Fantasy Baseball Snake Draft Strategy – Impact of Draft Order

The first piece of good or bad news for a fantasy baseball owner (at least in non-keeper, snake draft leagues) is their draft order spot.  Everyone has draft order preferences and they often vary by the year.  In some years where there was no overwhelming #1 pick, I preferred a middle pick so I did not have to wait a gazillion picks between turns.  I recall some drafts where I really liked the depth through pick 16 or so and wanted a late first round pick to grab two.  Last year, I wanted a top 2 pick because I felt Trout and Cabrera were clear #1/#2 and I did not want the agita of drafting Braunogenesis (of course I picked #3 and got Braun).

Anyway, I thought it might be an interesting exercise to estimate the value of draft order even though I understand this is typically randomized so this is less ‘strategy’ and more about ‘fate’.  This analysis is based on 12-team mixed leagues but I assume the same principles are in play for deeper league formats.

To estimate the impact of draft order, I took the 2013 End of Season Player Rater dollar value and matched the Player Rater rank with draft selection (e.g., the 1st draft pick is worth the value of the best player, 2nd draft pick worth the 2nd best player, etc.).  Below are the results (index of 103.1 means 3.1% better than average, index of 98.1 is 1.9% below average):

The above chart shows that the 1st pick is the most valuable and it goes downhill from there with #12 ever so slightly more valuable than #10/#11.  The driver behind why top picks beat bottom picks is because the differences in player value are greatest in the first round.  The best player (Miguel Cabrera) and 12th player (Hunter Pence) in 2013 were separated by $14 ($40 vs $26) where the 13th player (Cliff Lee) and 24th player (Jose Fernandez) were separated by $3.

The differences between adjacent picks (like #3 vs. #4)  will vary depending upon the value of the 1st round players but, in general, these differences should be negligible (e.g., in 2014, I’d estimate a minor gap at best between 1st and 2nd pick (Trout/Cabrera) but a larger gap than average between 2nd and 3rd).  Even the value of the best pick vs. the worst pick – based on this model – is only a 5% advantage.  That is nothing compared to what is seen at times in fantasy football when there are a small number of elite running backs.

But this model assumes that every team drafted the best possible player based on end of season stats (including pitchers) in the first round.  To figure out how this plays out in real leagues, let’s look at the results of our 64 Razzball Commenter Leagues in 2013.

Below are the average standings points per draft order spot in the 2013 Razzball Commenter Leagues:

This matches up pretty well with my estimated values per draft picks.  The top 4 draft order selections represent four of the five best draft selections and the bottom draft order selections were the least valuable.

The 4th pick proved slightly more valuable than the first 3 picks for one primary reason – Ryan Braun’s lack of availability.  He was only available in 2 of the 64 leagues by the 4th pick.  He was picked first in 17 leagues, second in 27 leagues, and third in 18 leagues, and fourth in 2 leagues.  The 4th pick ended up with less ‘black holes’ as Cano or McCutchen were taken 4th in 52 of the leagues and Trout lasted until the 4th in 5 leagues (the other 7 leagues had 3 Kemps, 2 Pujolses and 2 Brauns).

The slight dip at 5th/6th vs 7th/8th cannot be explained by the actual player selected in the 1st round.  I chalk this up to small sample sizes and that the 5th position is slightly more valuable than the 6th-8th draft positions.

The gap between most and least valuable draft order selection was 14% (107 vs. 93) versus the 5% seen in my estimates.  I think 14% is a more realistic gap between the most/least valuable draft order selections since my model mixed hitter+pitcher values while first round picks are almost exclusively hitters.  So while most of the late 1st round picks in 2013 finished close to predicted value (Fielder, J-Up, Tulo, CarGo all around $18-20), that’s still a $15-$20 gap from Trout/Cabrera and $5-$8 gap from Cano/McCutchen.

If we break down average team standings points by the actual player selected in the first round, the differences are greater since you are isolating the teams with great 1st round picks vs. those that drafted duds.

Below are the players who were drafted in the first round in at least 1/3 of Razzball Commenter Leagues.

Not surprisingly, Miggy and Trout owners fared best while Ryan Braun and Matt Kemp owners fared worst.  Just based on the first round selection alone, Miguel Cabrera drafters had a 22 point advantage over Kemp owners and a 21 point advantage over Braun drafters.

I wonder if Kemp hurt owners more than other injured players (like Pujols, TuLo, CarGo, Giancarlo) because 1) he did not perform that well when healthy and 2) his iffy prognosis meant you had to stash him on your roster all year vs. a Pujols who could be dropped once he went for surgery.

Conclusions

• Everything equal, you want the earliest pick possible in a snake draft.
• There are negligible differences between draft order spots that are close to one another.
• While higher draft order helps, it is who you pick that is most important (durr!).  Having the luck to avoid ending up with this year’s version of a Braun/Kemp is crucial for doing well.  (Maybe there’s a tiny bit of skill in here but I think it’s almost all luck.  Clearly players like CarGo and TuLo had known higher injury risk factors but delivered a lot more value than Kemp)