Ryan Braun's win value

[pogokat]

sweet a saber toothed showdown...

 

rluz---put that in your pacman suit and smoke it

[Brawndo the Thirst Mutilat]

No showdown intended. rluz has been my favorite poster here during the entire time I've been lurking, which has been awhile.

[Brawndo the Thirst Mutilat]

In fact, if he hadn't been here saying most of the things I wanted to say on most of the threads, I probably would've joined a lot sooner.

[rluzinski]

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.

 

With regard to the sabermetric topics found in this thread, I'm not sure we disagree at all. I think I might just be a little more forgiving with regard to why fans typically look at the "wrong" things for MVP voting. It's really combining skill, opportunity and luck and I don't have a problem with that. I accept that the MVP is ussually not the most skilled player in the league, whether the fans realize that or not.

 

BTW, MGL posted the best and worst 2007 UZR numbers by position and league in this post:

 

http://www.insidethebook.com/ee/index.php/site/comments/best_and_worst_of_2007_uzr/#comments

 

Braun didn't qualify but when I asked MGL about him, he had this to say:

 

"Braun would certainly be the worst, if he qualified. Even using zero UZR to make up the games he "missed" he is still probably the worst. That is a shame since he is such a good hitter, and his poor defense takes most (75% or so) of his hitting value away."

 

Costing your team 30 runs on defense is huge.

[Fatter than Joey]

So if you take "75% of his hitting value" away -- is he an above average or average player?

 

I wonder why Weeks isn't listed -- he must in the middle somewhere.

[rluzinski]

I take that as meaning he taking away 75% of hitting value, relative to a replacement level 3B. He still had positive value.

[endaround]

Braun's defense basically makes him an average player.

[Fatter than Joey]

To add fuel to the fire I see Tulo +20 is 2nd in the NL SS to Reyes.

 

So Russ, who would get your ROY vote? (assuming it is between Braun and Tulo)

[rluzinski]

I don't even know what the qualifications are for ROY. Just total value of this year's performances? Am I supposed to consider park factors, for instance? Tulo has a .720 OPS on the road, so I think his offensive value might be largely smoke and mirrors. Still, by all accounts, his defense is for real. Braun's defense was so brutal, however, that it's hard for me to say he deserves ROY.

[Fatter than Joey]

Am I supposed to consider park factors, for instance?

 

Sure -- I think VORP has built in park factors -- I think Braun was like 57 and Tulo was 36

[Brawndo the Thirst Mutilat]

Here's the thing I've never been able to get past about the runs-saved values I see with all the defensive metrics:

We know that run prevention is some combination of pitching and fielding, but we have a heckuva time separating the two except for 3TO events. On balls in play, it is very difficult to distribute credit/blame for the out or non-out between the pitcher and the fielder.

We know that run scoring is basically just hitting, with a tiny bit of baserunning on the margins.

Ryan Braun was an enormously good hitter this season. Historically good, actually, when placed in the context of his age. He was still only 36 runs above an average 3B. That's a function of how hard it is to contribute beyond what an average major leaguer contributes. An average major leaguer is (obviously) a very good baseball player.

So when I see runs saved fielding figures that have a player 20 or 30 runs above or below average for their position, I have a hard time believing that number. My guess is that at least 50% of what they're measuring there (was the play made or not?), and maybe a lot more, is due to some combination of the pitcher's contribution and dumb luck (not all opportunities are created equal, and I don't think that evens out for in the field quite like it does at the plate -- over the course of 600 PA, 2400 pitches seen).

Just using a conservative assumption that run prevention is 50% fielding, 50% pitching, Braun's -30 runs saved means that he is actually 2 times further removed from an average 3b in the field as he is at the plate (opposite directions, obviously). I just don't believe that is possible. I don't think he has enough opportunity to do that much damage, no matter how bad he is. For years, the SABR crowd underplayed the significance of defense because it was too difficult to measure...I think there is something of an overcorrection going on. I certainly could be wrong.

[Fatter than Joey]

For years, the SABR crowd underplayed the significance of defense because it was too difficult to measure...I think there is something of an overcorrection going on. I certainly could be wrong.

 

I don't doubt this is true to some extent.

 

Ryan Braun was an enormously good hitter this season. Historically good, actually, when placed in the context of his age.

 

He has been historically bad as well with the leather.

 

Braun's -30 runs saved means that he is actually 2 times further removed from an average 3b in the field as he is at the plate (opposite directions, obviously). I just don't believe that is possible

 

I don't know if you look at pitchers like Suppan, obviously our defense is wreaking havoc.

 

I just don't know how they keep him at 3b next year with a straight face.

[Brawndo the Thirst Mutilat]

Based on THT's zone rating, the entire Brewers defense was 48 runs below average. That number I believe. 30 for an individual player I have a much harder time believing.

[Fatter than Joey]

That number I believe. 30 for an individual player I have a much harder time believing.

 

Yeah -- I see what you are saying... esp. when I suspect Fielder/Hall/Estrada are all (neg) as well, and guys like Hardy/Jenks are probably only marginally (pos).

 

I think though Braun was so bad and so far away from the mean, that perhaps his impact is skewed. Sometimes metrics do not handle data from the extremes

very well.

 

I don't think there is any denying he is substantially bad.

[Brawndo the Thirst Mutilat]

For whatever it's worth, I would add that I'm sufficiently convinced that Braun is very bad to be in favor of moving him to the outfield. I just don't buy that his glove almost completely cancels his bat.

Winning = 50% run scoring, 50% run prevention
Run prevention = 50% pitching, 50% fielding (??? - this might be overestimating fielding's contribution)

Braun's share run scoring = (1/9)
Braun's share of run prevention = (1/9) x (50%) = 5.6%, at most.

All of those are bad, sloppy, back of the envelope guesses, but the basic logic is what makes me question fielding runs figures that rival batting runs figures in magnitude.

[endaround]

Last time I added up all of BP's stats, it got the Brewers to like -49 FRAA. and yes Braun is large out in front. And the run cost is just based off the cost of a baserunner and not making the out. So no its not context sensitive just like no on goes in and discounts home runs when ahead by a lot or trailing by a lot.

 

Previosuly I think Russ said the cost on an error is like .8 runs. Braun's errors alone then lead to about 20 runs.

[Brawndo the Thirst Mutilat]

Braun's errors alone then lead to about 20 runs.

But an average 3B doesn't make 0 errors over the course of Braun's innings. So that's not 20 runs below average. I don't know what the correct baseline is here, and I don't really care enough to look, because whatever their faults, the fancy metrics are eons ahead of using errors and fielding percentage to assess individual and team fielding.

The 0.8 figure is what is generally used for the cost of an "unmade play." The goal of all the zone based metrics is to figure out how many plays a player made above or below what an average fielder at his position would have made, given the same set of opportunities. It's the last part I don't think has really been solved yet. I'm not sure it can be, not totally anyway. None of them adjust for positioning. All of them rely on the subjective judgment of game scorers about what zone a ball was hit to, and what hit type it was. How do you tell the difference between a flyball, a "fliner", and a line drive, anyway? I don't have a better suggestion about how to approach this issue, I just think a fair amount of skepticism is warranted about the exactness of the numbers they produce.

Hitting, on the other hand, I have zero doubts about. If BP's stats say Braun's hitting was worth 36 runs more than an average 3B, then that is exactly how much, give or take like 2 runs (tops), his hitting was worth (context neutral). I offer this paragraph by way of contrast with the preceding one.

[TooLiveBrew]

Agreed, BTTM. Defensive metrics are getting better, but scouting is so much more accurate in assessing defense at this point (imho). It's nice to have some baby steps started toward defensive stats that actually do tell you somethign useful, but there are so many questions involved that it's clear they have a long way to go. I think once there are 5+ more seasons under the belt of defensive metrics, we'll really be able to begin to have both predictive and descriptive stats that are reliable(-ish) http://static.yuku.com//domainskins/bypass/img/smileys/wink.gif

[Brawndo the Thirst Mutilat]

TLB - I'd add to that assessment that sometimes the scouts are wrong and advanced defensive metrics help tell us when this is the case. It's taken until the last couple of years before any scouts started to have anything critical to say about Jeter's defense, while even the most primitive metrics (RF, ZR in it's original form), were telling us that he was a lousy fielding shortstop almost from the start of his career. When more advanced metrics came along, Jeter didn't look quite as awful, but he still was regularly near the bottom of all MLB shortstops. Jeter is a bad fielder. Jeter always was a bad fielder.

Then there's someone like Andruw Jones. Looked otherworldly good both to scouts and to metrics as a pup. Started looking bad in the metrics when he was still pulling down gold gloves every year and drawing comparisons to Willie Mays from scouts. Finally, some scouts started saying nasty things about his declining range (Jason Stark then labels him overrated in his dumb book). Meanwhile, IMO the very best advanced fielding metric available yet, John Dewan's plus/minus system, has him among the best CFs in baseball last year. At this point, I don't feel like I have any idea how good Jones is.

When all the metrics agree, and they disagree with the scouts, I side with the metrics.
When the metrics can't make up their minds, I'll go with the scouts.
When everybody agrees (as is sadly the case with Braun), there's not much room for debate except for on the "exactly how bad is he?" question.

[Fatter than Joey]

Well If you look at the link Russ posted and then the metrics that are linked in the comments you will see Ichiro is horrible in one and the best in the other.

 

It's shaky.

[rluzinski]

I agree that there are problems with even the most advanced defensive metrics but I don't think Brauns' terrible ranking is evidence of that. -30 runs means that Braun made about 40 less outs than an average 3B in 113 games. About one less out per 3 games. I wouldn't think that anyone who watched him this year would think that's not at least possible. I also don't find it surprising that a good hitter could negate a good portion of his offensive value with very bad defense.

 

Now, when I see a guy who's 5 runs from average, I consider that about average, since I don't think one year worth of defensive stats has a lot of accuracy.

[TooLiveBrew]

-30 runs means that Braun made about 40 less outs than an average 3B in 113 games. About one less out per 3 games. I wouldn't think that anyone who watched him this year would think that's not at least possible.

 

Ugh - this makes me sick. There's no way you can disagree with that statement. http://static.yuku.com//domainskins/bypass/img/smileys/frown.gif

[Brawndo the Thirst Mutilat]

One less out every 3 games is a lot bigger figure than it seems. Braun only had 219 BIZ (balls in zone), according to THT. That's 1.93 chances per game, or 5.8 chances every 3 games. Plus/minus systems also penalize fielders for failing to come up with an average number of out-of-zone plays (which they should). Braun made 21 such plays. Rolen, who is near the top of this list, made 39 in a similar number of innings. But he's excellent and we're concerned with average here, so let's split the difference and say Braun would have needed 30, or 9 more plays, to be average (a guess, but I hope a fair one).

 

In zone opps are defined as chances successfully fielded, on average, more than 50% of the time. Let's say that OOZ "chances" are converted more like 1/4 (another guess, don't know what else to do here). So Braun faced another 40, maybe, OOZ "chances" that are theoretically possible to field if you make an excellent play.

 

Add that to the 219 BIZ, and you've got 259 "total" chances. That works out to 2.3 per game, or 6.9 every three games. All this work is to show how few chances Braun actually had to cost the Brewers as many runs as they think he did.

 

"One less out per 3 games" means that an average 3B is successfully fielding something like 5/7 balls while Braun only converts 4/7 (it's actually closer to 4.5 and 3.5, but who likes working with fractions?). I did watch something like 80% of the games this season and, while I suppose that's possible, I just don't think it's very likely. 3B is a corner position, a hitting position. Average glove men there are better now than at some points in baseball history, but an average 3B still isn't exactly a wizard out there. Wes Helms shows up as just above average this year on the plus/minus estimator for THT's zone data linked early in the thread. I watched all his Brewers games too. There's no way that Wes Helms manages to convert 1 more chance than Ryan Braun every 7 tries.

 

That example is cherry picked for emphasis, but my basic point is that, when you're confronted with statistical outliers, it's a sensible reaction to wonder whether the measurement was entirely accurate. You needn't reject it completely - I don't doubt that Braun's defense was among the worst in the league. I generally make it a rule to maintain what I view as a healthy skepticism about individual fielding runs totals that are very high and very low, regardless of the system. If somebody's metric had Hardy saving 40 runs over an average SS, I'd be equally skeptical about that being the right number.

[Brawndo the Thirst Mutilat]

In zone opps are defined as chances successfully fielded, on average, more than 50% of the time. Let's say that OOZ "chances" are converted more like 1/4 (another guess, don't know what else to do here). So Braun faced another 40, maybe, OOZ "chances" that are theoretically possible to field if you make an excellent play.

 

This math is completely wrong. I don't have time to fix it - I've wasted way too much on that last post as is. So let me call myself out on being stupid before anyone else does. The error does drive down what I called the number of "total chances" in that post. Maybe it's more like 6/10 vs 5/10 every 3 games. I still don't think Wes Helms makes that extra play.