Sunday, September 21, 2014

La Russa Agile Insight #14 of 17: "Aliasing" Decisions in Zero-Sum Projects  

Good managers, in and Beyond Baseball, have to (what I call) "Alias" decisions and choose the right spots to autonomically pick the easy call while reserving their contemplative energy for the tough or crucial decisions. People managing Lean or Agile endeavors have to do this more deftly than more traditional ones, not less deftly as many believe. Baseball has great lessons for when to Alias and when not to

Aliasing is when you don't grind through a decision, meticulously measuring probabilities and possibilities. Instead, you count on the weight of history to automatically pick your option. Example: It's Chicago, it's August. You get up in the morning and get dressed for the day. Odds are it's going to be hot. Yes, you could check a weather report, or even three of them, but this is wasted ergs. You need to account for the very high probability of blistering heat in your choice of attire. On this day you have a key client coming into the office to meet with. You remember she always wears something green. That you should think about, perhaps make a choice to wear something green or think through why she might do that. Every millisecond you invest in processing weather prediction is a millisecond you're not investing in customizing the situation to better communicate with your client.

Organizations standardize commodity decisions (that is, alias) in non-zero sum choices (paper clip or staple? E-mail or Twitter broadcast? Discount or advertising?) and that makes sense. But insensibly, they tend to alias decisions in competitive or even zero-sum situations, where documented "standards" make it easy for competitors to predict their choices in advance, neutralizing and then creating a competitive deficit for the Aliaser.

In major league Baseball, the manager daily makes ~300 decisions in the ~5 hours of game preparation and the game itself. Given that volume over that amount of time, it's not only easy to alias decisions, it's a complete necessity. ┬┐What mollycoddled Fortune 500 CEO could make that many decisions in a week, not even trying to calculate their decision-load for a single work day -- and close to zero Fortune 500 organizations even have five days a year with the split-second, zero-sum decision environment a Baseball manager does?

Aliasing in and beyond Baseball is a necessity if you're going to be successful. But so is periodic contrarian self-examination, especially in an Agile environment or Baseball, where adapting to the exact situation on the ground is an unavoidable pre- req for success.

And that was a principle embraced by Hall of Fame-enshrined manager Tony La Russa, as documented in his most recent book, One Last Strike (p. 306-7). His narrative covers Game 5 of the 2011 National League championship series against the very tough Milwaukee Brewers.

Our bats continued to produce in Game 5, and as we'd done throughout most of the playoffs, we gave our starter an early cushion, with three runs ion the second inning. (Jaime) Garcia took that lead and ran with it pitching two more scoreless innings, and in the bottom of the fourth, when we were ahead 3-0, we got the first two men on base with no outs.

If you are knowledgeable about National League bottom-of-the-lineup tactics, skip to the next paragraph. If you're not, it will help to read this one. The Cards are at the bottom of their lineup, the weakest part in the general case, with the #8 batter being the player in the field that day with the weakest bat being followed by the pitcher at #9. And the autonomic decisions for the bottom of the line-up (both for the team at bat and the one in the field) propel certain patterns. At the early stage of a game, if there are no runners on base, there's going to be a low-probability of batting success at bat. With runners on base, the pitcher is going to (probably) bunt. This is close to automatic, which undermines its success probability because the team in the field adjusts (prevolves) to meet that challenge. And it also means the #8 batter, already likely to be the team's weakest not-pitcher batter, is going to degrade further because the pitcher and catcher, knowing #8 will strive to do something productive knowing the "useless" #9 is coming up next, will conspire to throw #8 borderline junk that is hard for any batter to succeed with. So, in the immortal words of the savage Old Testament diety and Glenn Frey, it's a losing proposition but one you can't refuse. But that's the standard operating procedure.

With the #8 batter and then the pitcher coming up next, the standard operating procedure -- #8 batter tries to make something happen, most likely without success and then have pitcher either try to bunt with one out or just try to make something happen, produces negligible returns. The aliased decision here is almost certain to underperform, but so is everything else. So most managers, in and Beyond Baseball, would just do the s.o.p. Not, in this game, La Russa.

Using some unusual strategy, I had position player Nick Punto (#8) try to move the runners over with a bunt, with (pitcher batting #9) Garcia on deck. Normally you wouldn't do that with the pitcher coming up next since most pitchers can't handle the bat as well as a position player. But I knew that Nick was an excellent bunter and that Jaime had a really good idea of what he's doing at the plate. In a way, I flip-flopped the usual roles of the eigth and ninth hitters. Punto got the bunt down and moved the runners to second and third base.

The thought behind this was that we needed to score in that inning, even if it was only one run. With just one run, we'd add to our lead and gain momentum. I thought Garcia could get the bat on the ball. I put the contact play on with Freese at third base. That meant when the hitter made contact with the pitch, Freese would head for home, forcing the drawn in infielders [note: they are drawn in because they are presuming the s.o.p. pitcher bunt] to make a play. Even if he was out, we would still have runner in scoring position (and the leadioff hitter, a most effective one coming up to bat).

La Russa's approach violated both standard operating procedure AND the blind faith of many math-centric sabermetricians that loves to point out the frailty of the s.o.p. The pure math neo-sabermetricians will point out that the probability of Punto succeeding lowers the Cards' probability of ultimately winning the game by about 2%, from 91% to about 89%, and if he fails in the sacrifice, the best the Cards can hope for is worse. The composite numbers argue against sacrifice bunts. But the event is not the composite average of all such events. There's a specific context. The move may work or fail, the additional run La Russa is aiming for will score or not in this case. But this even doesn't exist in a vacuum. The next time the Cards play the Brewers (or any team that has a scout at this game), the opposition will know that the #8 might bunt, and that will mean the opposition will need to invest energy in adapting tactics to meet the full toolbox of La Russa's approaches, diluting their ability to deal effectively with any one of them. La Russa may lose or win this one, but it incrementally increases his chances in subsequent ones, regardless.

This is more important in zero-sum situations, where one team's success requires another team's failure, and more and more in the arenas where Lean and Agile management approaches play a big role, zero-sum is the reality.

With Garcia hitting left-handed, the third baseman, although he was playing in for the possible squeeze bunt, was over towards the shortstop [note: another contextual advantage for the Cards] hole.Depending on how far the thrid baseman is away from the bag, the runner can get two to five extra steps closer to the plate. That's a really critical edge, especially with the contact play on, an edge you wouldn't have with a right-handed hitter. Garcia hit a ground ball to the shrtstopand Freese,with a good break, scored the all-important add-on run.

The Cards increased their chance for scoring one run and greatly diminished their probabilities of getting more than one run. Normally, net-negative, but in this context, a win. And further, in forcing zero-sum-competitors to increase their decision overhead many times in the future, a win earning continuing small yields.

Where to apply the La Russa non-Alias? Beyond baseball, I use my autonomic impulse to alias a decision to make a note to myself to examine it. I try to remember if the aliased decision has ever been something we could have achieved better results with if we'd done something different.

You can do this. If you use any kind of Kanban tool or even just index cards in your top right drawer (presuming you are a righty), document your aliased decisions. [Note: I usually use yellow -- for 'caution' -- cards for these aliased moves] You can invite your team to do the same thing. Reexamination is emotionally low-risk in the big picture sense, because team members rarely have personal investment in the standard procedures even if they are comfortable with them.

Baseball makes clear not only the value of Agile and Lean techniques, but it full of solid implementation ideas you can experiment with. You'll almost certainly never need the level of Agile skill even the below-average major league manager has, but if it works in that zero-sum crucible, it should almost certainly work in your own less competitive workplace.

This page is powered by Blogger. Isn't yours?

free website counter