Calculating WAR from OOTP stats

[cnield]

Has anyone ever tried figuring out how many wins above replacement (WAR) their OOTP players are worth? I've been trying to figure out how to do this by reading about WAR, but I'm not sure how to go about getting fielding and baserunning approximations based on the data available in OOTP.

[actionjackson]

Originally Posted by cnield

Has anyone ever tried figuring out how many wins above replacement (WAR) their OOTP players are worth? I've been trying to figure out how to do this by reading about WAR, but I'm not sure how to go about getting fielding and baserunning approximations based on the data available in OOTP.

I don't think the stats that are necessary for that calculation are in the game, but it would be cool to see. Either that or all those stats plus WAR itself.

[tombair]

It may be doable for offense and pitching but I'm not sure about defense.

Here are a couple links that explain how to calculate it for position players and pitchers.

WAR Lords of the Diamond (PositionPlayers) - Beyond the Box Score

WAR Lords of the Diamond(Pitchers) - Beyond the Box Score

[cnield]

While I was working on this problem, I came up with a very rough approximation that I'm now using to make up "UZR".



These are, of course, complete judgment calls. I welcome any ideas about how to refine this table. As far as a rough approximation, they seem reasonable to me, but I'm guessing that some improvements could be made.

If any roster makers see this, how do you guys convert major league defensive stats into ratings for your roster sets? Do you do something similar, but in reverse?

[tombair]

I'm not an expert on UZR (or WAR for that matter) but that looks reasonable to me. UZR is an imperfect measure (a poor fielder could have a good UZR and vice versa) while range (and the other ratings) represents actual talent. But since the entire WAR rating is just an "estimate" (if that's the correct word), I think your table looks good.

I think it'd be great if OOTP 11 has some sort of total player rating (WAR or something similar).

[JWay]

Originally Posted by tombair

I'm not an expert on UZR (or WAR for that matter) but that looks reasonable to me. UZR is an imperfect measure (a poor fielder could have a good UZR and vice versa) while range (and the other ratings) represents actual talent. But since the entire WAR rating is just an "estimate" (if that's the correct word), I think your table looks good.

I think it'd be great if OOTP 11 has some sort of total player rating (WAR or something similar).

I've been asking for this for two versions and being on the BETA team doesn't make my voice much louder

[RonCo]

Given only the data in the game logs it isn't overly hard to calculate a respectable zone rating based on actual performance. I would expect that at least a zone rating should be in v11.

[cnield]

Originally Posted by tombair

I'm not an expert on UZR (or WAR for that matter) but that looks reasonable to me. UZR is an imperfect measure (a poor fielder could have a good UZR and vice versa) while range (and the other ratings) represents actual talent. But since the entire WAR rating is just an "estimate" (if that's the correct word), I think your table looks good.

I think it'd be great if OOTP 11 has some sort of total player rating (WAR or something similar).

Yeah, I hope so too. I made a chart of range vs range factor as calculated by the game for all the league's center fielders with more than 40 games played in center, and there was very little correlation between the two numbers. So those stats seem fairly flawed.

[RonCo]

Range factor is a stat of very poor quality for several reasons.

[RonCo]

I'm sure Markus could do something much better than this with less work than it took me. He has the raaw data, afterall. I had to crash the game logs.

http://www.ootpdevelopments.com/boar...lab-diary.html

[actionjackson]

I'd even be happy with WARP1 using RARP, FRAR, and PRAR to calculate it in the game if we have to stay with Baseball Prospectus, considering that we're using VORP and Will Carroll's Under The Knife. Anything is better than nothing. WAR is I think the better of the two, but a metric that factored all contributions a player makes into one number would be cool. It would be nice if the AI could think WARically or WARPically so we wouldn't have to expend so much hand-wringing in epic threads on Mark Belanger etc, but it sure was fun wasn't it? Of course we'd need a manual override so MD could make sure he was just a defensive replacement, whatever team he played for.

[Malleus Dei]

There's a relevant discussion about measuring defensive prowess by Bill James here:

Fielding Bible

[Malleus Dei]

Originally Posted by RonCo

I'm sure Markus could do something much better than this with less work than it took me.

Of course he could. But a bunch of us will have to hold his feet to the fire before actually he will do it.

[RonCo]

I'm sure there will be about 58 XI Feature threads.

[thewupk]

Originally Posted by cnield

While I was working on this problem, I came up with a very rough approximation that I'm now using to make up "UZR".



These are, of course, complete judgment calls. I welcome any ideas about how to refine this table. As far as a rough approximation, they seem reasonable to me, but I'm guessing that some improvements could be made.

If any roster makers see this, how do you guys convert major league defensive stats into ratings for your roster sets? Do you do something similar, but in reverse?

The only issue I have is that certain positions can't net a +25 UZR. It's very rare for a 1B to get above +10 in UZR on a given year. So I think you would have to scale each position separately from 0-100. A 100 in CF would net you around +30 (something say Andruw Jones would be busting out in his prime) and a 100 at 3B would likely net you around +18 or so which is what Zimmerman has done 2 of the last 3 years and he's widely regarded as the best defensive 3B in baseball.