867: NFL Combine Winners and Losers

SPEAKER_2 (0:02)

This is The Late-Round Podcast with your host, JJ Zachariason.

JJ Zachariason (0:11)

What's up, everyone? It's JJ Zachariason in this episode 867 of The Late-Round Fantasy Football Podcast. Thanks for tuning in. There's definitely a large group of people out there who think the NFL Combine doesn't matter, that you can't actually take anything away from the event, and that it does more harm than good to your analysis. I disagree with that wholeheartedly. I'm someone who's scouting through analytics, as 99% of you know. And I think when I say that, a lot of people assume that that means that most of the numbers that I work with come from the NFL Combine. But that couldn't be further from the truth.

As you heard over the last two weeks on the show, age-adjusted college production is far, far more important than anything that you'll get from the Combine. You can listen to those shows over the last couple of weeks to give you an idea of what I mean.

Most of the NFL Combine is irrelevant to me. Most of it. There are things that I'm still looking for and things that I'm still inputting into my models. There are things that are still important at the event. Not everything at the Combine has to be important for the Combine to matter.

Now, within my wide receiver prospect model, I don't have any inputs from the NFL Combine, not even height and weight inputs. I don't think that means that those things don't matter, though. It's just that draft capital itself, that's an input, and draft capital is likely capturing what it needs to capture at the Combine. In other words, if a wide receiver runs a really fast 40, that doesn't get signal in the model by itself, but that wide receiver probably improved his draft stock, and that's the piece that still goes into the model. So I can't definitively say that none of the things at the Combine matter at wide receiver. Running back's a little bit different. Running back weight is a model input, and the Combine gives me and us a standardized way of reporting that data. Of course, some players are gonna be a little bit heavier or lighter based on the time of year, how much breakfast they ate that day, all that kind of stuff. But over a large sample, it's a great way to have numbers from players in the exact same environment, because some of these schools, they report some wild height and weight measurements. Now, on top of that, the running back model uses speed score, which is a metric that takes a player's 40 time, and it adjusts that for how big they are, how much they weigh. Because a smaller player should be running faster than a bigger one. And over the last like 48 hours, I've changed the way the model reports speed score, albeit slightly. It's not that big of a change, but it is a change. Historically, the model has looked at speed scores in a pretty binary way. If a player was below the 90 mark, which is a pretty low percentile, he would get dinged. His overall prospect score would be pushed down. Rather than having things so binary, I don't really like doing analysis that way, I did more testing recently and discovered a new way of doing things. 92 is actually the speed score threshold that I'm looking for. But a player with a 91 speed score isn't scored the same way a player with an 81 speed score would be. The player with a 91 score isn't getting hurt as much in the model. Conversely, I'm now looking at high-end speed scores as well. A player with a speed score above 112, he's getting extra love in the model. And then a speed score of 122, let's say, that gets a little bit more love than one of 112

Now, to be clear, I'm working within extremes here. The majority of players, like 70% of them, they're scored the exact same way in the prospect model when it comes to speed score. Because their speed scores land in that 92 to 112 range. The rest are where this can come into play. And most of those players, it's not even changing their scores dramatically.

Now, at tight end, speed score is a much bigger input. The tight end model has production metrics, but athleticism matters more at that position than at running back and wide receiver. At least when we're trying to predict fantasy success.

I bring all of that up because today, I'm gonna talk through some combine winners and losers. And I'm gonna point back to some of these model inputs when talking about the players. You can see how a player's combine is then impacting their overall outlook based on the stuff that I look at.