Defensive Prowess and Sabermetrics

1) No. I’ll take the team of Manny Ramirez’s of course (as would most anyone), but I think there’s some serious selection bias in your choices there. We all know Wilson’s true talent was below his performance in 2007 and Manny’s was above his performance 07. Jack’s team could win a few games, especially if the pitcher could get groundballs, but the true talent is not roughly equivalent here.

2) That pretty clearly depends on the context of the run environment. In 2003? No, not at all? In 2011? Possibly. In 1912? Certainly.

3) I don’t think it’s ever safe to base a true talent evaluation on one season of WAR or any other single statistic, no matter how large the gap in that season’s performance. Guys have fluke years, just as much so at the plate as in the field, and just as much negatively as positively. You have to look at a whole lot more than just one season of WAR to get close to determining true talent level.

1) I have to agree with the previous poster that your choice skews the question since you take one of the worst years of Manny (the second worst year in ten years) and best of Jack Wilson.
With that said, if the defense numbers are accurate (BIG if, since I think that UZR is trash or something crooked), I would take 9 2007 Jack Wilsons over 9 2009 Manny Ramirezes (butchers in the field, and having one of their personal worst offensive year to that date).

2) If you’re talking about replacement level, then, yes. If the comparison is with an MLB averagely valuable player, then I think, no.
Guti was so bad with the bat this year.

3) I would only be comfortable with calling player A more valuable than B, when completely disregarding UZR, if A had an advantage in oWAR, AND if A and B were similarly good, average, or bad in defense (by the eye test).
For two players with obviously largely different defense abilities, I think it’s a tough call, since we can’t trust the UZR numbers, to make a judgement call on equivalent translation of offensive contribution numbers.

By the way, I suspect something is very fishy (or possibly actually crooked) with UZR’s blackbox or it’s source numbers (some people are compiling the raw data, right?).

I know Ichiro is having a poor year by his standards, but no way do I believe that Ichiro is the worst defensive rightfielder in the major leagues (together with another Japanese player, Fukudome, who actually can cover CF).
Their UZR numbers are actually around 30% worse than the worst fielding RF in the league (Bautista).

Something’s fishy with UZR numbers.

And another thing which stuck out was that in terms of UZR numbers, Choo (although I generally like him) was ranked by far the best defensive RF in the league, before he got injured. Really heads and shoulders above the pack in UZR numbers.

No way.

Something’s fishy (maybe even crooked) with UZR, in my opinion.

Asides from any possible racial aspects in play, WAR (including UZR) is a widely used number now to evaluate player worth.
It’s so simplified that it’s dangerous (if not upboard and accurate), since anyone can remember one number, and make judgments based on this number.

oWAR can in general be objectively checked and calculated by any fan with an interest.

UZR is some kind of blackbox, of which the numbers we can’t easily check.
Let’s say some player get’s a crooked boost of 2 WAR from skewed UZR numbers. That would be worth 10 million per year on the market, since you can be sure his agent will be cramming these WAR numbers in the face of any GMs in negotiations.

This is huge money we’re talking about.

It would be huge money to whoever is controlling UZR’s blackbox, and/or any lowly scorer or whoever is compiling the raw data.

Asides from whether anything fishy is really going on or not, nobody can deny the possibility/ existence of temptation (or rather the mechanism in place) for potential wrongdoing, or to put it more mildly, just some tweaking of the numbers (that doesn’t sound so bad, does it?).

It worries me that this number (UZR, and by connection WAR) that we can’t objectively check, is having an increasingly important role in baseball.

I’m also disturbed because UZR didn’t seem THIS BAD up to last year. Something’s suddenly screwy this year.
Maybe some of the agents suddenly realized the “new inefficiencies in the market” (=subjectivity of UZR).

If they were both an excellent defender and a threat to steal a base every time they did get on base, but even then it’s probably hovering around the baseline for acceptable. Guti is a great defender but not as much of a base stealing threat so I don’t think he’ll cut the mustard unless he can recover his bat a bit.

1) I’d feel comfortable saying that both teams would win a similar number of games if they played a whole bunch of times.

2) I would definitely think that’s possible. I’m going with above replacement level with no consideration of salary as “net positive”, but even using “average” as the benchmark, I think it can be done. I’d have to say that the player would need to turn in one of the best defensive seasons in baseball history to be average or above.

3) Maybe something like 3 or 4 WAR difference would be enough for me to say that one player was certainly better than the other. I’d think of it more in probabilities, something like a .5 WAR difference means one player has a 10% chance of having been better, a 3 WAR difference might be 90% or whatever.

Good questions all.

1) Assuming your WAR calculation was based on the true run value of defense then yes. The problem is with UZR and other defensive metrics, not with run prevention. Then again, for this thought experiment, if you’d chosen a hitter who generated value with walks and home runs, the defensive advantage would be mitigated.

According to your belief – and just your belief – what percentage of a baseball game is your team’s defense? So for example, offense is ~50% of the game (since when you’re up to bat you’re neither pitching nor playing defense, accounting for 50% of your team’s output). So of the remaining 50% of the game, what percentage is pitching and what percentage is defense?

I’m open to these being wrong.

1) Maybe it’s just me, but I don’t think it’s possible for a team of excellent defense, poor offense players to beat a team of excellent offense, poor defense players 50% of the time. There are things like chaining to consider (getting multiple hits/OBP in a row to keep an inning going), the role of pitching when compared to defense (even Manny Ramirez-like defense is going to make most of the outs in a game), and walks/home runs – which are not balls in play and thus not included in defense. It’s possible that I’m wrong, but I find it very hard to believe that the team of Manny Ramirez like players wouldn’t win most of the game despite their WARs being equal to the Jack Wilson team.

Side Note: One possible reason for this: You always need to score at least 1 run to win a game. No matter how good your pitching and defense is, you have to have at least some offense. And indeed, because an average pitcher would still give up a home run or two, even if you have amazing defense they would still need to come up with 2 or 3 runs a game to win. Because of that it’s possible that offense may be worth more than 50% of a game. But that’s just a passing thought.

2) Two parts to this. First, once again, I think it’s hard to imagine that any .500 OPS player could ever make up for his terrible offense with his defense, no matter how amazing a defender he may be. Gutierrez may be the greatest center fielder in the game, and – cognitively – I don’t think it’s possible for a player to make up for being that bad at hitting, let alone a net positive, which WAR believes he is. Even if his UZR is accurate, something seems amiss.

And to illustrate that, most people here agreed that defense is probably about 10% of a baseball game. Yet Gutierrez has had several seasons where he’s been worth – as has been argued not only by UZR but also by just watching him play – 20 runs, or 2 wins, sometimes even more. That is the equivalent of being worth 10 wins on offense (Offense being 50% of a game, defense only 10%). Since 1950, only Barry Bonds has been worth 10 wins on offense. Guti’s defense is exemplary, but it’s not once in a generation.

This is where Kermit and I disagreed. I think defense, while it’s undervalued in baseball, may be massively overvalued in sabermetrics. Kermit thinks that it’s not overvalued, but it needs to be adjusted more for park, where Gutierrez in Safeco is worth a ton, but Gutierrez playing in Boston is worth much less.

Either way – and I believe it was mentioned by Poochie in another thread and something I believed for a while – the way defense is added into WAR (even if UZR is “accurate” for what it’s supposed to be measuring) seems wrong, whether it’s because of park adjustments or defense simply being overvalued as an asset in sabermetrics, at least in terms of how much a player helped a team.

3) Most people agreed that it sounds fair to compare people at ~1 win. Even still, there may be a player that UZR has valued incorrectly low, and a player that UZR has valued incorrectly high, and it’s not too tough to see how a player worth, say, 3.5 wins may actually have had a better season than a player worth 5 wins.

3.5) In my eyes, UZR is singlehandedly destroying the validity of sabermetrics with much of the general public. It makes WAR nonsensical, it makes a team built on defense to seem like it should be better than it is (I still think defense is undervalued, but in baseball – not in stats), it makes a less-than replacement player look like an asset, and it makes a player like Jack Wilson in 2007 seem equal to Manny Ramirez in 2009 who – while he may be a butcher in the field – would still probably win more games if a team of him went up against a team of Wilsons.

Anyway, these are all my own opinions and I’m not trying to argue they’re fact. It’s just some thoughts I’ve been having on all of this and it didn’t make sense to write them all out without seeing what other people believe. Hopefully this type of stuff is interesting to others even if you disagree.

1.) Theoretically, yes, I could find it plausible since Jack Wilson wasn’t garbage offensively that year. More hits would drop for Jack since Manny wasn’t good defensively, so Jacks would probably be able to score around 4 runs a game, which would be able to match Manny, because he wouldn’t hit a dinger in every game and I wouldn’t doubt that a team full of Jacks could rob a few line drives to keep Manny around 2-3 runs per game that doesn’t have homeruns. This is just what I think (which is also probably wrong!) So I could see team Jack winning 50% of games, but I think team Manny would win more, but not alot more.

2.) No. I think he can be replacement level with a .534 OPS, but not a net positive. I think if he were around a .600 OPS he could be, just because I think he’s the best defensive CF in the Majors right now, but I wouldn’t really feel confident that he would be average overall unless he was around .650.

3.) For 1 season 1 win would probably be a big enough gap for me to comfortably think one player was better then the other.

I think a problem with really trying to figure out how good someone is defensively is the fact that all ball parks are different in basically every way. Since all the parks are different you have to grade all the players differently, which leads to alot of mistakes. It’s nearly impossible to judge the fielders then even when adjusting for park factors (especially when they change from year to year, so a player like Jason Bay who was thought to be awful in 2009 defensively by UZR had his numbers changed to reflect non awful defense). So I guess eye test + a combination of whatever defensive metric you like the most is for now the best way to evaluate a defender.

1) You mention 9 Mannys, 9 Jacks, and an equal pitcher, which I assume means we are talking about an AL team. The one position where you get offensive Manny and offensive Jack without the defensive counterparts is pretty significant, and is something I wasn’t figuring in. On an NL team I definitely like Team Jack better. On an AL team I might have to take Team Manny.

2) I understand that in your premise you stated that we are assuming both teams stay healthy, which makes Manny around 2.7 WAR but also makes Jack around 3.1ish (which you didn’t factor in). However, I don’t think it’s safe to assume equal health here. Wouldn’t you naturally expect the bodies of the more athletic defensive minded team to handle this much better than the less athletic offensive minded team? Isn’t a season at short, catcher and centerfield going to destroy those Manny’s physically, whereas the seasons at first, left and right should be a relative breeze for those Jack’s?

Even the guys at Fangraphs question UZRs validity, but yet they still use it and WAR as some sort of fact in their writings. I just don’t get the point of using it when it takes 3 years of recording to get a true measure of a players defensive abilities when you could probably get a better sense after watching that player actually play for a month or so.