Management by Baseball

The Baltimore Orioles' front office uses a technique different from but parallel to that of the Oakland As' approach: find a way to use regression analysis of existing, available data to discover proprietary measures that might give your team a competitive advantage in personnel selection. The Orioles' technique is one you can use in your own organization if it has the courage to battle the fear-inspired resistance likely to emanate from H.R. and Legal.

According to this Baltimore Sun story, courtesy of Baseball Primer, the Orioles have exhumed an approach they pioneered (in baseball) back in the middle 1950s: using standardised tests of mental make-up. I haven't yet been able to find a copy of the test they are using, but according to the story, it's used across baseball, so all teams have access to the same data. The Orioles' difference is that Dave Ritterpusch, their director of baseball information systems, has taken enough interest in the tests to do the same kinds of things with the available data as Paul DePodesta has been doing with on-field performance data. That is, Ritterpusch has been running what appears from the story to be regression analysis, comparing early psychological test results to historical information about draftees and what kinf of success ended up achieving.

The article makes it sound like the team has created some formulae based on patterns in the data to create indicators they have found predictive. That is, they believe that for each position on the field, differentiating even starting pitchers from relievers, their test-result patterns that lead to greater probability of success.

It's a chancier strategy than the As', because the baseball events the As are using are in the external, easily measurable, world, while the Os are testing for interior tendencies and people are mutable. They don't ignore the other facets, just use this regression analysis as an additional perspective.

BEYOND BASEBALL

It's finding a niche and exploiting it. Depending on the quality of the test and the quality of the analysis, I believe this is wonderful technique.

I use a host of tools in my own practice for gauging personality and aptitudes. Some pop-psych profile systems are actually very useful when constrained to a work setting (that is, normal people at work exhibit a narrower and more consistent range of personality traits than they do in their "real" lives, so the profile systems don't need to be as complex or robust). I can recommend a couple of widely-available examples that are useful on the personality side if you write me.

In large organizations, there's usually resistance from H.R. and Legal, because they're more worried about getting sued than they are hiring and nurturing the right people and making sure they're in the right position -- that is, like bad baseball managers, they are more concerned with avoiding mistakes (not-losing) than they are with exploiting opportunities (winning).

Overcoming that barrier is important, and if you have the mojo to make more mental kinds of profiles part of your toolbox, you can greatly increase the depth of your organization's strength. And on a cautionary note, in the wrong hands, these kinds of profiles can lead unhealthy organizations to over-optimize on a small number of "types", rendering them less able to evolve or be dynamic.

Given that they are competing in the AL East Drillion-Dollar division, I think that's the least-likely challenge the Orioles are going to face with this interesting model.