Jump to content

Sell me on WARP.


Lt Melmo

Recommended Posts

I'd like to think of myself as a stats guy... I'm not too well-versed in working with them, but I'm a big supporter of someone else doing that work for me. :) And for a while, I'd blindly quote WARP3(and VORP to a lesser extent) as if I had any freaking clue what it was. And I don't. I know what it's supposed to tell you, I understand the concept of replacement players... I normally scoff when the purists talk trash about stats, but one thing I can't really disagree with was that people seem to be using VORP, the biggest, baddest total player rating when they don't know how it's calculated.

So how is it calculated? Does anyone really know anything beyond the basic theory? And why isn't that info freely available?

Also, WARP uses defensive statistics as part of its basis. Which one? Is it FRAA? And isn't it even somewhat agreeable to the biggest statheads that there'll never be a truly adequate way to judge defense statistically, or at least one such way doesn't exist now? Why do we quote WARP so often when at its very core is some defensive stat that's inevitably flawed in some way?

How is it that Cal Ripken's best season(an incredible 17.0) is 2.5 wins more than anything A-Rod has done in a season? Cal's '91 season was great but A-Rod has a chance to become the home run champ and has had some pretty incredible seasons.

And lastly, the numbers used to quantify replacement level productivity for VORP... 80% for most, 75% for catchers, 85% for first basemen and DHs. These numbers just strike me as extremely arbitrary, vulgar estimates, for a stat so many people herald as one of the best and most reliable.

I'm not knocking it... though I have stopped using VORP/WARP to look at players. I'm sure there are some things I just don't know so it'd be cool if someone could enlighten me.

Link to comment
Share on other sites

WARP uses FRAR for the defensive side, and that's the big problem with that stat imo. It's good to use if the defensive side of it for an individual player happens to be pretty accurate. But in many cases, the inaccuracy of the FRAR screws it up imo.

Link to comment
Share on other sites

So how is it calculated? Does anyone really know anything beyond the basic theory? And why isn't that info freely available?

It's proprietary to BP, and as far as I know, they haven't released anything about WARP other than its conceptual framework. I'm quite sure there isn't anything out there giving you a thorough idea of how to grind through the math and come up with these numbers.

Also, WARP uses defensive statistics as part of its basis. Which one? Is it FRAA? And isn't it even somewhat agreeable to the biggest statheads that there'll never be a truly adequate way to judge defense statistically, or at least one such way doesn't exist now? Why do we quote WARP so often when at its very core is some defensive stat that's inevitably flawed in some way?

The problem isn't with the lack of quality defensive statistics. The problem with WARP is that it uses a non-PBP based metric for its defensive assessments. Which is fine for Honus Wagner or Cal Ripken, but for Jose Reyes or Hanley Ramirez we have PBP based metrics (UZR, RZR, other advanced zone ratings, PMR) that are much better than what WARP uses (i believe it is FRAA or FRAR, not sure).

How is it that Cal Ripken's best season(an incredibly 17.0) is 2.5 wins more than anything A-Rod has done in a season? Cal's '91 season was great but A-Rod has a chance to become the home run champ and has had some pretty incredible seasons.

First, the season that Cal had a 17.0 WARP3 was '91. He finished 3rd in the league in OPS+ and 5th in PA - which is big because WARP accumulates with playing time. BP has him at 93 BRAR that season. ARod has surpassed that total twice - '05 and '07. ARod had other years just as good ('96 and '00), but he had "only" 670 PA in those seasons.

But the big difference is defense. ARod rates as an excellent defender for most of his career. His best season was '00 where he was +47/+15 (according to FRAR/FRAA). But Cal rates as a MUCH better defender, with five seasons of +59/+20 or better, including '91 when he was +64/+28. So in addition to being elite with the bat, they rate Cal as ~2 wins, give or take, better with the glove. And that's just at their peak. Compare Ripken '91 with ARod '07 and Cal is +38 FRAR compared to ARod. That's huge when you consider that at his peak Cal was also one of the top hitters in the league, albeit not as good as ARod.

I'm sure someone else will chime in and correct me if I'm wrong about this. But in general, I don't really put much stock in WARP at all. I'll look at it, but without knowing how it works and without a better defensive system, it is pretty insufficient.

And that's not to mention the other problems. The bar for a replacement level player is probably too low given that they treat defense and offense separately. IOW, a replacement player is Manny Ramirez with the glove and Brandon Fahey with the bat, rather than an great glove, no hit guy or a good hitter who stinks defensively. Case in point, Nick Punto hit .210/.291/.271 on the year, rated as an average defender at a corner, but he still rated at 0.3 WARP (and 2.0 WARP3).

Link to comment
Share on other sites

I think BP's rate/FRAA/FRAR are probably the best way I've heard to evaluate defense in the pre play-by-play era, so it's probably the best tool for those kinds of like, Aaron vs. Mays vs. Williams vs. Ruth type debates available, even if it isn't exactly perfect.

It loves Ripken's 1991 so much because aside from being a shortstop who batted .323/.374/.566(2nd in SLG, 3rd in OPS, 1st in Extra Base hits, etc.) in Memorial Stadium in a league with an ERA of 4.09, it also believes he was a magnificent, historically great, Peak Ozzie Smith-level fielder that year. So, yeah, the only better season from a position player in the history of baseball comes from Babe Ruth.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.



×
×
  • Create New...