Ryan Braun's win value

[kattleflap]

Hey guys I was just curious, is there any way to find the statistical value of how many wins Ryan Braun would have been worth if he had been playing from opening day up till his call up? Would we theoretically won more games during those first 50 of the season or not? I'm not very good with understanding VORP or even sure if that would be the right stat to use, but I know there are some of the smartest baseball stat guys I've seen on this board, so if anybody has the answer I'd appreciate it.

[TerraceReserved]

For win value, WARP would be a better tool, as it includes defense. This year he had a 4.4 WARP, in about 450 AB. If he was up the whole season, and he got roughly 150 more AB's, he would of added 1.4 wins, for a total of 5.8 Wins Above Replacement. So theoretically we would have added one about 1 and a half more games had he been up from the start.

[RightFieldCoder]

He would have added 1.5 games above a non existant replacement. The other part of the equation was how did Graffounsell perform at 3rd base for those six weeks? If they managed 1.5 WARP during that time, it's a wash. Anybody have that answer?

[Jimbo]

Keep in mind the 24-10 start. I don't see Braun adding more wins at the start, just larger margins of victory.

[rluzinski]

Has BP improved their defensive metrics or are they still ages behind? If they are still using put outs, errors, etc... I'd ignore their defensive numbers altogether.

[RoseBowlMtg]

Not to be a wise guy but do they ever look at WARP for what actually happened? When Braun came the Brewers were 28-19 or a 59.6% winning clip. While with the Brewers the Brewers were 53-60 or 46.9% clip. To say he helped the team 4.4 games when they lost 16 games to .500 is interesting to say the least. To say he would have improved a 28-19 team by 1.4 wins when his only history was that he was on a team that was 16 games below .500.

 

Just curious how they justify that when his only history has been at a negative pace. Will a team of Ryan Braun's win?

[TerraceReserved]

He would have added 1.5 games above a non existant replacement. The other part of the equation was how did Graffounsell perform at 3rd base for those six weeks? If they managed 1.5 WARP during that time, it's a wash. Anybody have that answer?

 

Ahh, you're right. Graffy this year had 1.2 WARP with 232 AB's, and Counsell has a 1.3 WARP with 281 AB's. If we say they both got 75 AB's that Braun could of had, Counsell would up put up .35 WARP and Graffy .38, for a total Warpof .73. So Braun would of added about .7 wins during this time over Graffasell.

[rluzinski]

Just curious how they justify that when his only history has been at a negative pace.

That's not his history but the team's. It's like saying Bonds contributed negative wins to the Giants because they had losing season.

If you want to now what the result of his production was within the context of the actual games, you might want to look at something like WPA (win probability added). It looks at the probability that the player's team had to win before and after an AB. For instance, if the Brewers had a 75% chance of winning a game just before Braun hit a walk off HR, he'd recieve .25 WPA (1.00 - .75). You can get negative WPA as well.

Brewers WPA

Braun had a WPA of 2.1. Of course, that doesn't consider defense, which Braun was terrible at this year.

[1992casey]

Will a team of Ryan Braun's win?

That's an interesting question, and I wonder how one would go about trying to answer it. For the defensive aspect, I'd think you'd get balled up with the fact that you'd be considering stats based on nine third basemen.

 

[RightFieldCoder]

RBM: Warp is Wins Above Replacement Player. If I understand it correctly, a "replacement player" is the AAAA player at that position that you could find on most teams. So the Brewers were 4.4 games better with Braun in the lineup, but only compared to an average minor leaguer. If Braun had a 4.4 WARP in 492 PAs and Graffasell had a 2.5 WARP in 594 PA's, Graffasell would average out to a 2.1 WARP in Braun's 492 PAs. So Braun added at least 2.3 wins when he started playing. I say at least because having him up left Graffasell to be used at 2B (while Ricky was finding himself) and as a PH. The team would have probably had to use even lesser players in those roles if Graffasell was tied up at 3rd. The real amazing thing is how bad bad of a collapse would we have had if we hadn't called Braun up?

[Brawndo the Thirst Mutilat]

Has BP improved their defensive metrics or are they still ages behind? If they are still using put outs, errors, etc... I'd ignore their defensive numbers altogether.

 

FRAA still doesn't use play by play data or any kind of concept of zones. I'm pretty sure they have fancy (presumably, they are proprietary and not disclosed) ways of adjusting for things like park effects, pitching staff variation (K-rate, GB/FB ratio), and so forth, but they're still basically just using the number of plays made and guesstimating the number of opportunities.

 

They're generally considered the least reliable of all the advanced fielding metrics, but they do a good job of making the numbers constantly available and constantly up to date, so they still get talked about a fair amount. As far as I know, the main reason they've been resistant to changing their methodology is that once you go the play by play route, you lose the ability to include the bulk of baseball history, for which no such data exists.

 

I'm anxious to see what the better zone systems (especially Dewan's ) end up saying about Braun's defense, but based on THT's revised zone rating and OOZ (out of zone) numbers, it's not likely to be pretty.

[Oldcity]

I'm curious to see how he measures up in WARP to Tulo. Does Tulo's defense really make him deserving of the ROY over a player with vastly superior offensive numbers?

[RoseBowlMtg]

rlu thanks for the information and the link but tell me this: Does WPA really tell how the person did that year compared to other years?

 

For example:

 

...........HR, RBI, ave, obp, slg, WPA

Lyle: 19, 72, .276, .367, .449, ?

Lyle: 16, 87, .301, .385, .478, ?

 

I think Overbay's .301 year was better since he had 15 more RBI's, 25 pts higher in BA, 18 pts higher in obp and 29 pts higher in slg

 

but

 

When Lyle hit .301 his WPA was 1.71

When Lyle hit .276 his WPA was 3.93

 

Why such a difference for an inferior year?

 

Thanks.

[rluzinski]

FRAA still doesn't use play by play data or any kind of concept of zones.

 

And that's why you never hear any serious sabermatricians (and no, I don't consider myself one) ever quoting them.

 

I'm anxious to see what the better zone systems (especially Dewan's ) end up saying about Braun's defense, but based on THT's revised zone rating and OOZ (out of zone) numbers, it's not likely to be pretty.

 

Check this out:

 

http://spreadsheets.google.com/pub?key=pfk_WuYpfduzefA8-s0_iYw&gid=0

 

I'm curious to see how he measures up in WARP to Tulo. Does Tulo's defense really make him deserving of the ROY over a player with vastly superior offensive numbers?

 

According to the above link, Tulowitzki was the best SS in the league, at +27.6 runs. Braun was the second worst 3B at -32.1 runs. Had he started the whole year, he would ahve easily been the worst. When each error is worth about .8 runs, it's easy to see how quickly bad defense adds up.

 

Why such a difference for an inferior year?

Without looking at the actual numbers, WPA is context sensitive, so raw stats don't tell you the whole story. If a player with mediocre numbers always did poorly when his team was already well on it's way to losing but always performormed well in clutch situations, his WPA would be huge. It tells you how much a player's performance influenced his team's wins and losses. It doesn't necessarily tell you how good of a player he his, however.

[Brawndo the Thirst Mutilat]

I'm curious to see how he measures up in WARP to Tulo. Does Tulo's defense really make him deserving of the ROY over a player with vastly superior offensive numbers?

 

BP has Tulo at 10.5 WARP, most of it from fielding. They have him as approaching Ozzie Smith level on defense. They have Braun at 5.3 WARP, and below replacement level on defense. That's really saying something, because FRAA has been widely criticized for setting replacement level too low.

So Tulo wins, and it's not even close. For me, this was the final nail in the coffin for FRAA. Braun's bad, but there's no earthly way he's bad enough to make him half the player Tulo was this season. Just no way.

[Fatter than Joey]

Warp is Wins Above Replacement Player. If I understand it correctly, a "replacement player" is the AAAA player at that position that you could find on most teams

 

The way I had it explained to me, is if you had a team of "Replacement Players" -- they would win 25 games in a 162 gm season.

 

I'm curious to see how he measures up in WARP to Tulo. Does Tulo's defense really make him deserving of the ROY over a player with vastly superior offensive numbers

 

Braun WARP 4.4

Tulo -- WARP 8.5

 

Basically Tulo plays a hard position very well -- Braun plays an easier position wretchedly. I would vote for Braun -- but I don't expect it will be a landslide.

 

There is a lot of misinformation floating around to the effect "(Insert HOF 3b) made a lot of errors but then got better" -- As good as Braun has been at the plate, he has been as bad on the field.

 

I'm anxious to see what the better zone systems (especially Dewan's ) end up saying about Braun's defense, but based on THT's revised zone rating and OOZ (out of zone) numbers, it's not likely to be pretty.

 

I agree -- Although sometimes you just don't need stats to see that a player is bad.

 

Has BP improved their defensive metrics or are they still ages behind? If they are still using put outs, errors, etc... I'd ignore their defensive numbers altogether.

 

With regard to ROY voting -- if this is the only metric the voters have, it will speak the loudest.

 

Also -- let's not forget Braun was banged up in ST -- so he wasn't a realistic option in April anyway.

[rluzinski]

Tulowitzki had about 35% more AB on the offensive side. As for defense, the above spreadsheet had Tulo's defense worth 6 wins over Braun. It's hard to imagine a larger defensive difference.

[Brawndo the Thirst Mutilat]

Thanks for the link, rluz. One really interesting thing I noticed. Fielder, Weeks, and Hardy are all below average in on in zone and above average out of zone (Weeks and Hardy are well above average). Knowing that the main shortcoming with the zone systems is that they cannot account for defensive positioning, I'm wondering whether the Brewers' inordinate amount of shifting is playing Loki with these numbers more than a little bit. Hardy, at least, should be the opposite - very solid in zone, but maybe a bit below an average major league SS out of zone. And how the bejeebus is Fielder above average out of zone?

[Brawndo the Thirst Mutilat]

FTJ and I posted different WARP numbers, but we're both right. He used WARP1 and I used WARP3, which adjusts for historical context (things like level of offense, length of season, etc.). Either will work fine to compare two players in the same season, WARP3 is better for historical comparisons, though it still relies on BP's dubious fielding numbers.

[rluzinski]

Knowing that the main shortcoming with the zone systems is that they cannot account for defensive positioning.

 

I can't disagree.

[Brawndo the Thirst Mutilat]

rlu thanks for the information and the link but tell me this: Does WPA really tell how the person did that year compared to other years?

 

For example:

 

...........HR, RBI, ave, obp, slg, WPA

Lyle: 19, 72, .276, .367, .449, ?

Lyle: 16, 87, .301, .385, .478, ?

 

I think Overbay's .301 year was better since he had 15 more RBI's, 25 pts higher in BA, 18 pts higher in obp and 29 pts higher in slg

 

but

 

When Lyle hit .301 his WPA was 1.71

When Lyle hit .276 his WPA was 3.93

 

Why such a difference for an inferior year?

 

Thanks.

 

That's why WPA hasn't really gained much of a foothold in the SABR community beyond just as a novelty, or a tiebreaker in MVP discussions to decide who was more "clutch." (nobody say that too loud though, 'cause it's probably the best argument for Prince to win the MVP this year). Here's a link with more info on WPA:

 

http://www.hardballtimes.com/main/article/the-one-about-win-probability

 

If you just want to know how many runs or wins a guy's bat was worth, you can go with VORP or any of the linear-weights based systems and they'll all tell you pretty much the same thing. This is how serious SABR guys know that they've solved offense, and it's why I think defense is still totally up in the air in terms of objective measurement.

[rluzinski]

Rosebowl asked how to measure the actual value of a player's performance to his team. WPA is one way to do that. If you want to know how valauble a HR was, you either have to relate it to runs or wins. If we are talking runs, you can still use linear weights but have to also take into account the base/out situation:

 

http://www.tangotiger.net/lwtsrobo.html

 

If based on wins, use WPA. A player can't be "most valuable" if he's never put into a position to win any games. It has little to do with the actual underlying skill of a player but neither does much of the awards dished out at the end of the season. If you want to use stats to estimate a player's skill, context is largely irrelevant.

and it's why I think defense is still totally up in the air in terms of objective measurement.

 

While they are not perfect, I think there's some real value in the current crop of zone based defensive metrics.

[Brawndo the Thirst Mutilat]

If you want to use stats to estimate a player's skill, context is largely irrelevant.

 

Right. I guess I just care a lot more about that end of the analysis, because I still think "clutch" doesn't exist as a consequential measurable skill. Frankly, I don't care much about MVP discussions. I care a little bit more about HOF discussions, but those generally ignore "clutch" (at least in SABR circles) because a player demonstrated over the course of a 15-20 year career that he was the same in clutch situations as all others. I do think WPA is the most intriguing attempt anybody's ever made to define and measure "clutch," I just don't give a rat's fanny about that endeavor.

 

While they are not perfect, I think there's some real value in the current crop of zone based defensive metrics.

 

I don't disagree. I just don't think they've come anywhere near the point where you can say with any confidence exactly how many runs a guy saved or cost his team. My approach is more like "If all these metrics agree a guy is good, he's probably good. If they all agree he's bad, he's probably bad. Even if scouts and/or my eyes tell me different." But when you're looking at a number, say -40 plays below average, what I read is somewhere between, say 10 plays and 60 plays (when in doubt, regress to the mean of an average fielder, since we just don't know). Until I see some more agreement between the metrics, that's as much confidence as I've got in them.

[rluzinski]

Most fans don't care about skill, they care about what a player accomplishes with that skill. It's why all the seasonal awards are based on accomplishments (RBIs, clutch hits) and I appreciate why it's that way (although I don't particularly care who wins them as a result).

[Brawndo the Thirst Mutilat]

rluz - The problem I have is that most fans also still think that RBIs and clutch hits are useful in a predictive fashion. WPA was a perfectly sensible thing to point out in response to Rosebowl's original context-driven question (in fact, I would have pointed it out if you hadn't already done so by the time I saw the question), but I wanted to answer Rosebowl's follow-up question on how Overbay's WPA could be better in a season that was clearly worse, both by conventional stats and more SABR-ish ones. Weird, flukish, non-replicable luck is the straightforward answer. I probably should have been clearer about that in my original post.

 

That's also why WPA is essentially useless in answering the original question this thread pursued - how many more wins would the Brewers have had if Braun had been up from the start of the season? Whether you're projecting forwards or backwards in a counterfactual fashion, I think you've got to use the skill-measuring metrics rather than the value-in-context-measuring ones.

 

In handing out MVP awards, voters should feel absolutely free to include lucky timing of production in the right situations as a criterion in their decision-making. But they can't feel that way because of modernity's love of merit and disdain for luck. The ancient Greeks had this figured out much better than we do -- if somebody was lousy at everything but lucky enough to succeed anyway they counted that as a virtue. "Favored by the gods," they said, and bore the lucky idiot no resentment (or at least, they were supposed to try not to). We have to go the opposite route and assume, perpetually, despite the glaring lack of evidence to support our view, that it couldn't have been dumb luck, just couldn't have been, because that just wouldn't be fair. And lo and behold, before you know it, some dumb sportswriter ends up calling Derek Jeter the very best he's ever seen play.

 

I'm certainly not suggesting you're in that camp. I suspect we really don't disagree much about the utility of WPA or many other questions of interest to the SABR-literate baseball fan. I just tend to be given to a bit of hyperbole now and then, and so say things like "defense is still totally up in the air with respect to objective measurement" when all I really mean is "Though progress has been made, I don't think we're yet in a position to treat runs-saved figures based on even the very best metrics as though they are remotely as reliable as runs-produced figures based on linear weights." It's quicker the first way. And more fun.