from Amazon
Monday, May 12, 2008
Quick Park Factors
This idea stemmed from the original Historical Abstract. Here’s what I wrote to David at Fangraphs, and I’ll provide further commentary:
I believe that you have adjusted the WE based on the run environment for the year, correct? Is that league/year or just year? And do you have plans to go to park at some point?
A quick stand-in for park is “team”, in that you simply look at the number of runs scored in that park, regardless of how many runs they scored on the road. The implication here is that the offense/defense split will be exactly 50/50, regardless of how good or bad a team’s hitting or pitching is. In one sense, it’s “ok”, since the universe is that park, and as far as the competitors are concerned, those hitters playing are the only ones in their universe. And of course, in another sense, it’s not ok for the obvious reasons. However, it is super quick to figure out the number of runs scored in a team’s park per 27 outs, each year. This version of a park factor might make it better than no park factor at all. I hate seeing the older Coors hitters getting too much of a boost.
So, this is how it works. Imagine you have a team, let’s call them Nos Amours that is equal in its offense and defense. The league scores 4.5 runs per 27 outs. If at Nos Amours home park (let’s call it Big Owe) there are 5.0 runs per 27 outs scored, then we know, for sure, that this is a hitter’s park. How did I figure that out, without knowing how many runs are scored on the road? Well, I said that the team had a defense that is equal to the offense. If there are 5.0 RPG at Big Owe, that would mean that the team scores 5 and allows 5, or scores 6 and allows 4. What would they do on the road? They’ll score 4.5 and allow 4.5, naturally (or a bit less, since Big Owe is not on the road, and the overall league average is 4.5).
But, what if you had a team that had a dismal offense and fantastic defense? And at Big Owe, there are 5.0 RPG? This would therefore be a hugely hitter-friendly park. If they score 3 and allow 3 in a road park where the average team scores 4.5 RPG, and at Big Owe they score and allow 5 runs, that’s a huge home park advantage.
HOWEVER, by using the “team” park-based approach, regardless of the team makeup, Big Owe is a 5.0 RPG park in a league of 4.5. So, a team that is really a 4.5 offense, 4.5 defense on the road is a 5.0 / 5.0 at Big Owe. And a team that is 3.0 offense and 3.0 defense on the road is evaluated at Big Owe as if 5.0 RPG is the norm. And therefore, the offense and defense will each get equal credit, even if we know better.
I hope that made sense.
Maybe one of you guys can give the one-line summary of this.
Anyway, this approach at least gives you a halfway decent park factor, and as mgl is fond of saying, some park adjustment is better than no park adjustment.