Determining Correct Hitter/Pitcher Investment For 12 Team Mixed Leagues

Unlike with real baseball, it can unequivocally be said that fantasy baseball is 50% about hitting and 50% about pitching.  Yet it is close to a foregone conclusion that fantasy baseball drafters should invest disproportionately in hitters vs. pitchers.  If someone were to draft a pitcher in the first round or three in the first five rounds, the average fantasy baseball player would scoff at them (scoff I say!).

Why is this the case?  Here are some theories…

1. Hitters have more perceived (and maybe actual) reliability than starting pitchers.
2. The possibility that a hitter could provide value across 5 categories increases their perceived value vs. starting pitchers who could only contribute in 4 categories.  (Note:  Only 11 hitters were at least average in all 5 categories according to 12-team 2011 Point Shares (which factors position) – Kemp, Ellsbury, Braun, Bautista, Pujols, Pedroia, Upton, Votto, CarGo, Gordon, Francoeur)
3. Most fantasy baseball writers/experts primarily play deeper league formats like AL/NL-only where playing time becomes an additional variable favoring hitters.  (e.g., worst case, you can always fill a pitching spot with a middle reliever whereas there are a finite amount of hitters who get 3+ games a week).  This influenced early adopters of 12-team mixed leagues and had a snowball effect on ADPs.
4. A perception that starting pitchers are easier to a) pick up via free agency and/or b) draft in bulk or stream pitchers and play the match-ups.
5. It’s a clear preference of fantasy baseball participants who feel more confident in their ability to successfully draft late-round pitchers and pitchers off waivers.

For this post, I’m going to focus on point #1 but I might do a follow-up post to focus on #4 (which I think has some merit – PARTICULARLY the ‘b’ point).

I am also going to focus on the ESPN default roster format (C/1B/2B/SS/3B/MI/CI/UTIL/9 P) for this post but it should translate very well to two catcher leagues as well.  This is the first year I’ve estimated Point Shares for the standard Yahoo! format so I can only theorize what the best mix would be for that format.

Lastly, it is a lot easier describing this hitter/pitcher mix for auction leagues vs. snake drafts.  By converting draft picks into auction dollars, however, we can handle both draft formats at once.

So here’s the question I’m trying to solve:

For 12-team MLB leagues, do hitters provide more reliability than pitchers and, if so, what would be the optimal hitter/pitcher investment during the draft to exploit any opponent draft bias?

The most well-known hitter/pitcher ratio is $180 hitter /$80 pitcher which – in the ESPN format – represents about $26 shifted from pitchers to hitters (13/22*260=$153.63).  This tends to be the average in expert AL/NL-only leagues – for instance, the 2012 AL LABR ended up with a $182.50 average and $181.50 median hitter investment (note: that’s a 2 catcher league where the average 2nd catcher got $4).  FWIW, our 2011 NL-only LABR draft was a $179/$81 mix.

The closest ‘expert league’ format to the 12-team MLB format is the 15-team MLB format.  The 2011 Tout Wars mixed league averaged $178 for hitters so the $180/$80 ratio seems to hold true.

Interestingly, when I converted the current ADP for ESPN and Yahoo to auction dollars, both ended up with a 167/93 hitter/pitcher mix.  The 2011 Razzball Commenter League ADP came out to 178/82.  So, consciously or not, the average ESPN/Yahoo snake drafter invests more in pitchers than the average auction drafter.

To test the  optimal hitter/pitcher mix, I did the following:

• Took my preseason projected Point Shares for 2010 and 2011.
• Adjusted the auction dollar estimates based on the following mixes:  153/106 (even), 160/100, 165/95, 170/90, 175/85, 180/80, and 2011 Razzball Commenter League ADP.  Re-ranked based on those adjusted $ figures.
• Created a weight for every draft pick based on the $ estimates observed in preseason data (to avoid over-crediting top picks because no one could predict a season like 2011 Matt Kemp)
• Multiplied each weight against the estimated $ value for each player based on 2010/2011 end of year Point Shares.  Any player with negative value for the year – either because of injury or incompetence – was credited with a zero.  I tried two different weights – one proportional to the ranking (20.0 for #1, 19.9 for #2….0.1 for #200) and one that mirrors auction $ proportions ($40 for #1…$5 for #200)
• Added up these weighted $ values for the top 200 players.

For instance, let’s say Albert Pujols was the #1 pick, was worth $35 based on end of year Point Shares, and the 1st pick receives a weight of 20.0.  Player ‘Barely Rosterable’ was worth $2 and was the 200th pick in the draft which receives a 0.1 weight.  Pujols would be worth $700 ($35*20) to the end sum whereas ‘Barely Rosterable’ would only contribute 20 cents ($2*0.1).  This might seem extreme but it ensures that rankings that projects a valuable player earlier in the draft should be credited more and vice versa when non-valuable players are ranked earlier.

This test does not factor in real-world drafting – just testing how successful each of the hitter/pitcher-weighted rankings perform based on league parameters.

Here are the summarized results:

*ADP is based on Mock Draft Central for 2010 and Razzball Commenter Leagues for 2011.

While I’m not sure these results are statistically significant, the success of the even hitter/pitcher split (154/106) vs. the extreme hitter/pitcher splits (175-180 hitter/85-90 pitcher) at least directionally refutes the theory that hitters provide more reliability.

It is more difficult to determine the optimal draft strategy for a real-like 12-team MLB draft given draft room dynamics.  The biggest challenge with 154/106 is that you constantly have SPs at the top of your draft board.  Even if Justin Verlander warrants a 1st round pick, you will get less value than if you can get him in the 2nd round or a comparable pitcher in the 3rd round.  (Here’s a recent post I read that goes into more detail on the topic of market vs. performance value in fantasy baseball.).

I have found that for 12-14 team MLB leagues, a 165/95 split provides the best balance of market and performance value.  This works best in auction leagues as it delivers not only a strong pitching staff but the fairly conservative hitter $ values help provide discipline in early rounds when some drafters pay premiums for top hitters, resulting in strong hitter bargains later in drafts.  The above would also hold true for snake drafts where drafters skew heavily towards hitters.

For typical snake drafts – where ADP seems to be indicating a more moderate skew towards hitters – I think you can succeed with just about any hitter/pitcher mix – as long as you are getting value with each draft pick.  (I still prefer 165/95.)  If you’re going 180/80, you can’t then exceed your already inflated hitter auction values/rankings and you have to be really disciplined in your pitcher selections.  I’ve also found that some fantasy baseball players feel more confident in finding waiver wire value in certain positions.  I think this is fine to factor into how you weight your hitter/pitcher mix but realize that the positions that seem to have the most depth on the waiver wire (OF, SP, RP) are also the ones where you will face the most competition (since there are more roster spots for these players).  My preference is to find positions where there will be little waiver competition – 2B/SS/3B and C in one-catcher leagues seem to be the best.