Photo: Imagn Images

Last week I wrote about how, in a world where there are hundreds of different "draft strategies" and "master plans" to choose from, there's only one that is guaranteed to be a winning strategy in 2023: building your draft around players who fall past their average draft position (or "ADP fallers"). And I walked through the details about why I think this is the best strategy in fantasy football, providing proof for why it's guaranteed to be a winning strategy in expectation.

(The executive summary for those who missed it: all players fall past ADP at roughly equal rates, so "ADP fallers" gets you the exact same players as everyone else, but it gets you them at a discount. Getting them at a discount means that— as long as players drafted earlier tend to outperform players drafted later (and they do)— you pair all of your players with better teammates and therefore have a better chance of winning than everyone else who drafts the same player.)

Conceptually, the approach is bulletproof. Not every team that drafts ADP fallers will win this year, but all "ADP fallers" teams combined are guaranteed to finish the year with a winning record.

A Real-World Breakdown of the Edge

After publishing, someone on Twitter pointed me to this analysis from 2018. It looked at all managers who entered at least 50 MFL10 leagues in 2017, then tracked the win rate of the top and bottom 21 managers in terms of ADP value per pick (essentially: who reached on players the most ahead of ADP and who grabbed players the most beyond ADP).

Does this 'value' hocus pocus work? These are the winrates of the top and bottom owners in value per pick playing at least 50 leagues last year. pic.twitter.com/KZrJnn4KTG

— Gentan Schulteis (@GentanSchulteis) August 14, 2018

For those not familiar with MFL10s, they were 12-team draft-only best ball leagues. Managers paid $10 to enter, drafted a team, and the league played out without any further input. Whichever team scored the most points for the season won $100.

The naive win rate in an MFL10 should be 8.3% (or 1 in 12). To post a profit for the year (and if you're playing in 50 leagues, you're probably trying to turn a profit), managers would need to win with at least 10% of their entries. So how did the "get your guys" managers and "chase value" managers do?

Some individual "reachers" turned a profit. 3 of the 21 teams that reached the most to secure their preferred players posted a winning rate above 10%. One even got as high as 14.3%, turning a 43% profit on his or her entry fees! But as a group, the "reachers" won just 6.4% of their leagues, on average, eating a 36% loss for the year. (Remember, a perfectly average player would lose 17%.)

Similarly, some "value drafters" lost money on the season; seven teams failed to win at least 10% of their leagues (though only two teams won less than the 8.3% they'd be expected to win by chance alone). But 14 out of 21 teams turned a profit, 8 out of 21 teams outperformed the best of the "reacher" managers, and three teams won more than 20% of the time, which means they more than doubled their entry fees. (The top performer won a whopping 26.3% of the time, winning an extra $16.30 for every $10 entry fee.)

Note once again that "ADP fallers" is only a winning strategy in expectation. Even the best "ADP fallers" manager lost nearly three-quarters of his or her leagues, and a few finished with a below-average win rate. But as a group, the value drafters more than doubled the win rate of the "who cares about value" drafters, winning 13.6% compared to 6.4% with a 36% net profit that perfectly mirrored the other group's 36% net loss.

Hopefully, if you weren't convinced before, you're at least open to the idea that winning in fantasy football really is this simple. (I happen to be a very big fan of simple, which is often underrated!) But just because it's simple doesn't mean it's easy, so let's get down to brass tacks: just how does one execute a "chase ADP value" draft strategy?

Option #1: Autodraft

I'm not joking; the easiest way to stockpile ADP value is to never show up to the draft in the first place and let the computer make all your picks for you.

If you've been playing fantasy football for a while, this idea might horrify you; when I started playing, using the autodraft pretty much guaranteed you were walking out of the first eight rounds with five quarterbacks and three players on season-ending injured reserve. But it's 2023, and site autodraft algorithms are actually— dare I say it?— pretty good. They no longer equally weigh drafts from three months ago in their ADP. They now have some minimum basic logic (like "don't keep drafting quarterbacks when the team already has three"). They're quicker at removing injured players from the pool.

Perhaps you've seen this in action. Every year I get bewildered messages from managers who are shocked that the team that never even showed up to the draft is somehow in first place. ADP fallers is the reason why! Autodraft algorithms are unsentimental value-maximizing machines.

Now, I want to be clear: when it comes to drafting ADP fallers, autodraft is table stakes. If you're going to manage your draft, "do better than autodraft" is the bare minimum you should shoot for. Better than autodraft is achievable with some prep and a good game plan, but if you don't have the time or the interest, there's no shame in letting the computer take the wheel.

If autodraft operated on fresher ADP (e.g. if it was quicker at removing injured players from the pool, etc.) then it actually would be optimal. https://t.co/7AV9PuA2dX

— Adam Harstad (@AdamHarstad) August 27, 2021

Option #2: Draft for Yourself

If you, like me, are not yet ready to cede your job to our new AI overlords, we can always manage things the old-fashioned way and pick our own players. When going this route, we need to grapple with the fundamental disconnect between theory and reality: "true" average draft position is unknowable. Conceptually, average draft position is a representation of where players are being drafted at this exact moment in time in leagues exactly like this one.

But we don't have that. Instead, we have data on where players have been drafted in recent weeks in leagues that are hopefully roughly similar to your own. So the first thing we must do is take that rough tool and shape it into something more useful.

While we don't want The Algorithms to draft our team for us, that doesn't mean we can't enlist their aid. Specifically, Footballguys' Draft Dominator tool has some very cool features under the hood that will make our job much easier. Most notably, the app automatically adjusts its ADP data to represent its best guess of where players will go in your specific league (listed in the app as "ADP+").

If your league offers 1.5 points per reception for tight ends, the app will infer that tight ends will be drafted higher than the ADP suggests. If your league offers 2.0 points per reception for tight ends, it infers they'll go even higher still. If your league only starts two receivers instead of the more-common three, it will adjust ADP values for receivers downwards to match.

The app doesn't (and can't) do a perfect job, mostly because it doesn't know your leaguemates. Are they the type that will read the rules and think carefully about the impact, or are they the type that will download a generic cheatsheet and draft off of it with no regard for the league particulars? If the latter, then the Draft Dominator's adjustments will prove too steep. (Though these too-steep adjustments will incentivize you to get players who are more valuable in your league's setup, so that's not the end of the world).

Most importantly, every draft room is different. I could create two identical leagues and fill one with friends from the industry and the other with my extended family, and I guarantee you quarterbacks would be drafted earlier in the latter league than they were in the former. If you've played with your current leaguemates before, you probably have a pretty good idea of their general preferences; it's probably best to trust your gut instincts as to which positions will prove more highly valued.

Finally, here are a few more rules to help you scoop up as much falling value as you can.

Rule #1: If every faller is at the same position, you've probably made a mistake

If every time you are on the clock, the biggest ADP faller is a running back, that means you probably misjudged ADP. Those running backs aren't "fallers"; you just thought they would be hotter commodities than they wound up being.