Small Sample Sizes - Pros and Cons

sas4 responded to my blurb on Carlos Villanueva yesterday. For anyone that doesn't want to scroll down, here's my comment:

Villanueva's been in the right place at the right time, scarfing two relief wins for the Brewers. But his walk rate is way up (he's already walked 12 in 19 1/3, or one more than he walked in 53 2/3 last year) and his low HR rate won't keep up with a 0.75 G/F. I can't say avoid. I like Villanueva long term, and still think he could get a crack at the rotation. But Yovani Gallardo might very well get the next shot, not Villanueva. Major league managers love having guys like this in the pen.

One thing I should address is that I'm generally an advocate for not reading too much into small sample sizes. A difficult thing about a FAAB analysis, though, is that I'm often stuck writing about players with limited amounts of major league AB or IP, either this year or even through their entire careers. If Villanueva pitches another 60 innings this year and reverts to his 2006 walk rate, he will finish with 24 walks in about 80 IP, or a 2.7 BB rate per 9 IP. This rate would be very good bordering on excellent, and it's entirely possible that Villanueva's early start is a hiccup due to the small sample size.

On the other hand, it's worth keeping in mind that it is worth at least considering what that sample size might mean.

In Villaneuva's case, out of his 12 walks, eight of them came in three games this year. He's pitched 13 times this year and walked at least one batter in seven outings. Of those seven games, he's walked three batters twice, two batters once, and one batter on four occassions.

His first outing this year was a disaster where he walked three batters, got no one out and all three runs eventually came around to score.

However, Villanueva's other multi-walk outings were also his longest outings of the year. In both of these games he pitched four innings and faced 17 batters. He walked three in one outing against the Cubs and two in an outing against the Cardinals. One of the walks against the Cubs was an intentional pass to the immortal Daryle Ward. In both of these games Villanueva also won, so the walks in these instances didn't have an impact on his ability to pitch.

All in all, a slightly deeper (but admittedly still cursory) look at Villanueva does indicate that he doesn't have as much of a problem with walks this year as I would have thought. However, it is something that does bear watching.

To conclude on a more general note, remember that small sample sizes aren't necessarily always bad. Sometimes managers or coaches do react to a month's or a couple of week's worth of statistics in a negative manner. I don't doubt that Ryan Howard's aches and pains are real, but I suspect that if he were hitting .300 with 8 HR that Charlie Manuel wouldn't be resting him this week.

Regardless of whether or not you feel like a small sample size is useful or useless, you do have to look at them for cues as to whether or not a player will keep his job or turn a platoon role into a larger role. Chris Snelling and Jack Cust - who I profiled yesterday in my A.L. FAAB round-up - will be judged in the next couple of weeks based on a very small sample size of AB. It's not necessarily fair, but that's baseball. Managers have to make snap decisions and have to work with the tiny amount of results they have. They can't wait a year, sit through a performance review, discuss a performance improvement plan with the Human Resources department, and then fire their employees. Their bosses are demanding immediate results and, with even the minimum salary in baseball the equivalent to that of a senior vice president in a Fortune 500 company, have every right to make that demand.