We don’t draft as well as we think we do. As a group. I’m sure some of you are from Lake Wobegon and are above average drafters, but the rest of us aren’t that good. The next few charts show that.
If we really knew what we were doing, this is the graph we’d get when comparing Average Draft Position (ADP) to Fantasy Points per Game (FPG, PPR scoring):
The x-axis is the ADP rank of a player at a particular position: the player with the lowest ADP at the position has a rank of 1, the 2nd lowest ADP has a rank of 2, etc. The y-axis is the median rank of all players at that position since 2002 in FPG.
If we really were good at drafting, the players we picked as the best at that position in drafts would finish the year #1 in FPG at that position; the guy going off the boards 2nd at that position would be #2 in FPG, etc. The RB taken first would be the best RB that season all the way down to the 40th RB taken ranking 40th in FPG when all is said and done. Using FPG compensates some for injuries and using the median instead of the mean (average) FPG rank minimizes the outliers who would pull down the mean1. The black line or Baseline is really the All-Knowing Line if we had Biff’s almanac to draft by.
But stuff happens and we don’t match the All-Knowing Line. Here’s how we’ve really drafted RBs over the last 17 years:
Look how few red triangles are at or above the baseline. The thin red line shows the typical performance of RBs drafted #1 to #40 at the position – it’s pretty poor. Note too there is no particular pattern which says on this band of ADP we typically find sleepers who outperform expectations.
It’s worth mentioning the top players at the position have a very hard time meeting expectations as a group: there is no upside for the #1 pick in this method, and very little chance for the top several picks to beat their draft position.
WRs aren’t much different:
Here are the RBs and WRs on the same chart:
I eliminated the baseline here since it just clutters things up – the two things I want to emphasize are:
- It’s fairly random at any positional ADP rank whether the RB line or the WR line is higher, meaning that we can’t say WR x always outperforms RB x. If there were a run of several WRs from #20 to #25, for example, that ranked higher than RB20-25, we could hypothesize that WR20-25 were safer picks than their running back counterparts.
- The thin red and blue lines are basically coincident down to ADP rank of around 15, while later RBs typically rank higher than lower ranked WRs.
This last point doesn’t necessarily mean that we should stock up on late round RBs over WRs, because I’ve not analyzed the actual ADP of those players or their performance relative to each other. These players are being compared to others at their position – we don’t know if RB35 is being drafted 100th and WR35 is going at pick #140, or if RB40 in FPG is outscoring WR40 in points. I would think it’s probably true that late round RBs as a group out-perform late WRs, and this is a bit of support for that idea, but it doesn’t come close to proving it.
My main point is RBs and WRs rarely live up to their draft position as a group. If you’re in a draftmasters format, this means you need to plan for drafting failure by getting depth or redundancy in some part of your roster and think about where to take risks because you can’t be deep at every position. And in managed full-season leagues, you have to really manage your roster through waivers and trades because your draft is probably not going to be enough.
1For example, of the 17 RBs drafted as RB1 since 2002, if 13 ranked #1 and the other 4 ranked 154th in FPG, the mean (average) rank would #37. But the median rank would be #1. Using the mean over-weights the performance of 4 RBs while the median puts more emphasis on the 13 who did just what they were drafted to do.
