This offseason, I conducted a study on Average Draft Position (ADP) since 2007. I found some insights that I hope will give you an edge in 2018 redrafts.
The Highlights
- Early round draft picks are even more valuable than you might think.
- Target upside after the third round.
- Overall ADP reports are no longer reliable as your Big Board.
- Position rankings are more important than overall rankings.
- Historical data is still the best method of predicting future outcomes.
One quick note: I’ll be referencing Leverage frequently in this article. Leverage is an Action Network metric that normalizes fantasy points for a given season. When you read “Leverage,” think “Fantasy Points + Positional Scarcity.” For a more detailed breakdown on Leverage and its uses, check out my article where I first introduce Leverage.
>> Sign up for The Action Network’s daily newsletter to get the smartest conversation delivered into your inbox each morning.
Draft Value Declines Logarithmically
Since 2007, the Pearson’s correlational coefficient between ADP and Leverage is 0.3738, which falls just short of the scientific standard for statistical significance (0.40). Pearson’s coefficient measures the strength of correlation between two variables based on the assumption of a linear relationship. In this case, it’s measuring how effective ADP predicts player value.
However, intuitively, we all know there is not a linear relationship between a player’s ADP and his relative value. Early round selections are worth much more than mid-round selections, and player value does not decrease evenly from pick to pick. It’s more of a logarithmic decline than a linear one.
So, in addition to Pearson’s coefficient, I also calculated Spearman’s coefficient for the same data. Spearman’s coefficient uses a moving-walk model rather than a linear one. The Spearman’s coefficient for the dataset was 0.6233, which is considered to be a moderate-to-strong relationship. The improvement from Pearson’s to Spearman’s confirms something we already knew intuitively: The relationship between ADP and player value is a logarithmic decline — not a linear one. This means that your early round selections are of paramount importance.
Hyper-Prioritize Your Early Round Selections
The players you select in the early rounds are likely the ones to carry you to the playoffs. Championships may be won and lost in the mid-to-late rounds of the draft, but it’s the early rounds that put you in contention in the first place. Accordingly, you should be discerning and critical of players ranked in this tier. Leverage as much draft capital as possible into positions with the greatest positional scarcity (first RB, then WR).
Moreover, due to the strength of the Spearman’s coefficient across all positions, it’s also advisable to target elite quarterbacks and tight ends if the price is right. There exists a meaningful predictive edge in selecting players at the top of each position. If you can grab a top-end QB or top TE once you’ve hoarded value at RB and WR, you should do it. Most people believe in the principle of accepting value when it falls to you, but very few people actually ever pull the trigger. Don’t be that kind of fantasy drafter. If Aaron Rodgers or Rob Gronkowski falls to you at an advantageous spot, don’t hesitate to select either.
After the Third Round, Pivot Towards Upside
Since draft value declines logarithmically, there is a larger-than-expected tier of players in the middle rounds that has approximately the same range of predictable outcomes.
Do you see that huge increase between the 80th and 70th percentile of players drafted? That correlation spike represents a major tier break in player value beyond the 80th percentile (usually the first three rounds). Beyond those first three rounds, the correlational coefficient improves dramatically.
When we see a massive spike in the coefficient like we do above, it signals to us that ADP is serving a much greater role in predicting outcomes at that inflection point. This means there’s significant predictable value in selecting a player before the 80th percentile. After the third round in most drafts, it becomes harder to discriminate player value, so you should pivot toward upside projections rather than median projections.
Overall ADP Is No Longer Reliable as Your Big Board
Below is a line graph tracking ADP’s ability to predict end-of-season production since 2007:
ADP’s predictive validity has tanked in the last four years. There are several possible explanations for this, but one draft trend immediately springs to mind. The proliferation of Zero RB (and other contrarian strategies) has introduced more variance to recent ADP reports. As more people are drafting players based on vastly different strategies, ADP reports are likely more variable — and thereby less sharp — compared to those of previous years.
No matter what has caused the decline in ADP’s predictive validity, one thing is certain: Overall ADP is not currently predictive of end-of-season success. Be careful with Big Boards built around ADP and public sentiment.
Position Rank Matters More Than Overall Rank
As a reminder, the all-time correlation between ADP and Leverage is 0.3738. By comparison, ADP Position Rank produced an all-time correlation of 0.4670. Given the sheer size of our statistical sample (1,771 drafted players over 11 years), this 0.0932 statistical difference is massive.
Because position rank is more highly correlated to success than ADP is, if you rate two players equally, you should consider selecting the player with the higher relative rank at his position. Just as importantly, you mustn’t look at raw rankings but rather the percentile of a player’s rank. The highest percentile position player has the best odds of performing at or above expectation. This point seems intuitive, but fantasy drafters ignore it all the time when a position run begins in a draft. If you find yourself in a position run, seize value at an under-drafted position instead.
How to Utilize ADP in 2018 Redrafts
To recap, positional rankings have produced stable correlations over the last 11 years — and so have historical leverage projections at an even greater magnitude. The one ADP metric that has plummeted over that time frame is overall ADP. So, here are my suggestions for you in 2018:
- Prioritize fantasy analysis grounded in historical data — it’s the most reliable method of predicting future outcomes.
- Positional rankings are still trustworthy and effective. You can trust the wisdom of the crowd to guide your rankings for each different position, so long as your “crowd” is populated by sharp analysts. And speaking of sharp analysts, I’d highly advise you use the Action Network’s Consensus Rankings — I’ll be using them this draft season.
- Do not trust overall ADP reports to act as your Big Board; develop one yourself. Public ADP reports should not be trusted to evaluate players at different positions. Instead, trust your own intuition (driven by an analytical, data-based mentality) to guide you. To quote Jonathan Bales quoting Peter Thiel, “‘The most contrarian thing of all is not to oppose the crowd, but to think for yourself.'”
Check out our fantasy football rankings:
PPR: Top 200 | QB | RB | WR | TE | D/ST | K
Standard: Top 200 | QB | RB | WR | TE | D/ST | K
Player Projections: QB | RB | WR | TE | D/ST | K
Printable PDF Cheat Sheet: Download now!
Player Profiles: Top 200 | QB | RB | WR | TE | D/ST | K
Photo Credit: Derick E. Hingle – USA TODAY Sports
Pictured Above: Alvin Kamara
