Park Factors Updated
For all our Wins Above Replacement metrics (WAR) we use five year regressed park factors. These are now updated to include 2010 data.
Our final 2010 UZR update will be later this week, but as of now UZR has been updated through September 30th and is only missing the last three days of the season.
David Appelman is the creator of FanGraphs.
Where can we see the park factor values?
+1, I would love to be able to see them.
Question: When to park factors enter the WAR equation? Is wOBA park adjusted already or does this happen later?
wOBA is not park adjusted; the “Batting” under the “Value” section is, however (unless I’m mistaken).
Park factors also come into play for pitching:
http://www.fangraphs.com/blogs/index.php/pitcher-win-values-explained-part-six
I noticed Lee and Doc both slipped a little with the update. Dave, did you take note of any significant changes to the WAR leaderboards? I recall when B-R updated theirs a few weeks ago, Sabathia leapfrogged King Felix.
There were really no significant changes because we use 5 year regressed and one extra year isn’t going to make a whole lot of difference. At the very most a player may have changed 3 runs with rounding (some Yankee players since the park is new and the home/away numbers were considerably different in the two years), but that’s about it.
David – how do you handle new parks (like NY on Minn)?
Do you just regress it with previous years of the old park to get the 5year data set? If a new park is substantially different do you go back and update old WAR data once a 5 year data set can be obtained on a new park?
To echo others, can you provide a table with the park factors you have calculated?
Patriot gives a table of park factors on his site, so you can go there. I’m also interested about Minnesota. What park factors were used before this update and how did it change the numbers afterward?
The new Twins park was set at 100 at the beginning of the season and it ended up as 99 this year.
New parks (or parks with significant changes) start out fresh at one year of data and are regressed 40% in year 1, 30% in year 2, 20% in year 3 and then 10% for years 4 and 5.
New park data is never mixed with old park data.
As for a table with park factors, I’ll try and give that a permanent spot this off-season.
Thanks for your response.
Have the UZR and +/- final updates happened yet? Thanks