from Amazon
Friday, December 03, 2010
Subject Matter Expert
I used to follow football alot (CFL more than NFL). And, I think if I were involved with the data analysis more, the reason for what I’m about to show you would have jumped out at me right away. I am positive that someone who doesn’t follow football would have been quite thrown by the data.
Here’s the data:
-0.01: average points following a short pass attempt
+0.45: average points following a long pass attempt
As soon as I saw this, the first thing that went through my mind: “How is that even possible. No, there’s a bias, or there’s bad data there.” That’s the way a critical thinker should think. And that’s why you rely on subject matter experts (SME) to see you through this data. And so, SME to the rescue:
It’s important to note some serious limitations of a direct comparison between deep and short passes. First, any pass attempt means that several things have already gone well. The pass protection held up long enough to allow a throw, and a receiver and a passing lane was judged to be open. Deeper pass attempts naturally require more time to develop, so the pass protection must last longer.
That is a big reason. The true classification of a deep pass should be based on “intent to throw deep pass prior to snap”, because prior to the snap, the defense “looks” the same for short and deep passes. (Well, not totally. Really, you would want an extra parameter that says: “defense looking for deep pass, short pass, inside run, outside run"… something like that.) If you don’t control for all the parameters, the data is biased.
This is just a general call out to all the data recording outfits to record everything you see, regardless of how subjective. Please. Without it, we have to infer what may have happened. Or worse yet, your academics who are not SME will presume there is no bias in the data, and that “all other things equal”.
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date