Column 1 2 3 4 or 5

If You Use Sacrifice Bunt, You're Sacrificing Percentages

By Kenneth Broder
Los Angeles Herald Examiner
May 8, 1988

When Gene Mauch retired as manage of the Angels this season, baseball lost its greatest practitioner of "little ball."

And I lost about half a dozen columns on the folly of his favorite weapons: the sacrifice bunt, the hit-and-run (or run-and-hit), hitting behind the runner and the stolen base.

The conventional wisdom venerates this approach as "percentage baseball," inferring that some statistical analysis has been used to shape the strategy. It hasn't. Except for a few years in the mid-'60s, this kind of play hasn't made sense since Babe Ruth pulled on a pair of cleats.

Mauch's kind of "percentage baseball" has probably done more to keep baseball stat freaks out of the ballpark and in the computer room than any other manager in history, although 20 years ago astute baseball men like Paul Richards were saying, "It has gotten to the point where they play you so well on the bunt that you can hardly afford to sacrifice any more. I just can't justify giving away a third of an inning trying to sacrifice."

Pete Palmer, co-author of "The Hidden Game of Baseball," decided to find out precisely what effect sacrificing an out to move up a runner has on the potential for scoring runs. But instead of hunkering down behind first base to watch muscle-bound Brian Downing push the ball 30 feet up the line, he ran computer simulations of every game played between 1900 and 1977.

First, he calculated the potential for scoring runs in any given situation (bases empty, one out; bases loaded, two outs, etc.) Then he calculated how these changing situations affected the odds of winning ballgames. It all gets rather complicated, as usual, but its usefulness becomes apparent as Palmer's co-author, John Thorn, walks you through the changing odds in a one-inning example.

A team entering the bottom of the seventh inning trailing by one run has a .343 win probability. The lead-off batter walks and the probability rises to .413. If a sacrifice fails, leaving one out and a man on first, the win probability drops to .348. If it succeeds and the man reaches second with one out, the win potential is .403. Incredibly, a successful bunt leaves you with a worse chance of winning the game. Hardly percentage baseball.

But don't tell that to Gene Mauch. His '87 Angels team led the league in sacrifice bunts (a typical Mauch characteristic) with 70, although that number was down 20 from the year before.

If Palmer and Thorn are on to something with their study of the sacrifice, it follows that managers like Gene Mauch may have paid a heavy price for their folly.

While the authors of the '88 Elias Baseball Analyst don't address Mauch's proclivities directly, they do take a stab at ranking the greatest managers of all-time. Gene Mauch doesn't make their Top 25 list, but then, their methods wouldn't make any statistician's Top 25 list of technique.

Essentially, they used three criteria: "A team's record in each of its previous seasons; its pattern of progress, for better or worse, during that period; and the volatility of the standings during the era in question." They decided Billy Martin was far and away the greatest manager ever. They feel he is personally worth 7.45 wins a season to a team, 5 wins a year to teams expected to win and a remarkable 12.6 wins a year to teams expected to be sub-.500.

Billy Southworth ranks second and Sparky Anderson is third. Whitey Herzog ranks eighth and Earl Weaver is 10th. Tom Lasorda didn't make it. The authors point out that Weaver would have ranked fourth if he hadn't attempted his ill-fated comeback a few years ago.

The Analyst study attempts to deal with the totality of a manager's performance: How much better or worse does he make his team? If hitters hit better for Mauch, or pitchers pitch better, or players perform better in the clutch--all that would be reflected in the Analyst rating. So would his choice of personnel, lineup and pitching rotation.

But let's narrow that focus, for the moment, to a manager's on-field generalship and two key measures of success: How well a team translates hitting into scoring and how often a team wins compared to how often it should win.

Mauch burst on the scene in 1962, his second full year in the league, when he took a Phillies team that had lost 107 games and turned it into a .500 club. And, in fact, he did perform some kind of minor miracle because the team was outscored that year, 759-705.

Item 3 on Bill James' list of Known Principles of Sabrmetrics states: "There is a predictable relationship between the number of runs a team scores, the number they allow, and the number of games that they will win." James figures the ratio between wins and losses is the ratio between the square of their runs scored and runs allowed (oh no, the dreaded Pythagorean Theorem).

Apply this method to Mauch's 1962 team and you end up with a predicted won-loss record of 75 and 86 instead of the 80-81 record it really had. Over one season, this could be attributed to any number of factors, luck being a prominent one. But over an entire career, it just might reflect how well a manager attains the victories his players earned. In Mauch's case it's a wash. In seasons he managed a full year, his teams played .490 ball, just what the run scoring ratio would indicate.

Of the seven managers I looked at, only Lasorda lost appreciably more games than one would expect, about 2.5 a year. His teams lost more than expected in seven out of 11 seasons. Why? That's hard to say. But it might have something to do with the fact his teams usually score less than they should.

Since we're using Bill James concepts here, let's use his Runs Created formula to determine how many runs Lasorda Team should put on the board:

Runs = (Hits + Walks) x (Total bases) / (At-Bats + Walks)

This formula is more accurate when stolen bases are included, but lacking caught-stealing stats this is the best I could do. Stolen bases are highly overrated, anyway (60 percent success is needed just to break even). Trust me. Or don't. While Brett Butler is stealing his way into your heart, he's running the Giants out of ballgames.

In Lasorda's 11 full seasons managing, his teams scored about 106 fewer runs than predicted. Since he's had some good running teams, you've got to figure not counting stolen bases cost him a few runs along the way. (Although it hasn't seemed to hurt Whitey Herzog's record much.) So maybe Lasorda's close to break-even. But check out his predecessor. Walter Alston, in a career spanning 23 full seasons, had teams that scored 385 more runs than would be expected. But they could only parlay that into a .001 advantage in expected won-loss record.

The findings on most of the managers looked at in this brief study tended to be inconclusive (they nearly matched projections in run scoring and wins) except for Mauch. His projected runs were off a whopping 405 in 22 full seasons.

Show me a team that couldn't use an extra 20 runs a season. Maybe he makes up for his "little ball" liability with expert handling of pitchers. Maybe? He sure has turned a lot of losing franchises into winners. But it does make you wonder if the stat freaks haven't got his number.

AUTHOR'S DISCLAIMER -- This is obviously not the last word on the merits of Mauch. But the author has a day job and a puppy to play with and can't possibly be expected to meet the maniacal standards of statistical devotion established by the true stat freaks. If you've dabbled at all in this sort of thing, have a few comments to make, or better yet, a study of your own to propagandize, send it in. This column can be your forum, too.