Management by Baseball

In my management consulting, I find the most prevalent management weakness in thinking about big projects is what I call ‘binary thinking’.

Dot-bombs tended to spend money as though they could sell stock for $120/share, so no-one ever said "no" to any request about spending, even on stuff like Kona coffee beans at $40/lb. or hardwood paneling from endangered Hawaiian trees at 22 times the price of oak. But when things went bad enough to catch the attention of funders, no one ever said 'yes' to any request...no coffee, no sodas in the fridge, no administrative help anywhere outside the executive offices, no testers to test for bugs in software about to be sold to customers, and people had to buy their own pens and pencils (all these 'yes' and 'no' examples from my experience, btw).

Now, in the permafrost economy, ‘no’ is the prevalent answer, and while a lot of the things people said ‘yes’ to before needed ‘no’s, the pure binary yes/no, good/evil, black/white simplistic thinking is very shallow and costly.

Baseball has had a binge of this. For years, a favored strategy of team managers was the sacrifice bunt, a move where the offense just-about-surely surrenders an out in exchange to advance baserunners. In the deadball era from about 1902 to 1920, many games were 2-1 and 1-0 affairs, so yielding an out (that is, surrendering the chance for a big inning) in exchange for a scarce resource (a run) made good sense; if you can win a game 2-1, why worry about how best to score 5 runs in an inning?

Managers tended to use this strategy they’d learned themselves as players, even after games most often became 4-3 and 5-2 affairs. The bunt was just the way “things were done”.

When SABR started becoming influential and people with strong math backgrounds started attacking probabilities, some of its earliest studies were on the efficacy of the sacrifice bunt. Researchers like Pete Palmer 'proved’ the sacrifice was a net-negative approach to winning. And it seemed to be, if you added up and made composite all the cases then averaged. It’s standard in SABR now to ridicule the bunt as a strategy in all cases.

But every bunt has its own context: a hitter with some measure (good or bad) of hitting skill, the hitter’s bunting skill, the pitcher's effectiveness, the infield’s fielding ability, the score, the inning, etc. There are hundreds of basic contexts for a given bunt, and most of them will be net-negative and some will be net-positive.

Tim Tippett, developer of the most advanced baseball simulation I know, Diamond Mind Baseball, has done an informed study on the bunt in his Diamond Mind blog, called Sacrifice bunting revisited (Monday August 18 entry). His conclusion is a mathematically-informed equivalent to the common sense argument I made in the last paragraph: there aren’t just two outcomes to a sacrifice bunt (successful where batter is out and runners advance one base; unsuccessful where bunter fails, no one advances), there are many. And when you factor in all the possibilities, the sacrifice bunt, (for many good reasons you should read his article to understand) is a net-positive tactic in specific context, and because the act of having the sac bunt in your arsenal forces the opposition to sub-optimize by spreading resources over more possibilities (one of the key Allied invaders' tricks leading up to D-Day).

There’s a lot for organizations to learn from this. Saying ‘yes’ all the time, or ‘no’ all the time is certainly simple, especially for the weak-minded or lazy, but reality is finely-graduated with many levels along the continuum. Every decision has a context. Good managers keep that in mind.