"Replacement level" isn't just theoretical

[The Wolf]

The A?s and Brewers Define Replacement Level | FanGraphs Baseball

[Déjà Bru]

Yes, I hate to say it (again) but the indistinctness of the calculation is its chief drawback to me. When I see WAR in the news (the baseball kind, not the [insert current hotspot location here] variety), what is the methodology and is that the one which is being employed in this game? I cannot take such things on faith alone. I must be able to know, understand, and even calculate the statistic myself or it has little meaning to me.

[The Wolf]

Originally Posted by Déjà Bru

Yes, I hate to say it (again) but the indistinctness of the calculation is its chief drawback to me. When I see WAR in the news (the baseball kind, not the [insert current hotspot location here] variety), what is the methodology and is that the one which is being employed in this game? I cannot take such things on faith alone. I must be able to know, understand, and even calculate the statistic myself or it has little meaning to me.

So learn that. It's easy.

[Déjà Bru]

Originally Posted by The Wolf

So learn that. It's easy.

Learn which, Jim? Sure, somebody might do a good job of explaining it, their version. I'm only interested in this one, the OOTPB version. Can you or someone else explain exactly how it is calculated? Especially the "R" part? And even so, can we feel comfortable that when this game says Jose Bautista historically has a WAR of "X" and ESPN says he has a WAR of "Y," we are talking about all oranges?

But there I go again being a:

• Village idiot
• Dinosaur
• Other

Hey, maybe this should be a poll?

[Fyrestorm3]

Originally Posted by Déjà Bru

Learn which, Jim? Sure, somebody might do a good job of explaining it, their version. I'm only interested in this one, the OOTPB version. Can you or someone else explain exactly how it is calculated? Especially the "R" part? And even so, can we feel comfortable that when this game says Jose Bautista historically has a WAR of "X" and ESPN says he has a WAR of "Y," we are talking about all oranges?

This is the issue that I have with WAR. I like the stat, and I think it could be a very useful one. I personally can't really wrap my head around the idea of a "replacement player," though that hasn't stopped me from using it. And, in fact, I find that it's great to use in OOTP, where you can be absolutely certain that every player in the game has been put through the same algorithm to determine their WAR. But in real life, I'm iffy on it, simply because there's multiple ways of calculating it. That right there suggests that the calculations leave far too much room for error to use it as anything more than an ancillary stat. If it could be narrowed down to a single formula across the board, I'd be much more okay with its use on the whole.

[Déjà Bru]

Originally Posted by Fyrestorm3

This is the issue that I have with WAR. I like the stat, and I think it could be a very useful one. I personally can't really wrap my head around the idea of a "replacement player," though that hasn't stopped me from using it. And, in fact, I find that it's great to use in OOTP, where you can be absolutely certain that every player in the game has been put through the same algorithm to determine their WAR. But in real life, I'm iffy on it, simply because there's multiple ways of calculating it. That right there suggests that the calculations leave far too much room for error to use it as anything more than an ancillary stat. If it could be narrowed down to a single formula across the board, I'd be much more okay with its use on the whole.

Yes, the benefit of this seems to be in the comparison of players, relative to each other. If you trust the maker of this particular WAR stat, and I do, then it becomes useful to some folks in that fashion, I suppose. It just does not sit well with me because it is . . . [drum roll] . . . arbitrary in the sense that you are just supposed to accept it without really knowing what in the heck it is.

[The Wolf]

Originally Posted by Fyrestorm3

This is the issue that I have with WAR. I like the stat, and I think it could be a very useful one. I personally can't really wrap my head around the idea of a "replacement player," though that hasn't stopped me from using it. And, in fact, I find that it's great to use in OOTP, where you can be absolutely certain that every player in the game has been put through the same algorithm to determine their WAR. But in real life, I'm iffy on it, simply because there's multiple ways of calculating it. That right there suggests that the calculations leave far too much room for error to use it as anything more than an ancillary stat. If it could be narrowed down to a single formula across the board, I'd be much more okay with its use on the whole.

All that you have to do is use a consistent version and you're fine.

Come hang out on Fangraphs or Baseball Prospectus and you can learn all that you need to know about WAR, regression, projections, sample size, and a thousand other useful things. I have learned enough about sabermetrics that I can actually argue successfully with the big names, and all of that knowledge transfers usefully and directly to watching MLB games and playing OOTP.

[The Wolf]

Unifying Replacement Level | FanGraphs Baseball

[The Wolf]

WAR Comparison Chart - Baseball-Reference.com

[The Wolf]

fWAR and rWAR | FanGraphs Sabermetrics Library