Monday, January 5, 2015

Fantasy Focus: Resolved! Trust the Projections

I have always been a bit standoffish with projections.  They just seemed to be to focused on regression to the mean, so that all players end up looking similar in the end.  They are notoriously poor at predicting breakouts and are iffy on young players who don't have a track record yet.  But the projection systems have gotten more sophisticated over the years.  They now are able to take things like  injury history, minor league records and career trajectories into account.  Last year I did a comparison of projections for Giants players with their actual 2013 results and the correlation was uncannily good.

Last year, I drafted the fantasy team I wanted.  I did my own research and my own "eyeball" projections.  I participated in several mock drafts.  The mock drafts were very similar to my real draft.  I got the players I wanted at approximately the price I wanted.  There were very few, if any, surprises.   with just a couple of exceptions, the players I drafted performed up to expectations.  There were a few who were worse than I expected and some who were better.  The team did not perform well.  Maybe I wanted the wrong players!

After thinking it over this offseason, I have resolved to follow the projections in ranking my draft board.  I haven't decided to use ZIPS or STEAMER yet.  Probably doesn't really matter.  They get similar results.  Of course, if I have specific information about a player that goes against the projection, I will act on it, but I plan to basically rank my players according to the projections.

The next question is which metric to use.  One of the amazing things about baseball in general and fantasy baseball in particular is the difficulty in finding one statistic that tells you the value of a player.  In real life, WAR comes close, but WAR does some things that do not make it useful for most fantasy baseball leagues.  WAR accounts for defense.  Most fantasy leagues do not.  WAR adjusts for park and league factors.  Most fantasy leagues do not.  WAR highly values OBP.  Most standard 5X5 leagues do not.

In the standard 5X5 fantasy baseball scoring, the 5 offensive categories are BA, Runs, HR's, RBI's, SB's.  4 of those are counting stats.  3 are directly related to power, particularly HR's, since HR's are the only offensive event in which a single player both drives in a run and scores a run.  OBP is nice, but only indirectly impacts Runs and SB's with virtually no impact on BA, HR's or RBI's.  OPS is a popular single offensive stat, but it relies heavily on OBP which is of minimal benefit in standard fantasy formats.  I finally settled on SLG% as the best single stat to rank fantasy hitters with.  SLG% is essentially BA + IsoP which is what will have the most impact on at least 4 of the standard offensive categories.  You can use SB's to move players up and down the list, but as I pointed out in my last post, fantasy managers must more and more depend on "SB specialists" to win that category.  SLG% should tell you the most information about what players are likely to help you in the other 4 standard categories.

With that in mind, here are the top projected SLG% for hitters in 2015 using STEAMER projections:

1.  Giancarlo Stanton- .568
2.  Miguel Cabrera- .557
3.  Mike Trout- .538
4.  Jose Abreu- .530
5.  Troy Tulowitzki- .527
6.  Paul Goldschmidt- .527
7.  Edwin Encarnacion- .517
8.  Carlos Gonzalez- .505
9.  Andrew McCutchen- .504
10. Anthony Rizzo- .503
11. Kris Bryant- .496
12. David Ortiz- .496
13. Yasiel Puig- .495

I would probably downgrade Tulo and CarGo a bit for injury risk, CarGo more than Tulo.  I am a bit worried about Stanton as he would not be the first slugger to have his career take a dive after taking a pitch in the eye.  I don't know if I would have the guts to spend a first round draft pick on a rookie even if I knew they would be starting on the active roster out of spring training no matter how good they are, but Kris Bryant certainly deserves close attention.  I would move Trout and Cutch up a notch or two because of the added SB potential.  Otherwise, I think most fantasy managers are going to want to roster at least one of these 13 players on their team in 2015.


  1. I like all the new sabermetric (and what-not) stats, and I use them when I discuss baseball these days. But, honestly, I still think in ERA and slugging % to myself because, I'm a dinosaur and back in the days, those were the about the only stats we really had that could measure a players ability to help a team win... Now you a lexicon to figure out what people are talking about.

    And yet, here we are. It's 2015 and someone going back to the dinosaur stats... Makes me smile...

  2. I am more K vs BB, H/9 and HR/9 for pitchers and OBP, WAR, dWAR and K vs BB for hitting. Things like ERA only go so far because it depends on the park, defense behind you and strength of schedule. I would hate to be a pitcher for the Blue Jays and to play the Orioles, Yanks and Sox in those hitter parks. NL West doesnt sound so bad. I know WAR and dWAR have their flaws but I have spent countless hours sorting those 2 stats and they hit IMO more than they miss.

    1. But in fantasy baseball, unless a player is changing teams, you don't want to correct for the park or the team context because the player is going to be in the same lineup and playing in the same park. You WANT those inflated hitting stats that come with Coors Field! You WANT that suppressed ERA that comes with Petco Park, AT&T Park and whatever that mess in Oakland is called these days.

    2. Plus, if you use a good projection system, it already corrects for the BABIP and HR/FB ratios for you.

  3. Sorry. I meant that I look at those type of stats when evaluating players. I guess I overlooked the fantasy baseball thing. I havent played it for years.