Draft Strategy

Predicting MFL10 QB ADP

Some will say I’m a dreamer, but others will simply say I was right.

Last year, I proclaimed the death of late-round QB for MFL10 bestball drafts, writing in May:

Now, fantasy players have abandoned all caution, driven mad by the late-round success of last year’s quarterback crop. This year’s patience at the position is unprecedented. . . .

Quarterbacks are cheap, dirt cheap, and I believe that trend is here to stay.

At that point, quarterback ADP looked like this:

Late Round QB is the New Black

By the end of draft season, however, quarterback ADP looked like this:

2016 MFL10 QB ADP

Clearly, despite your attempts to persuade me otherwise, I was not right. I was right, however, in hedging, “Maybe my theory will look like hot garbage by August.” QB ADP rose dramatically throughout the summer, with late-round picks eventually going earlier in drafts than they had in the past two years.

Perhaps only in an effort to salvage my pride and my reputation, I want to highlight one prediction I got right. I wrote that recency bias was likely to “drive even casual fantasy owners to overrate their ability to find elite passers at the tail end of the draft, deflating QB prices throughout the draft season.” In fact, that was precisely what happened. Drafters went hard after QB16 and beyond, pushing their price higher than in recent years. Everyone thought they’d find the next Blake Bortles, Tyrod Taylor, or Kirk Cousins — all drafted outside the top-20 in 2015 — and they were willing to pay more to try.

Likely as a result of that late-round emphasis, early quarterbacks were drafted later than they had been before. Not only did drafters feel more confident in their late-round choices, but we likely all carried subconscious fears with us after the devastation the 2015 season imparted upon high-priced QBs.

This year, however, was not like 2015 for fantasy quarterbacks. Early-ish QBs carried the day. No quarterback drafted outside the top 153 picks gave you a win rate higher than 10 percent. Put differently, every quarterback drafted outside the first 13 rounds in 2016 lost you money, on average.1

2016 QB ADP and Win Rates

Let’s Predict the Future

This year, I expect recency bias will drive the quarterback train yet again in early MFL10s. After all, as RotoDoc explains, “ADP is largely driven by prior year results.”

Let’s put a spin on that idea and look at prior year results for the relationships between QB ADP and PPG. In other words, how well have we predicted QB performance each year?2

 Season Correlation Between QB ADP and PPG
2016 0.501
2015 0.0391
2014 0.4234
2013 0.4534
2012 0.575

I won’t mince words. We all sucked in 2015. To an unprecedented degree. The community’s preseason ranking of QBs bore nearly no resemblance to their final production, to a degree that was unprecedented in the modern era. We whiffed on some early QBs, and late-round heroes dominated the top-12 by season’s end.

And there is some reason to believe that failure affected how long we were willing to wait on QB in MFL10 drafts this past year. In 2016, QB1 stayed on the board for nearly twice as long as he did the prior two seasons.3

 Season Correlation First QB Taken Year N+1
2016 0.3898
2015 0.1039 41.64
2014 0.3743 21.8
2013 0.4534 22.68

From that small sample, we can hypothesize that the wild outlier 2015 season shook our confidence in early QBs and made many of us overconfident in our ability to spot late-round gems, resulting in a dramatic wait for the first QB off the board.

If 2015 begins to feel like an aberration inside a much larger sample — which it was — we can also predict that early QB ADP will rise back to more traditional levels given the relatively typical 2016 season we just had. Over at Fantasy Football Calculator, early QB ADP bears out that hypothesis:

2017 Overall ADP 2016 Overall ADP
QB1 18.5 34.5
QB2 26.9 38.7
QB3 33.9 53.2
QB4 43.7 58.3
QB5 47.5 61.1

My expectation is that 2017 MFL10 drafts will kick off looking much like 2014 and 2015 at the quarterback position, meaning we can expect two or three drafted in the first three rounds. Sadly, I doubt we’ll be stealing Drew Brees in the sixth or seventh round the way we did last year.

Subscribe to the best.

  1. For a full discussion, check out the superb work of Mike Beers on players and MFL10 win rates.  (back)
  2. These samples included all QBs drafted 1-32 each season.  (back)
  3. We only have reliable MFL10 ADP back to 2014, so I cannot draw out a longer historical trail.  (back)
By Joshua Lake | @LakeTwoQBs | Archive

No Comment

Leave a reply