Thursday, March 7, 2013
The sacrifice bunt isn't dead yet
By David Schoenfield
Sabermetricians may have predicted the death of the sacrifice bunt, but it's still alive and well.
"There is a place for the sacrifice bunt and it's deep in your closet." -- Earl Weaver
Before sabermetrics, there was Earl Weaver, who managed the Orioles to five 100-win seasons and four American League pennants. Weaver believed in playing for the big inning, and thus his disdain of the sacrifice bunt become famous, as much a part of his personality as arguing with umpires and feuding with Jim Palmer.
His 1979 Orioles team that reached the World Series, for example, had 42 sacrifice bunts, tied for fewest in the AL, and 100 less than Gene Mauch's Twins. His last team in 1986 had just 33 sacrifice bunts, below the league average of 46. Weaver had actually evolved on the bunt through the year; his 1969 club had 74 sacrifice bunts, just one less than the league-leading Angels. In 1973, the first year of the designated hitter, the Orioles ranked fourth in the AL with 58 sac bunts. The Yankees, under Ralph Houk, had just 27. There were apparently Earl Weavers before Earl Weaver.
Weaver's theory on the sacrifice became one of the rallying cries of the early days of the sabermetric revolution, with managers who bunted (especially in the first inning) mocked by statistical analysts for still believing in old-school, one-run strategies. Why give the other team a free out? The numbers said playing for the bigger inning was the smarter move. But there isn't a revolution unless it invokes change. How much has the game changed, at least in this one area?
Sacrifice bunts have gone down, although maybe not as much you may think. I broke the totals into three-year periods (skipping the 1981 strike year):
1960-1962: One sacrifice bunt every 85 plate appearances
1970-1972: One sacrifice bunt every 84 plate appearances
1980-1983: One sacrifice bunt every 93 plate appearances
1990-1992: One sacrifice bunt every 99 plate appearances
2000-2002: One sacrifice bunt every 116 plate appearances
2010-2012: One sacrifice bunt every 118 plate appearances
Note that in the first two periods, there was no designated hitter, so you get more sacrifices simply because AL pitchers were still hitting. Even though Bill James started publishing his "Baseball Abstract" nationally in 1982, there wasn't a big change in sacrifice bunts until the offensive explosion of the mid-'90s. It wasn't sabermetrics that changed managerial philosophy so much as the game itself forced managers to forgo one-run strategies that are less valuable in a higher run environment.
Another way to examine this is by looking at No. 2 hitters. The old conventional wisdom was that your No. 2 batter was a guy who could bunt or move the leadoff hitter along if he got on base; you didn't necessarily put one of your best hitters there, rather a guy who fit a certain mold. And that mold was often detrimental to scoring the most possible runs. This became another facet for sabermetricians to mock managers over. We now know, that for the most part, Weaver was right; those 27 outs are precious. We know the optimal batting order would have your best hitter batting second or fourth, not third, and certainly that your No. 2 hitter should hit more than a singles hitter who can bunt well.
Now, sacrifice bunts by No. 2 hitters have slightly decreased. Again, three-year averages:
1960-1962: One sacrifice every 63 PAs
1970-1972: One sacrifice every 65 PAs
1980-1983: One sacrifice every 67 PAs
1990-1992: One sacrifice every 73 PAs
2000-2002: One sacrifice every 82 PAs
2010-2012: One sacrifice every 84 PAs
The overall number of sacrifices by No. 2 hitters has gone down slightly, but the ratio compared to total sacrifice bunts is relatively unchanged. No. 2 hitters are bunting a little less often only because managers in general bunt a little less often.
In fact, the whole conventional wisdom that managers used to bat a light-hitting "get the ball in play" guy second doesn't hold true either. For each period (again, skipping 1981), let's look at the OPS+ for No. 2 hitters. An OPS+ of 100 means No. 2 hitters across the majors were collectively league average for that year, above 100 means better than average, below 100 means worse than average.
No. 2 hitters today aren't any better than in years past. In fact, their relative production has decreased from 20 years ago. The revolution hasn't really reached this spot in the batting order. In fact, you arguably used to see more outside-the-box thinking in years past. Thanks to the Baseball-Reference Play Index, we can search to see who hit No. 2 in any given year. In 1984, for example, the collective OPS+ of No. 2 hitters was 108. Regular No. 2 hitters (200 or more PAs) that year included Dwight Evans, Ryne Sandberg, Carlton Fisk, Tony Gwynn, Alan Trammell, Dwayne Murphy and Gary Ward. In 1992, when the OPS+ was 109, we had Sandberg, Gwynn, Edgar Martinez, Tim Raines, Terry Pendleton, Lou Whitaker, Rafael Palmeiro, Roberto Alomar and Dean Palmer.
Yes, there were some bad No. 2 guys in those decades. The 1970s featured such light-hitting No. 2 guys as Larry Bowa, Jerry Remy, Denny Doyle, Tim Foli and Jack Brohamer. Bud Harrelson had over 1,000 PAs hitting second in the '70s and hit one home run. The '80s had Sandberg, Gwynn, Evans, Wade Boggs and Robin Yount, but also Rodney Scott, Jose Lind, Rafael Ramirez and Craig Reynolds.
But managers are still doing things like that. Dusty Baker gave Wilson Valdez 101 PAs at the two-hole last year. Buck Showalter kept J.J. Hardy there all season even there he finished with a .281 OBP hitting second. If anything, managers seem to use the No. 2 spot as a revolving door; only eight players received as many as 400 PAs hitting second compared to 16 hitting third. The No. 3 hitter gets his permanent spot in the lineup but anybody can apparently bat second.
It shouldn't be that way. The No. 2 spot is more important than that. In fact, Jayson Stark wrote yesterday that Davey Johnson thinks Bryce Harper may end up hitting third, which certainly speaks volumes about Harper, but the Nationals' best lineup would probably have Harper hitting second (and Jayson Werth hitting leadoff instead of Denard Span). But that's nitpicking. Span, Werth and Harper are all good hitters.
It's those managers who hit Wilson Valdez second -- who still think bunting in the first inning is a good idea -- that leaving analysts wondering when the revolution will get down to the field.