Argh! I thought this was done, but the fat lady got laryngitis…
These are two papers. There are 3 things to say about them:
1. They are principally the former “Monkeys” paper split into 2. It was two long, and took too long to get to the “Estimator,” which I see as the principal contribution of it. So “mWAR” is the estimator part and “Weaver” the demonstration of the invalidity of a Pythagorean-expectation-based estimator part.
2. The “mWAR” paper does involve one significant change, however. It jettisons the “Jamesian Forecast” component of the former “Hybrid Estimator.” On reflection, I concluded that the JF component was likely invalid and added too little to matter anyway. If you read the last version of the paper, you’ll see that the estimates of the latent skill levels of the analyzed managers (all 512) are very similar to what they were under the old Hybrid version of what is now the mWAR estimator.
Now I swear I’m moving on (or maybe back; there are some old projects that need updating; there actually is no fat lady when one is trying to make sense of empirical evidence).
