You tell me. What do you think the chances are of Jason Bay posting another -1.3 RAA in a full season? Extremely unlikely?
yes, but more likely than Holliday. If we just tossed it out because it's measurement error or whatever, we'd say both guys are essentially the same. The bill james projections up on fangraphs now say .380 wOBA for Bay and .395 for Holliday. That sounds more reasonable than a wash, given that Holliday was better to the tune of 50 runs a couple of years ago.
I agree with your last bit though. I haven't really looked into it, but there's a point where signing Holliday is a bad idea. I'm still holding out hope that he doesn't want to hit in Citi field, Coors scares off a few teams, the Yankes re-sign Damon, the red Sox keep Bay, and we're able to give him a little more than some other teams since it won't mean losing a draft pick.
Mike Cameron is starting to sound like a really good idea.
cards2468 wrote:if you go by baseball-reference's defensive metrics, Holliday appears to be a terrible defender, however I've learned that b-r's defensive metrics are almost useless, and according to UZR, Holliday is definitely a decent defender.
That was my basic point--how much do you trust UZR. It's a fine tool for informational purposes. But using it as the *major* factor between Holliday being valued as a $25MM/yr. player vs. Bay being valued as a $15MM/yr. player, I think it's worth questioning it's value, since it is still a somewhat subjective measure of defense.
What defensive statistic isn't? Until a good pitchFX type system for fielding is implemented, I think I'd mostly trust scouting reports and UZR. Baseball-reference claims Matt Holliday is 13 runs below average at his position over his career while Manny Ramirez is only 7 runs below average over the past 7 years. I have a hard time believing Manny is that much better, or even better in general, than Matt Holliday defensively, so I tend to put a little more trust in UZR which has Holliday at a UZR/150 of 6.9 over his career while Manny has been at -12.1 over his career.
I'm not going to say that any defensive metric is perfect. Neither is any offensive metric, but they are a bit more reliable. In my opinion, you can look at both factors--offense and defense--and even convert them to the same units (runs) to compare. But I think it's starting to get sketchy when you start combining two sets of data into one final number, when you have 90% confidence in one set of data, and only 50% in another. Random percentages, but you get my point. I'd rather see them separate.
I think teams operate that way to an extent, or IOW there's a discount on defense. Either this makes sense or it's something to exploit.
Whatever the case, you do need to come up with one number if you are going to place a value on a player. Also, like I mentioned previously, when you are dealing with years of UZR data summed up in that /150 number, that helps things quite a bit.
haltz wrote:I mean in terms of throwing out 'outliers'.
If you are just making a general point about UZR, it correlates better year to year than ERA.
I guess I don't know enough about UZR and UZR/150 to be able to tell much. I know that a large positive number is good but I guess I don't know the difference, at least as it translates on the field between +4 and +16, for example. I know that Pujols was a +16 one year (2007) and a +0.8 this past year. He's also been an +8.5, +4.5 and a +1.0.
I suppose I could safely conclude that he's a better defensive first baseman than someone who is chronically in the negatives, but other than that, I don't know. Is he more like a +16 or a +0.8, or somewhere in between? Would throwing the +16 and +0.8 out affect the opinion or conclusion about where he lies? I guess I would have assumed defensive statistics to be a little more consistent than offensive. In basketball, they say something along the lines of defense never having an off night. Defense in baseball is obviously quite different but I've always assumed that if you get a good defensive player you're going to get good defense out of him. Maybe +16 and +0.8 are both good defense, just different levels of it. I don't know. But it seems that it would be dangerous to assume that someone having a +16 year defensively is definitely a better defender than someone posting a +0.8. Perhaps it's safe to say they had a better year defensively but does that mean they're definitely better defensively?
haltz wrote:Mike Cameron is starting to sound like a really good idea.
My sense from watching Cardinals-Brewers games this year was that our brain trust thinks Cameron can be pitched to. Maybe it's the 17 strikeouts in 52 at-bats talking (and that's against the 2009 pitch-to-contact Cardinals), but Cameron won't get the La Russa "He doesn't give away at-bats" seal of approval that Holliday got.
That weird shirt untucking stuff that I don't understand could cause friction as well.
You tell me. What do you think the chances are of Jason Bay posting another -1.3 RAA in a full season? Extremely unlikely?
yes, but more likely than Holliday. If we just tossed it out because it's measurement error or whatever, we'd say both guys are essentially the same. The bill james projections up on fangraphs now say .380 wOBA for Bay and .395 for Holliday. That sounds more reasonable than a wash, given that Holliday was better to the tune of 50 runs a couple of years ago.
I agree with your last bit though. I haven't really looked into it, but there's a point where signing Holliday is a bad idea. I'm still holding out hope that he doesn't want to hit in Citi field, Coors scares off a few teams, the Yankes re-sign Damon, the red Sox keep Bay, and we're able to give him a little more than some other teams since it won't mean losing a draft pick.
Mike Cameron is starting to sound like a really good idea.
Not that much more likely than Holliday. You would have to think a RAA below 10 for either player is aiming for the skinny part of the bell curve. I have no idea what happened to Bay in 2005 (actually I do know--I drafted him), but since then he's been chugging along in the 30's.
I hope that we sign Holliday because we desperately need a bat. But I know that means paying him $18MM+, at the very least. That makes me cringe. For a couple years, sure we could do that. To wrap that up for 7 years--ugh, I don't know. It was hard enough going through Rolen and Edmonds' unproductive years at the end of their big deals. Holliday's salary is going to be twice as expensive, and he won't be an elite defensive player at 3B or CF.
haltz wrote:I mean in terms of throwing out 'outliers'.
If you are just making a general point about UZR, it correlates better year to year than ERA.
I guess I don't know enough about UZR and UZR/150 to be able to tell much. I know that a large positive number is good but I guess I don't know the difference, at least as it translates on the field between +4 and +16, for example. I know that Pujols was a +16 one year (2007) and a +0.8 this past year. He's also been an +8.5, +4.5 and a +1.0.
I suppose I could safely conclude that he's a better defensive first baseman than someone who is chronically in the negatives, but other than that, I don't know. Is he more like a +16 or a +0.8, or somewhere in between? Would throwing the +16 and +0.8 out affect the opinion or conclusion about where he lies? I guess I would have assumed defensive statistics to be a little more consistent than offensive. In basketball, they say something along the lines of defense never having an off night. Defense in baseball is obviously quite different but I've always assumed that if you get a good defensive player you're going to get good defense out of him. Maybe +16 and +0.8 are both good defense, just different levels of it. I don't know. But it seems that it would be dangerous to assume that someone having a +16 year defensively is definitely a better defender than someone posting a +0.8. Perhaps it's safe to say they had a better year defensively but does that mean they're definitely better defensively?
That's the way I see it. No different than offense.
the difference of 12 runs is like 16 plays over the course of the season, or one play every ten games. Tough to tell by watching. it also could be that the player got unlucky with the balls in his zone, either just had an off season or they were harder to field than other balls that were classified the same way.
Pujols was also dealing with only 250 balls in his zone. Think about it this way: he could hit 500 baseballs and the spread of liners, flies and grounders is the same throughout (just an example), but in one sample of 250 10 balls were lucky hits. He didn't hit them all that well but they found a hole or blooped in. In the other sample 10 balls that he centered up and smoked were hit right at fielders. In the first sample he recorded 160 hits for an average of 320, but in the second he recorded 140 hits for an average of .280. Didn't do anything all that differently, but that's a big deal in terms of batting average.
Most critiques of UZR can be applied to offensive stats as well. The batter is dealing with seven players in seven zones, who may make a miraculous play or it might bloop in or he might tag a ball right at someone. This evens out over time, just like it does for UZR. Except for UZR you need more years to get the same number of chances (depends on the position).
It's not quite that simple, but I think that's a helpful way of looking at it.
haltz wrote:
Whatever the case, you do need to come up with one number if you are going to place a value on a player.
The only singular number you need to come up with is the dollar figure on the contract. WAR tries to get to that figure by treating offense and defense as equal parts in the equation. Even though they *are* even in a baseball sense, I think you have to apply some common sense in how much faith you have in the individual components.
What that ratio or whatever is, I don't know. Making outside-the-box comparisons such as Bay = Mike Cameron are OK when you're just making comparisons on a baseball forum. But if I'm a real team devoting real dollars to these players, I'd have a bit higher pucker factor trusting UZR as much as offensive metrics. Which is probably why ultimately Jason Bay will earn a lot more than Mike Cameron next season.
InvincibleCakeEater wrote:That's the way I see it. No different than offense.
But what is Albert, UZR-wise? The +16, like two years ago, or the +0.8, or somewhere in between? And what's to say what he will be next year? Another +0.8 or back to the +16? And how do you stack him up against someone who was a +1.0 last year? Is he better, because he was a +16 a couple years ago? And what's to say that the +1.0 player won't become a +16 next year, like Albert once was? Seems like a lot of variance, although I don't think we talk about defense like it's a wildly fluctuating variable.
