WAR calculation

[DODO]

So I've noticed that since OOTP has added WAR as a stat that it has seemed a bit off. What specifically stood out to me is that it seemed to underrate SS and C and overrate 2B and CF.

I just finished a season and tried to calculate WAR. This is what I did:

Offensive Runs=((wOBA-lgwOBA)/1.125)*PA
lgwOBA = league wOBA, had to figure that out myself. But used the given wOBA rather than calculating that and PA=plate appearences.

Defensive Runs = ZR
Given by the game

Positional Runs = Calculating WAR for Position Players | FanGraphs Sabermetrics Library
This is what's going to be off on my calculation. You cannot bring up a sortable sheet of IP per position, so instead I had to charge all of their defensive IP to the position they were currently listed as. So this wasn't completely accurate.

Replacement Runs = Calculating WAR for Position Players | FanGraphs Sabermetrics Library

So you have Offensive Runs + Defensive Runs + Positional Runs + Replacement Runs = WAR

The results, for my season, was that myWAR was higher than ootpWAR on average by 0.93 for 162 games played (So, if ootpWAR was 2, mine was 2.93).

And there definitely was a positional difference. Here was the average WAR difference based off position based off 162 games played (Positive: Mine was highers, Negative: OOTP was higher. Rounded to .05 for easy sake):

C:+1.85
1B:+.60
2B:+.55
3B:+1.00
SS:+1.25
LF:+.90
CF:+.70
RF:+.65

As we can see, that's quite a discrepancy, especially for catchers. If we want to look at what exactly the difference is between them just from a positional standpoint (calling .9 the average for easy sake):

C:+.95
1B:-.30
2B:-.35
3B:+.10
SS:+.35
LF:+.00
CF:-.20
RF:-.25

So the difference between C and 2B (the biggest discrepancy) over 162 games is 1.3 WAR, which is a pretty significant amount.

Also, I checked out pitcher war. I haven't seen anything about how to calculate a pitcher's position runs, so I decided to see roughly what OOTP was using for that. Based off using the Offensive Runs, Defensive Runs, and Replacement Runs, Pitchers needed a +40 positional run in order for myWAR to equal ootpWAR. So that seems quite a significant boost as well.

Obviously my way is not perfect due to the positional rating problem but it did support my hypothesis. I was curious if anyone working at OOTP could share just how they calculate WAR and, more specifically, their positional numbers.

Thanks

[Déjà Bru]

Dodo, I would not waste much time on this. WAR and VORP (and some other sabermetric statistics) are inherently imprecise and subject to variability as to method of calculation. The makers of this game use a proprietary set of formulas which may not be exactly what you have learned and use. Which is why I don't put much stock and faith in them, really. When I do pay attention to them, I use them as RELATIVES (comparing one player to another) rather than ABSOLUTES.

[RchW]

Yeah, since all SS and C are equally affected then it makes no difference when evaluating them. It may prevent SS and C from winning an MVP though. The game seems to favor highest WAR in that case.

[DODO]

Originally Posted by RchW

Yeah, since all SS and C are equally affected then it makes no difference when evaluating them. It may prevent SS and C from winning an MVP though. The game seems to favor highest WAR in that case.

MVP awards is what brought it to my attention actually, haha. But I also think it's important during free agency or extensions or anything in regards to contracts, really, when you're deciding between players at different positions. It does not make a difference when comparing same position players though, yes.

And I agree WAR is not absolute but it is the best "all in one stat" available. And there will be variations, but that's why I asked how they calculate it because I'm curious as to how it's different (Much like Baseball-Reference WAR is different than FanGraphs WAR, but one can easily look up what the differences are).