883: Using Running Back Prospect Scores in Season-Long Leagues

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, 883 of The Late-Round Fantasy Football Podcast, sponsored by Underdog Fantasy.

Thanks for tuning in. As almost all of you know, I have a prospect model and a year two model. The prospect model, called the ZAP model, helps me project how well a player is going to perform during the first three years of his career. The year two model takes inputs from a prospect's first season in the NFL, and it tells me which players are best set up for their second and third seasons in the league.

I'm confident in those models because I've tested them. I also now have a few years of being an analyst while using them, and I've seen good results, anecdotally. The models changed a bit this offseason, but the backbone is still there.

I know that the ZAP model, the prospect model, is more predictive than rookie draft average draft position when looking at a player's first three years in the league. What I don't fully know is whether or not we can use these models for our season long leagues, our redraft leagues. So of course, I did a little data analysis to find out. I looked at historical average draft position data or ADP data and sort of layered prospect scores and year two scores on top. And I want to share some of those results.

Today, I'll be talking running back. Next week, I'll do the same thing, but at wide receiver. Now, before digging in, I want to remind you about a special underdog fantasy offer. If you've never played over at underdog, you should change that immediately. They've got best ball drafts, a great pick them game and more for you to try out. Yours truly is going hard with best ball drafts this year, in particular, and I'm creating content around that experience as well. You can use that content to your advantage when drafting, and you can also use the Late-Round Draft Guide. The guide is currently available for pre-order over on late round.com. But for anyone who's not on underdog, if you sign up today using promo code late round, you'll get two things. One, a really strong deposit offer, but two, on July 1st, when that draft guide is finished, it'll be emailed to you. You get a free Late-Round Draft Guide. So head to underdogfantasy.com or download the Underdog app. When you go to sign up, use promo code late round and make it a deposit. You'll get that deposit offer along with a Late-Round Draft Guide that'll be delivered on July 1st. Again, that's promo code late round.

So the first thing I need to do is touch on the ZAP model itself. The model spits out prospect scores for players who either went to the NFL Combine or who were drafted. And the model dates back to 2011 Scores are on a scale from zero to 100 But because there were so many players in the model who were undrafted or who went in the late rounds of the draft, anything below a score of like 70, that's when things start to be a little concerning. Or I should say, those scores don't hit at a very high rate. Scores of 95 or better, those are typically reserved for the elite players like Bijan Robinson, Jameer Gibbs, those types of players. And then the 90 to 95 range is filled with very solid backs. Players in the 80s can kind of go either way.

The year two model is built pretty similarly. Any score from 95 to 100 is gonna capture an elite player for the most part. Anything in the 90 to 95 range, you've got a really solid player. But with the year two model, when you get below 90, you run into more issues. The reason you run into issues below 90 is because the model is more predictive. It's more predictive than the prospect model. You're now looking at a year of actual NFL data with that model, with the year two model. So we have a sample of the player playing in the league. That's huge. Now this is important to wrap your head around because I'll be referencing different prospect score ranges throughout today's show. I'll do my best to help you out so that you don't get lost, but that's the gist of how the models work and how the scores are distributed.

Now my goal with the analysis was to just see if there was anything glaring when it came to where we've been drafting running backs in the past in our redraft leagues and how strong those running backs prospect scores were or year two scores. Like let's say hypothetically that when you drafted RB2s in the past, guys in the RB13 to RB24 range, let's say all of the running backs in that range who had prospect scores under 90, never score 10 or more PPR points per game in a season. That would be interesting, wouldn't it? It's not true, but it would be interesting, right? I wanted to see if there were any trends like that.