An Effect of Shifting, or an Effect of April

Life is funny sometimes. While working on a post about an interesting little quirk in the 2015 increase in runs per game, Neil Weinberg tweeted out the following.

Fun early season blip in MLB. R/PA is up from last year, but wOBA is the same. Probably April, but fun to think about possible reasons.

— Neil Weinberg (@NeilWeinberg44) April 30, 2015

This was the exact quirk that I noticed that led me down the particular rabbit hole I’m about to drag you into. If you go to the league leaderboard, you’ll notice that the league as a whole is hitting .249/.314/.390 this year, good for a .310 woBA. That is almost an exact match for the .251/.314/.386 line the league put up last year, and by wOBA rounded to three decimal places, it results in the same .310 mark.

The fact that April 2015 offense is equal to total 2014 offense is interesting, because it suggests that offense might be ticking up this year, ending the trend of the last few years. April is almost always the most pitcher friendly month of the season, with cold weather knocking down baseballs that will generally fly out of ballparks in warmer months; additionally, most teams haven’t yet had to rely on their pitching depth yet, as the full toll of injuries manifests itself more later in the year than it does at the beginning of the season.

But that’s not the part that got me interested; it’s the part that Neil noticed. League batting is basically the same as it was last year, but league runs per game are up a decent amount, going from 4.07 R/G last year to 4.26 this year. That’s nearly a match for the league’s run scoring levels of 2011, back when league wOBA was .316. So offense is up a bit from a runs perspective, but not up much at all from an individual outcome perspective. What’s driving that difference?

As is usually the case when there’s an imbalance in run scoring relative to wOBA, the answer lies in context-specific performance. Just like a team can score more runs than you’d expect by bunching their hits together, the entirety of MLB has done just that so far this year. Here are league averages for the last 10 years with the bases empty and with runners on.

Empty	Season	PA	BB%	K%	BABIP	wOBA	wRC+
	2006	103246	7%	17%	0.298	0.328	94
	2007	102844	8%	18%	0.298	0.327	94
	2008	102838	8%	18%	0.297	0.324	94
	2009	102834	8%	19%	0.297	0.327	95
	2010	103039	8%	19%	0.293	0.316	93
	2011	103831	7%	19%	0.293	0.314	95
	2012	104403	7%	21%	0.295	0.310	93
	2013	104627	7%	21%	0.295	0.312	96
	2014	104252	7%	21%	0.295	0.305	93
	2015	13775	7%	21%	0.286	0.301	90
—–	—–	—–	—–	—–	—–	—–	—–
Men On	Season	PA	BB%	K%	BABIP	wOBA	wRC+
	2006	84825	10%	16%	0.306	0.336	99
	2007	85779	10%	16%	0.309	0.337	100
	2008	84793	10%	17%	0.303	0.334	99
	2009	84245	10%	17%	0.302	0.331	97
	2010	82514	10%	18%	0.302	0.327	100
	2011	81410	9%	18%	0.297	0.319	98
	2012	79776	9%	19%	0.299	0.322	100
	2013	80246	9%	19%	0.300	0.316	97
	2014	79675	9%	19%	0.303	0.317	101
	2015	10323	9%	19%	0.303	0.322	104

As everyone is aware, everyone hits better with men on base. The first baseman has to hold runners on, opening up a bit hole on the right side. The pitcher has to pitch out of the stretch instead of the wind-up. Worse pitchers put runners on than better pitchers, so the men-on-base split is skewed towards a lower quality of pitcher in the first place.

And, in a variable that is becoming more important now than it has been before, teams shift far more often with the bases empty than with runners on. If shifts harm offensive production — and they likely do, or else teams wouldn’t be rushing to adopt them with such vigor — and if shifts occur more with the bases empty than men on base, we’d expect to see a relative decline in offense compared to what we see with men on base. The historical BABIP difference between these two situations is 5-10 points; this year, it’s 17 points.

And while it is way too early to draw any conclusions, that is kind of what we see over the last few years. For most of the last decade, wRC+ with the bases empty was around 95, dipping down to 93 once and getting as high as 96 another time, but in the 95 range. Last year, it hit 93 again; this year, it’s at 90. Because we’re dealing with just a fraction of a season, we’re more likely to get significant deviations just due to random chance, so don’t just look at that 96-93-90 trend and conclude that the shift is crushing offense with the bases empty. Especially because BABIP was unchanged with the bases empty the last few years.

But that was always a little bit weird, because it does seem like shifts work to some degree, and the fact that league BABIP wasn’t moving even as teams became better at defensive positioning was not easily explained. The best explanation I heard was that perhaps there was a counterbalancing factor that was hiding the shifts impact on BABIP, so while the shift was driving BABIP down a bit, something else was driving it back up; maybe the increased velocity in the game leading to more hard contact.

But this year’s results are maybe more what we’d have expected to find. BABIP with the bases empty has cratered to .286, while BABIP with men on base is the same as it was last year, and right in line with historical norms. If shifting had a big impact, and was measurable at the league BABIP level, this is where we’d see it. And so far, this year, this is where we see it. Teams are getting about the same amount of hits as always, but they’re getting them with men on base instead of with the bases empty. Thus, runs are up while wOBA is the same.

Of course, the shift effect wouldn’t explain an increase in offense with men on base over the last few years, so this is all more likely to just be early season randomness than anything else. In a month, I’ll check back in, and this will probably have all regressed towards the mean. I wouldn’t wager on this effect continuing on in May.

But in April, it’s a thing that has happened, and not just to the Red Sox. Pitchers have been lousy at stranding runners, so runs are up even as offense is mostly unchanged. If MLB is happy about the higher levels of run scoring, they probably shouldn’t be congratulating themselves just yet; this likely won’t last. But if it does last, then maybe we’ll finally have some league-level evidence that shifting is suppressing offensive performance, rather than just highlighting individual players whose pull tendencies have made them worse.