Where traditional scouting and statistical analysis intersect?

Down in Fort Myers this past week, I had the opportunity to sit down with Twins Assistant GM, Rob Antony, and pick his brain about the organization’s use of statistical analysis. In his office in Hammond Stadium we covered a variety of topics with the din of the gathering crowd at the grandstand below. While a more detailed interview will be forthcoming on Monday, one portion of the chat was worth further examination.

Throughout the half-hour sit-down, I peppered in several questions from Doug Decatur’s GM IQ test from his book Behind-the-Scene Baseball. Decatur had worked as a statistical consultant for the Cincinnati Reds, Milwaukee Brewers, Chicago Cubs and Houston Astros and concocted a series of questions based on statistical thinking that all GMs (or aspiring GMs) should have some understand of. This was to gauge the Twins’ current comprehension of accepted statistical theories.

One such topic was runs batted in. Bill James wrote in his 1987 Baseball Abstract that RBI is “more subject to illusions of context.” Essentially, RBI are heavily contingent on hitters ahead of your 100-RBI guy to reach base regularly while a high slugging percentage is a product that is, theoretically, independent of other players. Using his quiz, I created a hypothetical scenario asking that if Antony were going to sign a free agent, would he go after the guy with a higher RBI total or slugging percentage?

Antony replied that he would prefer the player with the higher RBI total.  “Because you win with runs,” he said. “And I want that guy because you also have the correlation with a better batting average with runners in scoring position – he’s the guy that can step-up, the guy you want at the plate.”

Because Antony readily admitted that the organization had been reliant on traditional methods and had not dabbled too far into statistical analysis, the response was not at all surprising. At the same time, knowing that the statistical community commonly views RISP batting average as a small-sample size anomaly, I wanted him to elaborate further.

“I think guys are pitched differently when they have a chance to do damage and they can’t make adjustments. Then, sometimes the guy with a bunch of home runs and few RBI with nobody on base, they challenge him, and you look  and a lot of those guys do their production with the team behind and they tack it on and  enjoy a solo home run in the eighth inning.”

At the time of James’s publication, pitch data was not available to the general public and Antony’s contention throws a new wrinkle into the discussion. Do pitchers handle hitters differently with runners in scoring position?

According to Inside Edge data, hitters do witness a slight uptick in the amount of non-fastballs thrown while were tending to stay away from the strike zone once runners advance to second:

2009 MLB total:	Fastballs	Non-Fastballs	In-Zone
Non-RISP	65%	35%	52%
RISP	60%	40%	48%

 

How about more specifically? What barrage of pitches did Justin Morneau, who has 623 RBI since 2003, face once in RBI territory?

Morneau:	Fastballs	Non-Fastballs	In-Zone
Non-RISP	63%	37%	47%
RISP	59%	41%	38%

 

While the pitch type varies little, pitchers obviously respected Morneau’s power potential in 2009 and attempted to get him to swing at bad pitches. This had minimal effect as the Twins first baseman managed to produce a .908 OPS with RISP and a .866 OPS without RISP.

This is a steep contrast to someone like Delmon Young. After hitting in 93 runners in 2007 with the Devil Rays while procuring a rather pedestrian .408 slugging percentage, the Twins acquired him only to see his RBI total never reach his ’07 number. Reviewing his splits, you see that pitchers did not try to get too cute with the Twins’ left fielder:

D. Young:	Fastballs	Non-Fastballs	In-Zone
Non-RISP	59%	41%	49%
RISP	54%	46%	51%

 

Judging from the amount of pitches he sees in the strike zone under potential run-scoring conditions, a fair conclusion to draw is that pitchers have little fear of Young’s abilities.  Whereas Morneau contended with having to sort through over 60 percent of pitches thrown off the plate in order to find a good one, every other pitch to Young was in the vicinity of the strike zone.

Still, Young’s.425 slugging percentage in ’09 appears to validate that slugging percentage is a better statistical device than RBI. Was there anyone that actually fit Rob Antony’s profile for high slugging but a poor hitter under an RBI condition? Down south, Texas’s Michael Young personified that theory this past year. Young put up monster slash-numbers (.322/.374/.518) while jacking 22 home runs but drove in just 68 runners. Pitchers approached the Rangers’ third baseman with more non-fastballs and fewer strikes when runners had reached second and Young wilted:  

M. Young:	Fastballs	Non-Fastballs	In-Zone
Non-RISP	65%	35%	51%
RISP	52%	48%	45%

 

Under other circumstances, Young had a .959 OPS but that declined to .655 in scoring opportunities. Young’s ability to make contact dropped radically as well when pitchers bombarded him with off-speed offerings. Pitchers showed him different looks and he could not adjust with them.

While the opportunities afforded to a clean-up hitter and an eight-hitter differ as greatly as a graduate of Harvard and one of McDonald’s Hamburger University, there is truth that pitchers do handle hitters differently. As Antony noted, good hitters adjust. Even though I would maintain a preference for acquiring a hitter with a track record of better-than-average slugging percentages over the gaudy RBI guy, Antony’s preference for the opposite based on information from the traditional side of the fence does exist.  Once again, this is another area in which traditional scouting and statistical analysis can intersect.