from Amazon
Sunday, August 07, 2011
“Trust me! This uber-stat is great!”
Dean Oliver comes with a background that is unmatched among most, if not all, analysts in any sport. Dean, under the auspices of ESPN, released a Total QB Rating, one that would supplant the current NFL Passer Rating, of which we all know has some problems.
And there is the issue. Suppose the NFL, to much great fanfare, rolled out a “product” called the Passer Rating, and gave their top 10 list. They even wrote an article to support it, giving just enough details to tell you what it was about, but not tell you how to calculate it yourself. What you end up with is not enough to satisfy the analyst and not enough to satisfy the fan. Without convincing anyone, who can you actually sell this to? And that’s what ESPN did here. The commenters in that thread, and there is almost 1000 posts, seem to feel mostly the same way.
Maybe there was a time where you could just trust Dean to do it right, and maybe there was a time where you could just trust ESPN to do it right. That time no longer exists, if it ever existed.
A great way to introduce a black box metric is the way MGL has done it in the past with the early versions of UZR and Super LWTS. Here you can see the idea behind UZR, as well as adjustment tables. There’s enough information there that if you wanted to recreate UZR, you’d get pretty darn close.
So, that’s my call out to Dean, and any other black box analyst. Don’t make it part of the parameters to introduce a black box stat as “trust me”. Do what MGL did with UZR, and gives us a substantial part of the details.
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date