The draft slot advantageby Michael Lerra
October 16, 2008
If you're playing in a league with strangers, you probably won't get to choose your draft slot. Most leagues have the default set to randomize the draft order, and follow through in the "snake" format. So in a 12-person draft, you could end up with picks 1, 24, 25, 48, and 49. Or 2, 23, 26, 47, 50, etc.
If you're drafting with your usual league though, you may have some choice into which draft slot you get. Many leagues will draw names out of a hat, and the first person whose name is drawn will get the first chance to pick their spot in the draft order.
Most fantasy players are familiar with the notion that there are "tiers" of players. Players next to each other in the preseason rankings may in fact have quite a difference in value. If the seventh through twelfth players in a draft are widely considered to be of similar value, then surely one would try to get the 12th pick in the draft if possible, because getting the 13th pick to lead off the second round is far better than the 18th pick.
But what if the elite players don't have clear tiers? Or if there is a clear top tier of three players, but those draft slots are already spoken for by the time your name is drawn? Is there an advantage to particular slots in the draft order?
In trying to answer these questions, I first turned to theory and mathematics. Your fantasy league pre-ranks each player, and for the most part they do a good job at identifying where players will get drafted. (This is partly due to good forecasting, and partly due to self-fulfilling prophecy). It's a pretty safe bet to expect players to get drafted in the neighborhood of their ranking.
So imagine a worst-case scenario. You're drafting against a bunch of folks who are letting Yahoo make their picks for them, simply based on rank. You want the guy who's ranked 83, but you're twelfth in the draft and you you get picks 61 and 84. In this case, you need to use the 61st pick on the 83rd best player (22 spots early!), because he won't last until your next pick.
A similar worst-case for someone drafting in the sixth position would see them eying the 90th-ranked player, while holding the 78th and 91st picks. In this case, they need to use the 78th pick, drafting the 90th-ranked player 12 spots early. But this is the worst case! The sixth drafter will never have to pick someone up more than 12 spots early, while the first-and twelfth-place drafters may need to make choices up to 22 spots too early! It's pretty clear there is an inefficiency.
So theory tells us that the early and late first round picks force players to make inefficient choices. What about practice? I have draft slots and season-long rankings from four leagues in which I've participated. From this data, I've eliminated my own four teams as well as any team that made less than 10 moves across the season, the reasoning being that they were not as committed as other players to optimizing their lineup and giving their best at winning.
I then coded each draft slot according to it's efficiency (lower is more efficient): First and twelfth received a code of "6", second and eleventh received a "5", all the way to sixth and seventh which received a "1". This way, the better draft slots would get lower numbers, and would expect better places in the final standings; higher numbers would expect worse places. I then correlated this code with the actual place finish of each team; the result was "r" = .07. This is a very, very weak correlation, probably not statistically significant, and caveats certainly apply, a small sample size of only four leagues being the biggest.
I then went back to test the importance of the original "tier" theory. We know the greatest spread in talent is at the top of the rankings; the difference between the second and third ranked players is far greater than the difference between the 149th and 150th. So theoretically, players who get the early round one picks would have an advantage. Using the same analysis data set as above, I correlated draft position with final team standing. The result was an "r" = .22. Substantially bigger, and interesting! The positive correlation here says that as draft position moves towards 12th, team standing moves towards 12thas well (and vice-versa, the first drafter is more likely to attain first place).
When put together, the two correlations paint a nice picture. The larger impact of draft order on team standing is that higher picks tend to place higher in the standings. In fact, squaring the correlation tells us that 4.8% of the variance in team standing is due to draft position. This supports the notion of tiers, as well as the importance of getting one of those best-of-the-best early round one superstars. Having a draft slot closer to the middle of each round explains just 0.49% of the variance in team standing is due to the benefit of being in the middle of the draft.
Bottom line: Try to get the highest draft spot you can, and take advantage of your tiers. But if there is a tiebreaker needed, try to get close to the sixth or seventh slot in the draft to help keep your picks efficient.
Mike is located in Cambridge, MA, working in the search engine marketing field, and waiting for fantasy baseball to get as big as poker so he can "go pro". Feedback on all pieces is welcome and appreciated!
<< Return to Article