View from the Basement: An Interview with SABR President Vince Gennaro
A few weeks back, I had the pleasure of attending an annual charity event in Boston entitled Saberseminar with the theme Sabermetrics, Scouting and the Science of Baseball. The weekend’s worth of presentations benefitted The Jimmy Fund and featured some of baseball’s top coaches, statisticians, scouts, doctors and scientists.
Though many of the sessions were SABR-conference worthy, the tenor was quite informal which added to the enjoyment.
One of the more traditional talks was presented by SABR President Vince Gennaro. The topic was Evaluating Pitcher Similarity. As a daily fantasy baseball player, I found the talk fascinating with some definite implication to daily fantasy baseball. As such, I asked Mr. Gennaro if he would answer a few questions pertaining to his research and he cordially obliged.
There are several studies that demonstrate a hitter’s history versus a pitcher is non-predictive but yet we hear of MLB managers making lineup decisions based on the data as well as broadcasters frequently referencing the numbers. At the recent Saberseminar in Boston, you presented some work that may be applicable to batter versus pitcher match-ups. Can you share the Cliff Note’s version with us?
I’ve always believed the career batter-pitcher matchup history is among the poorest uses of stats in baseball. In my opinion, it’s closer to baseball trivia than to baseball analytics. I’ve approached this important issue–one which GMs and managers face every day–from the standpoint of creating a predictive model that is grounded in the reality of the game. The framework that I am still fine-tuning is based on my hypothesis that how a batter and pitcher perform against each other at any point in time is a function of 5 major factors:
1. The style of the pitcher–how does his pitch repertoire, velocity, release point/arm slot, movement, pitch sequences, etc., match up against a given hitter’s strengths and weaknesses?
2. The style of the hitter–how does his swing plane, ability to handle an inside pitch, willingness to employ a 2-strike approach or hit to the opposite field, etc. match up against the pitcher’s strengths and weaknesses?
3. The quality of the pitcher–Clayton Kershaw and Drew Pomeranz profile very similarly in terms of “style”, however it is important to adjust our pitcher/batter expectations for the quality of the pitcher
4. The quality of the hitter–Some hitters systematically fare well against “top pitching”, while other hitters feast on the weakest pitchers. In other words, not everyone hits best, avg and worst pitchers proportionately. Yadier Molina handles top pitching exceptionally well, but hits only marginally better against the weakest pitchers. Josh Hamilton gets neutralized by top pitchers, but crushes weak pitching. This needs to be factored into the batter-pitcher matchup expectation.
5. The ballpark–the dimensions, altitude, prevailing winds, micro-climate, amount of foul territory can all have a profound effect on how a batter and pitcher perform, so this needs to be incorporated
The goal is to create a template that would reflect an expectation about any batter-pitcher matchup, tailored to each ballpark.
Down the line, do you feel your results may aid the daily fantasy player in finding favorable hitter versus pitcher matchups?
I could see this information being valuable to a fantasy player on multiple levels. First, giving the fantasy player a sense of which pitchers match up poorly against an opposing team’s lineup. Also, it would inform the daily player about individual matchup expectations. The caveat of course is that the outcome–the daily fantasy play–is by definition, a small sample size. However, it should provide an edge.
Other than the results of your research, what are a couple of other factors you’d recommend to identify a favorable hitter versus pitcher matchup?
Short of having my research in hand, I suggest a strong focus on ballpark as a factor. A medium fly ball to RF at Yankee Stadium may be a HR and at AT&T Park, it’s a routine fly out. Beyond that, I don’t think there are too many shortcuts that would create suitable proxies for my in-depth, detailed analysis I listed above. There are two other factors that I’d like to incorporate into my analysis, albeit down the road in a second phase: a fatigue factor and a streak factor. The fatigue factor would be to develop some expectation if a player (pitcher or hitter) is at the top (or bottom) of their game, based on some workload measure, leading up to the game in question. For example, after adjusting for other factors, I’d like to see how a hitter performs on the last leg of a 9-day road trip, when he’s playing a day game after a night game, and has had no days off. The “streak” factor is an interesting and controversial one and may be very difficult to quantify. I’m just speculating, but perhaps it means that the longer a player performs above (or below) his baseline level of performance, the more likely he is to “correct” (regress to his mean) in the upcoming games. Both of these questions may be relevant, but are clearly a second phase for me.
On behalf of the DFSEdge staff, I’d like to thank Mr. Gennaro for a few moments of his time. If you’re interested in following this and other interesting research topics, be sure to visit http://www.vincegennaro.com/ where you’ll find links to several other studies along with a link to Mr. Gennaro’s blog where you will find a more detailed discussion of his work on clustering pitches by similarity.
As readers of DFSEdge are no doubt aware, we de-emphasize batter versus match-ups when presenting our daily strong plays, instead focusing on handedness and venue. Even from the brief synopsis above, it should be apparent why I am very intrigued by the work and the possibilities of incorporating the results into a daily projection model. Mr. Gennaro said it best; the sample of projecting performance for one day is extremely small. All we’re looking for is an edge.
We also downplay streaks, largely emanating from research by our own Derek Carty. Mr. Gennaro cites streaks as a future research topic; it will be interesting to follow that and deduce if there are any actionable conclusions for the daily player.
Not that I needed another reason, but having someone with the resume of the President of SABR willing to share some of his valuable time for a daily fantasy baseball strategy site gets me a bit jacked. It’s obvious that the NFL and those that cover the NFL get it with respect to the symbiotic relationship fantasy football has with the NFL. Unfortunately, MLB doesn’t quite get it yet so to know the sabermetric community recognizes the relevance of fantasy baseball is a step in the right direction.
When Todd isn’t attending Saberseminar, he can be found at Mastersball where he is a Managing Partner in charge of the Platinum subscription content. Todd’s a veteran of Tout Wars and LABR as well a multiple time National Fantasy Baseball Championship league winner. Follow him on Twitter @ToddZola where he is very proud to have his very own trolls, though Jay Bruce’s second half struggles have rendered them conspicuously quiet.