Collette Calls: Monitoring the Small Sample

Wednesday afternoon, Peter Gammons tweeted out the following tidbit about Justin Masterson:

Justin Masterson's change in 2-seam and 4-seam usage w/ backdoor and back foot sliders why LHHs are .146/.222/.195/.417

— Peter Gammons (@pgammo) April 17, 2013

Shortly thereafter, Ken Rosenthal tweeted out something very similar to which I replied that Masterson had been using his slider more frequently to left-handed hitters than he had historically done according to his pitcher card at BrooksBaseball.net. This then sparked a civil discussion back and forth with Joe Sheehan about sample size.

We know from the work that Russell Carleton, now of Baseball Prospectus, did years ago just how much of a sample size is needed for a stat to become meaningful. Both Sheehan and myself are rather familiar with that research and we both preach it frequently, even if one of us was singing the praises of Bryan LaHair 114 plate appearances into last season only to watch him have a .638 OPS from that point forward.

My point in the discussion was all change has a starting point. If a pitcher's strikeout rate needs 150 batters faced to become meaningful, it does not mean we ignore plate appearances 1 to 149. Rather, it means we watch the player's process during that span to see if something has changed or if the improved numbers are another example of statistical randomness. Last season, Fernando Rodney faced just 42 batters in April but had seven saves and a 0.87 ERA.

In fantasy land, Rodney is being picked up in 95 percent of leagues after his second save because people are just that desperate for saves. In real baseball, there was quite a bit of skepticism in those numbers because it was hard to forget the many years of Rodney being a bad pitcher. Yet, in hindsight, April served as a starting point for the changes Rodney made in his delivery and his position on the rubber that led to better short-term results that eventually became better long-term results. Discussing, tweeting, or writing about such changes to deliveries, positioning of hands and feet, or swings is not confirmation bias of statistical recency as much as it is looking for starting points for change.

Masterson has historically been a poor pitcher against left-handed hitters so if he were to become even league average against them, it would be a big deal as he is already good against same-handed hitters. If I were to come out and tell you that Masterson's increased slider was going to allow him to do that in 2013, I would be no better than the snake oil salesmen of yesteryear as we cannot be predictive based on data from three starts including one against a very poor Rays team. What I can, and should be doing is pointing out changes in hitters and pitchers as something to watch for. Pitchers adding new pitches can help them find short-term success. Hitters making changes to their swings can help them get better results. In fantasy baseball, we do not have the luxury of waiting until a pitcher faces 150 batters or a hitter getting 300 plate appearances before making a judgement call. We play on a week to week and sometimes daily basis in making the judgment calls on whether a player's recent performance is real or fake. Take for example Hisashi Iwakuma.

Iwakuma spent the first part of 2012 as a reliever before moving into the rotation on July 2nd. He pitched effectively in the second half and Paul Sporer sold me on him this offseason during our podcast chats. I ended up with Iwakuma in four different leagues including both of my expert teams. Iwakuma pitched six shutout innings yesterday going toe to toe with Justin Verlander before exiting with a blister issue. With that start, Iwakuma now has thrown 20 starts as a starting pitcher and has a 2.44 ERA, a 1.08 WHIP, a 20 percent strikeout rate, and a six percent walk rate. He has faced 483 batters as a starting pitcher so we have a significant body of work to review.

Iwakuma's strikeout rate is slightly better than the league average for starting pitchers and his walk rate is better than most starting pitchers. The ratios are also rather strong but the caution flags are also present. The flyball pitcher has a 1.1 HR/9, and has put up a .258 BABIP while stranding 85 percent of his baserunners. Those two rates are the reason why his 2.44 ERA is much lower than his 3.96 FIP. In reviewing his player card, we see that Iwakuma is using his fastball seven percentage points more than he did last season.

Iwakuma's 2013 differs from last season in a few ways. Last season, he had a 53 percent groundball rate yet so far in 2013, just 39 percent of his batted balls in play have been groundballs while his flyball rate has gone from 30 percent to 44 percent. His overall pitch usage has changed in that he has thrown his fastball six percentage points more frequently than he did last season and is throwing a higher percentage of strikes in 2013. Last season, he would throw a lot of chase pitches with this breaking ball and splitter as he threw pitches in the zone just 52 percent of the time. This season, Iwakuma has found the strike zone 58 percent of the time while striking out the same percentage of hitters as he did last season but walking much fewer to date.

Is this a pitcher who has more confidence in his fastball? Maybe, or it could be an influence of the catchers he now works with. Jesus Montero is terrible behind the plate so perhaps Iwakuma lacks the confidence to bury his slider and splitter too much as he fears Montero will not be able to block it. He also throws to Kelly Shoppach who is a proponent of establishing a pitcher's fastball and forcing the issue with hitters as he did in helping James Shields turn things around in 2011. The early-season numbers for Iwakuma are strong, but in a small sample size. Yet, if we stretch back into last season, we see we have a very serviceable pitcher that has been a bit fortunate so picking him up now may help in strikeouts and WHIP, but the ERA could be in for some trouble as his strand rate and BABIP regress even just a few points.