Printable View
Just because it's difficult to do accurately, doesn't mean you should give up trying. Basically, all you can do is use the best resources you have and make decisions based on that data. If it turns out to be wrong, oh well.Quote:
Originally Posted by dickay
And this is where sample sizes come in. If the question is about veterans that have a few years of defensive statistics - then we can be fairly comfortable in using statistics.Quote:
Originally Posted by dickay
If you want to tell me that Elvis Andrus is a fantastic shortstop - then the statistics cannot really support that yet. We have some evidence that he might be good, but the scouting reports are probably more important than the stats at this point.
Similarly, Prince Fielder was a bad defensive fielder who was statistically average last year. Did he become a better fielder - or was that an anomaly? The statistics can't answer that question - so you move on to a scouting report.
But the question is whether we can make meaningful projections of defensive ability. The answer is probably yes, but, especially for younger players, their defensive value should be regressed a long way toward 0.Quote:
Originally Posted by haveacigar
Not quite true. If a player has a bad fielding percentage for his position, it tells you right away that he has bad hands, or else makes a ton of throwing errors. If he's booting/throwing away the ball that often, I don't think anything else he does defensively is really going to make up for that; he's a poor defensive player.Quote:
Originally Posted by haveacigar
Unfortunately, the opposite isn't true. A guy can have a great fielding percentage but still not be a good defender if he has the mobility of a cigar store Indian and the throwing arm of a 3-year old girl.
I really really don't like defense, I prefer pitching and hitting and consider them much more important. Your defense can suck huge dick, but if your pitcher is great and striking lots of guys out, it won't matter. Similarly, you can have a shitty pitching staff and awful defense, but have a lineup of big boppers, and win all your games 13-12.
That's why I rank them this way, from most to least important:
1. Hitting
2. Pitching
3. Fielding
And of course, at a very very distant 4th, baserunning
Or, if a player has a bad fielding percentage, it means he is getting to more balls and making tough plays, sometimes going for errors. If a guy's range allows him to make 20 more outs then the other guy, but he makes 3 more errors, he's going to have a lower fielding %. But the 17 extra outs are worth the 3 errors.Quote:
Originally Posted by dps
Take Brandon Inge for example. His fielding % is probably far lower than Chipper Jones at 3B. He's also a way better fielder.
I think most people consider it in order of importance:Quote:
Originally Posted by metsguy234
Pitching
Hitting
Fielding
Which is why fielding costs way less on the FA market than pitching or hitting.
That said, I can't think of a WS winner lately that was poor defensively.
The way I look at it. Of course defense is important but it is not everything. We must remember that a player can't be just judged based on his defense but also his other skills. I don't believe either that defense should be weighed more heavily than offensive numbers either. I don't believe anyone thinks that though. I've just heard a lot of arguments from people, not on here, but people who I talk to and defense seems to start to be the more weighed stat now a days in people's minds and not even thinking of overall offensive production. This is just something I have been thinkin about.
The Yankees last year were 20 runs below average, the Red Sox were second last in the league in 2004 (and 35 runs below average) - so I guess it depends on your definition of poor.Quote:
Originally Posted by haveacigar
It's all about balance. That's why WAR is a good starting point for judging players because it at least attempts to quantify them and come up with an overall rating for the player. It's why I'd rather have Matt Holliday than Jason Bay if I were the Red Sox, but at any price? No, and certainly not at the silly price that Scott Boras is hawking his latest client who he can demonstrate to you is the greatest player ever to play the game, if you take the time to read through his snow job presentation. Please Mr. Boras, do us all a favour and save the f**king trees and let the teams do their own evaluations and come to their own conclusions. Your client will be paid what your client is worth, not what you think he's worth. Of course, all Boras has to do is find one idiot GM amongst the thirty and once again his ship will come in. Possibly (still early) teams are starting to come to their senses. If I hear the word collusion I will go postal. Since when is deciding to run a team more efficiently colluding with the other 29 teams? That is beyond the scope of this thread however. Hey metsguy, just how many of those 13-12 type teams have you seen win a World Series? Yeah, thought so. :DQuote:
Originally Posted by rschusta24
Thanks. I had a feeling I'd be wrong, but I didn't feel like doing any research. :oQuote:
Originally Posted by kenny1234
I was gone the whole day and missed out :(
Anyways, I've skimmed most of the thread and thought I'd throw some stuff out there.
- Defense is very important, however, it isn't THE most important thing to consider, obviously. It's a balance between all aspects of baseball. However, defense is very important and even though defense is starting to finally be valued it's still the least valued and the easiest to pick up on the cheap.
- One thing to consider about the correlation between pitchers and defense is take the defense COMPLETELY out of it. What would the pitchers K/9 be? 27/9. Defense affects a lot.
- Like everyone has said, it's bad to use just one season of data or ignoring other stats and scouting reports. Hence why the 'stats vs scouts' thing is complete ********. I'm not sure if there are any sabermetricians out there who think stats always tell the complete story. I will say though, that just looking through UZR/Tango's Fan Reports/+/- that they all generally do a good job at ranking players.
- Range almost always reigns supreme over other areas of defense (arm,err,DP), however, always keep in mind the position you're looking at and use common sense. Go over to fangraphs and look at Bobby Abreu who acquires almost all of his value in defense through his arm. That +7 with his arm saved him from absolutely terrible, but going forward I would expect that to be maybe a +2 at best.
- For players without lots of time in the majors, while some have said you should regress it toward 0, I disagree with that. If they come out of nowhere and put up some crazy +25 or something, then yeah, but if it matches up with scouting reports and Tango's Fan Report it's worth just assuming he's well above average. Take Michael Saunders for example, he put up a +19.6 UZR/150 in 312 innings in LF. Now, while I'm not about to declare him a +19 UZR fielder I will say that having watched him (against Gutierrez/Ichiro/Chavez/Langerhans which helps) and read scouting reports on him that I'd put him at a +8 UZR level. Always use common sense when evaluating players.
- Catcher defense is still a long ways away from being quantifiable. DevilFingers has a good article on it though and has done the best job I've seen at trying to quantify what a catcher does.
- First base is also another weird position. Some first basemen keep the ball more than others do. Generally you should avoid using UZR at 1B if there's a scouting report from a scout you trust.
- UZR DOES have some problems, most notably when a shift is on. Chase Utley gets some bonus to his UZR because Ryan Howard is a pretty terrible fielder. By moving Utley more toward Howard he makes more 'out of range' plays. Utley is still a gold glove fielder, however, but it's worth keeping in mind.
- Defensive metrics still have a ways to go and may never be perfect, but you should never just blow them off. Teams like the Red Sox/Rays/Mariners/etc.. wouldn't be using them and having success if they didn't work (Note : These teams DON'T use UZR; they more than likely have a nonpublic stat)
People talk about the Tango fan report as confirming the statistical results. I just think that the people that visit the Tango site are people that read sabermetric blogs and keep up with the statistics. They then go to Tango's site and parrot back the statistical results. The strong correlation is absolutely meaningless.Quote:
Originally Posted by 200tang
As an example, every Blue Jays fan that I know would tell you that Wells is a good to great centerfielder, mainly because he won a gold glove 5 years ago. Statistics say that he isn't, and the Tango fan report agrees with the stats, not with my admittedly anecdotal evidence of fans. The same thing with Jeter in New York - people thought he was better than he is until the data proved them wrong. The fact that he is now a below average shortstop in the Tango fan report is not proof that the stats are good - it is proof that the stats have affected people's opinions of players.
Since to my knowledge UZR is not used in BBM and Range Factor is.... is there much of a difference between the two? Im guessing Range Factor is the best available Defensive stat BBM has.
+7 arm is probably not to be repeated, mainly because it hasn't been that high before - but the bigger question for Abreu is the range. In the last 3 years, his range number has been -3, -29, -15. Given that variation, it is hard to say what he will be like in the field next year.Quote:
Originally Posted by 200tang
This also depends on position. Ichiro has played both center and right. In right, his arm is considered roughly average (even if that goes against my intuition) but his range is very good. As a centerfielder his range is about average, but his arm is very good. That makes sense when you consider who makes up the comparison group. But if you are considering him as a centerfielder, it would be wrong to discount the value of his arm just because you prefer range.
That is how most people see it, but it is probably actually in this order:Quote:
Originally Posted by haveacigar
- Hitting
- Pitching
- Fielding
- Baserunning
Quote:
Originally Posted by SirKodiak
Idiot.
1. Bunting
2. Grit
3. Mentoring
4. Leadership
5. Slap-hitting
6. Stolen bases
7. Quick at-bats to keep your pitcher ready
8. Lots of ass-slapping to show that you enjoy each other at teammates
9. Moustaches
10. Loyalty
-Dayton Moore
I didn't mean to make it sound like you should totally discount the arm, but I think generally you're going to go with the guy who has the better range than the better arm even if they grade out about the same in total runs saved.Quote:
Originally Posted by kenny1234
I also find Bobby Abreu to be a somewhat weird player. He profiles as someone with above average speed, but it doesn't seem to help him at all in the field. Sort of like Jacoby Ellsbury, although I think Ellsbury is better than the crazy bad -18.6 UZR he put up this year, but with that speed you'd figure above average. I understand there's more to fielding than just being fast, but you would imagine they'd at least grade out as average. Really helps show why playing defense isn't EZ MODE.
Three other things; one question and two additions.
- A lot of people will post the examples of UZR and +/- disagreeing on certain players as a reason that the metrics don't work, but I can never understand this. Usually they're pretty close to each other and in the event that they do disagree then you just need to learn how each one comes to it's conclusion. They're not the exact same, hence why they're not named the same, so it makes sense that they aren't always going to agree. For another example, take tRA and FIP. Both good stats that most people use and have no problems with, but let's look at Tom Glavine. He's been a guy who has always been looked at as a meh pitcher from 02-09 according to tRA. The best tRA he's ever put up at the ML level was 4.72. However, if you look at FIP it paints a much different picture. In that same span he's bested his tRA 4 times by .5-1 run, which is quite a bit when talking about pitchers. Rarely do people ever bring up the weird pitchers that these systems disagree on as a reason the metric sucks, so why defense? I think mainly because it's harder to quantify defense, but that's no reason to completely dismiss it.
- Another fun person to look at is Alfonso Soriano. Looking at '07 he compiled a +18.4 UZR, however, looking deeper in to the numbers reveals his entire value was tied in to his arm (14.3). I think a lot of that stemmed from the fact that he was new in LF and people tested his arm a lot.
Getting back to the Ichiro example, I think a lot of people still remember him throwing out Terrence Long at 3B. In recent years it seems fewer and fewer people have been trying to run against Ichiro because of his reputation. While Ichiro may not rack up as many 'kills' as other fielders, I wouldn't be surprised if he's in the top 3 in the league in stopping runners.- And my question, does UZR on fangraphs include stopping runners from advancing? I read an article about it in the last off-season so I know there's a way to measure it, but I'm just not sure if fangraphs relies solely on kills. Hopefully someone out there knows.
I'd put a higher priority on mustaches and a lower priority on stolen bases.Quote:
Originally Posted by KowboyKoop
11: Looking good that one time I was watchingQuote:
Originally Posted by KowboyKoop
Not only that, but if you use a defensive metric that uses range factors in any manner, it's likely that better pitching will actually yield worse defensive stats. Why? Becasue, generally, better pitchers will strike out more batters. The more outs the pitchers record by Ks, the fewer plays the position players will be making in the field--and since range factors are just plays made (usually expressed on a per 9 innings basis) that'll lead to lower range factors.Quote:
Originally Posted by HoustonGM
Yeah, definitely. Although, it wouldn't make the player 'worse', but more it would give them less opportunities. I assume that's what you meant, but just thought I'd clarify :).Quote:
Originally Posted by dps
Like I said before, too, defense also affects K/9. A crappier defense means more chances for the pitcher to record more strikeouts. Defense and pitching really do go hand in hand.
Actually, expanding more on this, I think it depends on the player. A really good player (Gutierrez/Jack Wilson/etc..) probably won't see much of a dip. They're so consistently good that they almost always make those extra plays. Moving down the spectrum we have the 'average' defenders. Average defenders would probably be hurt the most because they go from a negative to a positive UZR year to year. Then we have the absolutely terrible players (Jason Bay/Adam Dunn/etc..), these guys would actually BENEFIT from getting less opportunities. They rarely make good plays so limiting the amount of bad plays they make helps their UZR out.Quote:
Originally Posted by dps
This is actually a large reason, along with shifts, that teams create their own defensive statistics along with using scouting.
I said worse stats; I would have thought that made it clear. How many K's the pitchers get logically shouldn't have any effect of the fielder's defensive abilities. The problem lies in the stats we have available to measure defense. We have more tools to accurately evaluate defense than we used to, but they aren't as good as the tools we have to evaluate hitting and pitching.Quote:
Originally Posted by 200tang
If you didn't run a defense out there and the only way to get an out was to strike him out, his K/9 would be 27. So, yes, defense DOES have an affect on K's. That's obviously the extreme, but to say it doesn't is incorrect.Quote:
Originally Posted by dps
[Edit : I should clarify that the effect is probably small on most teams, but the less outs the defense makes the more chances of a strikeout.]
One of the reasons K% is superior to K/9Quote:
Originally Posted by 200tang
if you have a smaller sample size i don't see how it necessarily makes it worse. yes, errors or mistakes are compounded until a larger sampling is inserted but there are also less opportunities for those errors and mistakes to occur. what i was saying is more that good pitching makes what seems already to be a very large margin of error much larger. I'd be interested to see these "UZR" stats for teams that have had the top five pitching staffs the past few years. If all the top pitching staffs have relatively high UZR's it could mean that defense helps pitching, it could mean that pitching helps defense, OR it could mean that the stats are flawed.Quote:
Originally Posted by 200tang
Well, it would make it worse for better defenders, because UZR is a counting stat, and fewer chances = less opportunity to get a high score.Quote:
Originally Posted by dickay
that seems like a flaw right there. so a ground ball pitcher is going to make the OF ratings worse theoretically??Quote:
Originally Posted by haveacigar
Sort of, but luckily I'm not sure if there's any rotation/bullpen out there where this would even be a problem over the course of 3 years of data. Like has been said before, just use common sense when looking at defensive metrics.Quote:
Originally Posted by dickay
http://www.fangraphs.com/leaders.asp...n=2009&month=0
Last year there were only 15 qualified pitchers with a GB% over 50 (5 of them were at 50.X). Unless you have a team of Joel Pineiro's it's not really worth bringing up. Also, if it's anything worth noting (probably not), three of them were on the Cardinals and Colby Rasmus still managed a 13.7 UZR/150, Rick Ankiel 0.8 UZR (Across all three OF positions), Matt Holliday -0.9 UZR (SSS for all three, obviously)
I think that the UZR data compares each fielder to how an average fielder would have performed. So, for an average fielder, the pitching staff makes no difference. For everyone else, getting less chances than average would bias the results toward zero. A groundball pitcher will make good outfielders look worse and bad outfielders look better - but I'm really not sure of the overall effect.Quote:
Originally Posted by dickay
And how would you measure the "top pitching staffs"? Using what? ERA? Common sense says that a good defense will lower a team's ERA because they get to more balls and thus, less hits. I do not understand what you're trying to get at here.Quote:
Originally Posted by dickay
I see what he's trying to do, but unfortunately it's circular. You can't separate defense and pitching when they both significantly influence each other.Quote:
Originally Posted by HoustonGM
well, maybe i went about it the wrong way. HGM was stating that a handful of teams improved pretty much solely because of their defense. I said its kind of short sighted to put it all, or even most towards defense IMO and then got into the line that their pitching had a big part of it. yes it turns circular i guess, but basically this thread has proved what I had suspected. That these newage defensive stats have significant margin error and are still not a very accurate measure of a players defensive ability.Quote:
Originally Posted by haveacigar
When a team's pitching staff is pretty much the same, and the offense doesn't change much either, yet the team does significantly better and also made clear and obvious defensive upgrades... is it really a stretch to attribute a large part of their success to the defense?Quote:
Originally Posted by dickay
Defensive Efficiency is a very simple team defensive metric. All it is is the percent of balls in play that a team turned into outs. The 2008 Rays were 1st in the majors in DE at .710 (71%). In 2007, they were dead last in the majors at .656 (65.6%). They were almost exactly the same offensively in both years - 774 runs in 2008, 782 runs in 2009. Their difference was entirely on the defensive side of things (pitching/defense). They allowed 944 runs in 2007 but lopped off nearly 300 in 2008, allowing just 671.
Their starting pitching staff was pretty much the same except for the addition of Matt Garza. Their bullpen was overhauled. They allowed nearly 300 less runs. I think that it was due to a combination of the better bullpen and the much better defense. Turning 6% more balls in play into outs is a very significant amount - as evidenced by them skyrocketing from dead last to first in the league in that category. Maybe I'm crazy, but I think changing 6% of balls in play from hits to outs will cause a team to be significantly better even if everything else stayed the same...which, in the case of the Rays, is largely true. The only significant difference was the bullpen, and I don't think 2 or 3 relief pitcher upgrades will cause a team to allow 300 less runs than the year before. Their radically improved defense was unquestionably a significant reason for their turn around and this is honestly the first time I've ever heard anybody view that skeptically.
And, for the record, the other teams I mentioned:
2009 Mariners DE: .712 (2nd in the majors)
2008 Mariners DE: .682 (26th in majors)
2009 Rangers DE: .699 (6th in majors)
2008 Rangers DE: .670 (30th in majors)
2007 Rockies DE: .701 (6th in majors)
2006 Rockies DE: .684 (20th in majors)
Is there a stat that doesn't have a margin of error? I can only think of a few, like K%.
they are/were also a very young developing team. same players doesn't equal same performance year in and year out. a guy has a good year pitching and you want to attribute nearly all that to better defense behind him? I'll pass.Quote:
Originally Posted by HoustonGM
so did their defense improve that much, or are the stats flawed? not just when they did good, but when they did poorly? Is it proof of a small sample size? Is it improved pitching that helped the defense? Or is it truly that their defense improved that much and that alone improved their pitching and win totals?Quote:
Originally Posted by HoustonGM