from Amazon
Wednesday, September 15, 2010
What park factors to use for awards…
There was some discussion on BBTF about what park factors to use if you want to adjust CC and Felix’ stats for the CYA.
Some prominent posters over there think that one year, 2010, park factors (how the park “played” this year) are the ones to use, since you are not trying to figure true talent or performance going forward.
While the latter is true, I strongly disagree - with one qualification. That qualification is that a one-year park factor does include, to some extent, weather conditions for that year (and the schedule and opposing players I guess). However, I think that is a small consideration, and can easily be accounted for by weighting each year in computing multi-year PF’s.
Ignoring that (the weather and opponent issues), the one-year park factor is irrelevant in determining how a pitcher (or batter) performed for that year - we only care about a park’s true park factor (for that year). That is true for an award and not just for a true talent estimate or a projection.
For example, let’s say that the “actual” park factor for 2010 in Safeco was 1.00, which is entirely possible, since we are only dealing with runs scored and 81/81 games (off the top of my head, I would say there is a 15% chance of that occurring by chance for a park that has a true PF of near .90). And let’s assume that the weather had nothing to do with that - it was simply a fluke, and as I said, not an unusual one at that. Let’s also assume that true PF for Safeco is .92.
If we want to determine how well Felix actually pitched at Safeco, what do we use? I contend we use the .92 and I am pretty sure that is correct. Why do we care that the runs scored at home when Felix wasn’t pitching was exactly equal to the runs scored in other parks, in 2010? We don’t. And of course if Felix’ own Safeco PF (only when he pitched) was 1.00 also, the sample is so small as to be practically meaningless.
BTW, that is the same mistake that many people (including prominent analysts) make when adjusting team records for strength of schedule. To do that, you use the estimated true talent of opponents, and not their actual records. I can easily prove that using actual records is wrong. Say you have a two-team league and one team is 100-62 and the other is 62-100 (of course). If you adjust each team’s record for their strength of schedule using the actual records of each team, guess what you get? You get each team being a .500 team. That is true regardless of each team’s record in a 2-team league!
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date