Have been considering this idea a bit more.
Some further thoughts are;
^the idea has merit, but in reality is probably highly time and data intensive, with limited data on directors actually available
^to achieve assessable data, one would probably need to assess all the easily accessed company fundamental and financial data fields in conjunction with director attributes
^in effect, one would need a huge kickarse dataset with some kick arse algos to sort the chaff from the wheat
^the final result? probably iffy results. Human teams are involved. Corruption, greed, sabotage, manipulation etc etc could all distort outcomes and are unmeasurable parameters
^it could be done, at great cost, time, effort etc, perhaps individual appointments checking on proposed investee companies might be the easiest continued norm.
I wish there was a "duds and avoid" list of directors...