Today I put the Tape Detective costume in the closet of my mom’s basement, and instead put on my nerd hat. I looked at which stats help (and don’t help) predicting WRs for fantasy football, and why that’s the case.
If you find any of this useful, you should get into the best ball streets with promo code “UNDERBLOG”. Any first-time deposit will be accompanied by an additional $10 courtesy of the fine folks of Underdog Fantasy (aka me). Use this link to go straight into the sign-up sheet.
Fantasy Points Per Game: It’s simple, yet effective. I have all of last year’s half PPR per game stats here, along with their expected half PPR points per game based on their adjusted usage. It’s the two stats I use the most when building projections and rankings.
Yards Per Game: It gets even simpler. Because touchdowns are fluky and receptions aren’t as important in half PPR scoring, yardage metrics are better predictors. Yards per game will get the job done for most non-young WRs who are playing with the same QB.
Targets Per Game: Targets are indeed a skill, but they are slightly less predictive than fantasy points and yards per game. Some WRs are just less effective on their targets than others. That difference is why it falls into the “kinda matter” bucket.
Yards Per Route Run: It’s a fantastic efficiency stat and it is helpful for determining how good young WRs are, but it is less predictive than fantasy points or yards per game in general. Removing a WR’s longest reception of the previous year added some predictiveness because long YAC plays are volatile. That’s why Hunter Renfrow had the decent yards per route run season as a rookie even though he’s a replacement-level slot WR.
PFF Grade: Their grades closely align with yards per route run. There could be something to creating a stat that combines fantasy points per game and PFF Grade. I like the idea of combining production and a film grade.
Touchdowns: It’s no surprise that touchdowns hold less predictiveness than fantasy points, yards, or receptions. They can be fluky, which is why you’ll see my touchdowns over expected column pop onto Underblog soon.
Avoided Tackles: We love them for RBs, but it’s not very useful for predicting WRs, particularly if citing avoided tackles per catch. What a WR does before and at the catch point is more important than how much elusiveness or physicality he has after the catch. The latter is baked into fantasy points and yards per game anyway. Just use those stats.
All “Per Reception” Data: This can be yards per reception, yards after the catch per reception, or others. The only “per _______” numbers we should care about are “per season”, “per game”, and “per route run” when it comes to predicting a WR’s fantasy points.
Drop Rate: With a correlation of -0.13, drop rate has almost no impact on next year’s fantasy points. In fact, total drops (not drop rate) are positively correlated with next year’s fantasy points. The only time drops matter are when a coach removes that player from the starting lineup. That’s somewhat rare, however.
Contested Catch Rate: Awesome for highlight tapes. Less awesome for predicting fantasy points. Contested catch rate has a correlation of 0.05 to next year’s fantasy points per game. Focus on WRs who get open, not who win on less-valuable sideline fades.
Average Depth of Target: This is nothing but a “how he’s used stat”. aDOT is useful in assessing weekly volatility (that’s helpful in DFS and for spotting “better in best ball” WRs), but it alone has zero predictiveness in projecting next year’s fantasy points.
% of Snaps in the Slot: When we hear news of a WR playing more snaps in the slot in training camp, remember that WR slot rate has a slightly negative correlation (-0.11) with next year’s fantasy points. Where a WR is lining up is solely a “how he’s used stat”. It alone is not helpful for predicting fantasy points. Myth busted.
Only WRs with 100 routes run in both “SeasonA” and “SeasonA+1” qualified. There were 1,007 qualifying WRs since 2010. The “SeasonA” stats are predicting half PPR points per game in “SeasonA+1”.