Judging from my Twitter feed, Rickie Weeks has become a divisive figure as of late, with the Brewers’ fanbase split over whether he’s lost his usefulness as a player, or whether he just needs more plate appearances to come out of a slump mainly driven by bad luck. I delved into the numbers some to try to examine this more. Looking at his basic stats, it’s quickly apparent what the problem is; he’s batting .159. .159! Batting average is absolutely overrated as a stat, but here it’s very important to look at, because considering Weeks’ triple-slash is .159/.297/.297, it’s clear that the unacceptably low OBP and SLG are consequences of the low batting average. Both his ISO-D (OBP-AVG, a measure of patience) and his Power Factor (TB/H, a measure of power) are above both his career average and his career year of 2010. This means that if he were hitting his typical .270, he’d be posting a fantastic line.
But, of course, he’s not hitting his typical .270, and playing the “if” game won’t get us very far. So the important question to analyze is, “Why is Rickie Weeks only hitting .159?” Well, there are two reasons: he’s striking out more than usual, and his BABIP is way down. Those familiar with sabermetric thought know the latter isn’t much cause for concern, but the former can be. I will deal with each of these in turn.
As has been widely documented and discussed, BABIP can fluctuate wildly during the course of a season, and an abnormally low or high one is likely to regress to around the hitter’s career average. Rickie has average a .305 BABIP in his career, but right now it’s sitting at .212. That should go up, to say the least. Digging a little deeper, it seems one of the main drivers of his low BABIP is his 21.2% infield fly-ball percentage (IFFB%). This rate is the percentage of his fly balls that stay in the infield; clearly, this sort of ball in play is almost an automatic out, and an increase in these will easily lower BABIP. Is Weeks’ increased IFFB% a bad sign? Probably not. In 2007, baseball researcher Pizza Cutter (AKA Russell Carleton) did a study what number of plate appearances is required for various statistics to become reliable (h/t to Jonah Keri for linking that in his Grantland column.) He found that IFFB% takes 350 PAs to stabilize–as Rickie is only sitting at 165, I’m not particularly worried. I suspect much of this is due to how IFFB% is defined; it’s the percentage of fly balls that stay in the infield, so the sample size is the number of fly balls the batter has hit. For reference, in 165 PAs, Rickie has hit 33 fly balls. Clearly, the 33 is a much smaller sample than the 165.
So, here’s the big one, and the thing that fans seems to criticize Weeks for the most. Before I start talking about sample size and significance, a note. Rickie Weeks is a hitter that strikes out a lot, he’s always been a hitter that strikes out a lot, and he’s been very successful as a hitter that strikes out a lot. In 2010, when he was not only the best second baseman in the NL, but also a top 10 player in the NL, he struck out in 24.4% of his plate appearances, nearly one out of four. That’s the highest full season strikeout percentage of his career, in the best season of his career. Rickie Weeks can strike out a ton and still be a very productive player; however, there is such a thing as too much, even for him. If he had his career BABIP of .305 but his current strikeout rate right now, he’d only be batting .200. Rickie can strike out a lot, but not this much. So, again, the question is: does this strikeout rate portend bad things for Weeks’? This time, the answer is more complicated. Pizza Cutter’s study does find that strikeout rate stabilizes after 150 PAs, so Rickie having an elevated strikeout rate does cause some concern. At the same time, though, we have to consider questions of magnitude. Weeks’ career K% is 23%. Is his current 29.7% high enough to assume something has changed? To try to answer this, I decided to do a little number crunching of my own.
Pizza Cutter’s article is an empirical one, based on major league data. I come from a theoretical background, so I approached the issue slightly differently. Suppose you have a coin and you’re not sure if it’s a fair coin, or a double-headed coin. Say you think there’s a 5% chance it’s double-headed. If you flip 4 heads in a row, should you assume it’s double-headed? Well, there’s a 6.25% chance of tossing a fair coin 4 times and getting 4 heads, so it’s actually still more likely that it’s a fair coin. We can construct a similar experiment if we think of a plate appearance as a coin toss, where striking out is “getting heads”, and any other outcome is “getting tails”. We can assume the real strike out rate is x%, and find out what the range of strikeout percentages we’re 95% likely to get with a real rate of x% (95% is the standard commonly used in clinical trials). If we assume Rickie Weeks true strikeout percentage is his career rate of 23%, we’d expect him to end up between 17.0% and 29.1% 95% of the time after 165 PAs. If we use his 2010 K% of 24.4%, we get a 95% confidence interval from 18.8% to 30.3%. Currently, he’s at a 29.7% K%, so he’s basically right on the border of being outside his career norms. In other words, we’re just starting to get to the point where Rickie’s strikeouts are worrisome. Given that his plate discipline still appears to be strong, I suspect this K% will come down closer to his career norms. However, we are at a point where the numbers say it’s worth considering that there may be something wrong with his contact ability.