football side, as everyone knows, it's becoming a bigger and bigger thing, and you're seeing players like, you know, someone like Christian McCaffrey. His size adjusted speed was good, but not fantastic. But then when you have you know, just you know, six six six three code time and things like that, that's what really ends up popping and you're seeing it in the receiving game, which is becoming super important.

00:06:52
Speaker 2: Yeah, were there any particular running backs that stood out in a good or bad way based on their combine results?

00:06:58
Speaker 3: I mean, I think.

00:07:01
Speaker 1: It kind of depends on how you were viewing them going into this thing, But some of the top backs that were perceived coming out of it were some of the top backs that I had in the model, So Jonathan Taylor and AJ Dillon coming near the top, although Taylor, you know, somewhat significantly above Dylan for the fact that in other models that I had done they were fairly close. And I think there's some reasoning behind why them It might be the case. I mean, some others are some other popul names like cam Akers and others who had done well, But I think generally the class was not that strong. At least if you look at the success probabilities that were coming up here, most of them were in the Just like I said, this is strictly based upon combines, so we don't know about draft position, but most of them were in the high thirties, and then Taylor was the only one in the forties, So they're not really approaching some of the highest numbers that we've seen over the historical period, which were, like I said before, some of them well into the seventies.

00:07:53
Speaker 2: Interesting, very interesting. What did you find at the wide receiver position. Were there any particular measurables that seemed to matter a little bit more.

00:08:01
Speaker 1: Yeah, I hinted at this before, but size is probably more important from a fantasy perspective than from a real football perspective. So someone like Henry Ruggs, who people seem to be falling in love with it kind of more and more, and the combine helped as part of that process. After running that ridiculous forty time, he's quite a bit down on my list as far as the highest success probabilities, and it probably has something to do with the fact that that he's a smaller guy. And I think generally this is really a really interesting class when it comes to wide receiver size. I know that battle was something that you know, a lot of blood was spilled over over that a number of years ago about wide receiver size. But this is really an interesting test case because we have the top three receivers who are, you know, under two hundred pounds, and Rugs and Junior are not really close to two hundred pounds, whereas last year, you know, people weren't so high on the receivers last year, but you had a lot of guys like Metcalf and aj Brown and and Deeple even guys like people Sam you don't think it was being big, but he was two hundred and fifteen pounds. So these guys were well over two hundred pounds.

00:08:57
Speaker 3: This year.

00:08:57
Speaker 1: It's really you know, the highest names are under tw hundred pou So it's gonna be interesting to see if they can really be the target hogs that you need to be to have that fantasy success.

00:09:05
Speaker 2: So I'm assuming then of sort of the guys who looked good and the guys who didn't look good in the model, like a Chase Claypool might have looked good, whereas some of those leaner guys didn't look as good.

00:09:16
Speaker 3: Yeah, exactly. I mean Claypool is up there. I mean Mems is up there too. He's not that heavy.

00:09:21
Speaker 1: But to get hundred pounds well, you know, pretty decently over tw hundred pounds and put up the rest of the measurements that he did, it's just going to look pretty awesome. But by the standard. But again, like I said, these are guys where the success probability is more like a coin flip as opposed for the top guys. And so these guys are closer to being some of the top all time numbers that we've seen since two thousand and six, yet they're still relatively low in their success probability versus running backs because the combine just doesn't matter as much.

00:09:50
Speaker 2: Yeah, gotcha. So I personally don't have a model for tight ends because production just seems to matter a lot less of the position from what I've tested, So I really don't talk about tight ends that much on the show. It also doesn't matter as much from a fantasy football perspective, so I don't really spend as much energy and time on them. You may or may not have found the same thing. But what combine metrics stood out at that position in particular.

00:10:14
Speaker 3: Yeah, I mean again, it's it's speed more than anything else.

00:10:17
Speaker 1: And we talked about this this, you know, feedback loop between combine measurables.

00:10:23
Speaker 3: Draft position, what's important or not.

00:10:26
Speaker 1: Again, this is one of these situations where speed is going to influence draft position pretty heavily. I mean, you see guys that you know aren't necessarily overdrafted, but you know, a fast tight end can can move up boards. It seems like every year except for George Kittle for some reason. So I think this year, what we've seen from you know, what we saw from Albert Oh.

00:10:45
Speaker 3: And I'm just going to keep it out.

00:10:47
Speaker 1: Albert Oh, you're running that four four nine. He's gonna I mean, I can't see him not taking a huge jump as far as people's expectations. If he comes in and he starts to look good in some of these team visits or pro day things like that, you know, he shows that he's capable catching the ball, everyone's gonna think he's gonna be able to run up the seam and make big plays. So I feel like that could contribute to it to a higher draft position, which will help for success, But it doesn't necessarily translate that well to position. And in fact, the historical projections we're pretty much over as far as getting it right on some of the top ones that they had for historical projections, because it's just so hard at the position.

00:11:23
Speaker 2: Yeah, it's a really tough position to prospect. I think overall, there's just so many different things that a tight end does. That's a question that often gets asked on this show is why do tight ends take so long to sort of get acclimated to the NFL level? And I think that the very basic answer is that they're doing a lot more. There's a lot more to learn. It's a very skillful position, and I think that just adds on to why it's such a tough position to prospect. I think ending on a very very basic question here to sort of wrap things up makes sense for fantasy football purposes. Kevin, does the NFL combine matter?

00:11:55
Speaker 1: As much as I like the nothing matters nihilist that we have going on here, I would say it does matter, and I think that it matters in a couple different ways. One for a position like running back, it literally matters, So it is something that is significant, and you see significance in some of the results and what we've seen in the past. And I think it also matters as far as understanding the relationship between the combine and success, in particular at the wide receiver position. When I was looking strictly at draft position in my model, there the wide receiver there was one of the heaviest influences on things like the forty yard dashtime on their draft position and moving the draft position up. But yet it had some of the smallest influence when it looks at their actual value. So I think it's important in knowing how much it has been overvalued. So when you see someone like a Ruggs we mentioned earlier moving up boards quite a bit, some people are talking about it was potentially the wide receiver one. You know, things that happened with John Ross in the past, that that sort of stuff. I think it's just important to keep in mind on and kind of knowing what to fade, not only knowing what to pick up on and to apply.

00:13:01
Speaker 2: Yeah, that's a really good point. Awesome stuff though, Kevin, I appreciate you hopping on and talking about your article. If you guys aren't following Kevin in his work, you can find him on Twitter at Kevin cole PFF. You can also find me on Twitter at Late Round QB. Make sure you're following me on there, and make sure you're subscribed to this podcast, the Late Round Podcast, which you can find pretty much anywhere podcast can be found. Thanks for listening everyone, I'll catch you later in the week with this week's Mailback episode.