Manny to St. Louis

[KowboyKoop]

Of course.


There isn't as much fluctuation from system to system as you seem to think. It does exist for some players, and yes there's more fluctuation than offensive stats, but for most players, the advanced systems do peg the players at roughly the same level. And from year to year? Yeah, offensive stats do that too...


The +/- system does judge every ball on it's path, how hard it was it, speed, etc. That data, plus the video analysts, is used.

FRAA is based off the basic stats - putouts, assists, errors, etc.

And here's an introduction to UZR, from 2003. The system's been updated since.

I'm not really sure exactly what your point is. Is it to not look at the stats? We definitely should look at them. Is it to not trust them as much as we trust offensive stats? Yeah, of course. I'm just not sure what you're objecting to. It seems every time somebody mentions advanced defensive stats, you make it a point to post and say this same stuff you're saying now, if I recall correctly. I'm not sure why.

I've seen and read all that. I'm not saying we shouldn't look at any of it. I'm saying the data and methods they use aren't even close to complete enough for defensive metrics to be a reliably accurate judge of defense.

Offensive stats do fluctuate from year to year, but you can identify the reason why. Increased BABIP, walk rates, HR/FB rates, LD%, things of that nature. You can't do that for defensive metrics. The data for tracking every single batted ball, and the position of the fielder on every single play, the effects of the situation on the fielder's anticipation/instincts/first-step, arm strength, skill at turning the double play...there are so, so many more factors that go into fielding than hitting..at least in terms of measuring a player's ability. It's just far too incomplete right now. Again, that doesn't mean they are a waste of time...it just means that they are in the early stages of development or being perfected.

I talk about why defensive metrics are somewhat inaccurate/incomplete b/c so many people just throw out a number from one system and conclude "thus, player X is a bad fielder." That is generally inaccurate, thus I say it. That's why we're all here. Nowhere am I insulting anyone or offending anyone. The same topics get discussed over and over..so you're gonna get repeat replies/thoughts. Just the way it goes.

[HoustonGM]

Right. I just don't see why you have to chime in with that every time a defensive metric is mentioned.

[KowboyKoop]

Right. I just don't see why you have to chime in with that every time a defensive metric is mentioned.

I don't see why you have to worry about why I talk about what I talk about. You "chime in" in lots of threads and have generally the same viewpoint in one thread as you do in another when talking about similar topics. If someone talks about defensive metrics, I'll give my thoughts on defensive metrics. If someone talks about BABIP, I'll talk about that. If someone brings up the KC Royals, I'll talk about that. If someone talks about Keith Law wanting to move a 19 year old 3B prospect to catcher because of a strong arm, I'll talk about that.

[filihok]

Experiment

[200tang]

Honestly I doubt defensive metrics will ever be as good as offensive, but I agree with HGM. They still peg players close to how well they really field. And really if you're going to talk about defense with players, what else do you have to look at besides your own personal bias? Still not saying you should rely on defensive metrics, but I'm stumped as to what else you can look at and show other people why you think player a is better than player b.

[KowboyKoop]

Honestly I doubt defensive metrics will ever be as good as offensive, but I agree with HGM. They still peg players close to how well they really field. And really if you're going to talk about defense with players, what else do you have to look at besides your own personal bias? Still not saying you should rely on defensive metrics, but I'm stumped as to what else you can look at and show other people why you think player a is better than player b.

If you think that defensive metrics aren't as good as offensive metrics, then how can you say they peg players close to how well they do on the field? One system saying a player has a good UZR doesn't make him a good fielder, it means that according to UZR he has done well in what that calculates.

I think you should be able to completely eliminate personal bias. Offensive numbers have nothing to do with personal bias..a hit is a hit, a HR is a HR. The numbers are the numbers. Fielding is completely different. The data just isn't as complete right now.

As for what else to look at, I think we just need to wait until we can have more complete data on every batted ball and everything else on the field before we have a better picture.

This is the first step in that process.

http://www.fangraphs.com/blogs/index...tart-of-hit-fx


If defensive metrics had a lot of merit, then why would this be such a breakthrough? The development of hit f/x technology, which apparently will take a lot of time (5-10 years??) is what will truly lead to a breakthrough in defensive metrics. Until then, I don't put much merit in what the current numbers say.

Again, that doesn't mean all the work done up to this point is worthless.

[HoustonGM]

Defensive metrics, when taken in conjunction with one another, do give a decent picture of how well a player is in the field. Are they close to offensive metrics in terms of their accuracy? Not at all. Is there much more work to be done? Absolutely (and Hit f/x is a step towards that, just like Range Factors were in the 80's, and Zone Rating, and Fielding Win Shares, and +/- all were steps as well).

As for what else to look at, I think we just need to wait until we can have more complete data on every batted ball and everything else on the field before we have a better picture.

Are you saying that we shouldn't look at the current defensive metrics when trying to evaluate players? I don't think that's what you're saying, but that's what this statement sort of sounds like so I want to clarify.

[KowboyKoop]

Defensive metrics, when taken in conjunction with one another, do give a decent picture of how well a player is in the field. Are they close to offensive metrics in terms of their accuracy? Not at all. Is there much more work to be done? Absolutely (and Hit f/x is a step towards that, just like Range Factors were in the 80's, and Zone Rating, and Fielding Win Shares, and +/- all were steps as well).


Are you saying that we shouldn't look at the current defensive metrics when trying to evaluate players? I don't think that's what you're saying, but that's what this statement sort of sounds like so I want to clarify.

That's all I've been saying. You may not like that I've used the word "suck," but I clarified that I think they "suck" compared to offensive metrics. I've said like, four times now that I don't think defensive metrics are a waste of time. Of course they can tell you that Grady Sizemore is a better outfielder than Pat Burrell, but I don't think they can be relied on for a whole lot more. Look at them if you want, but you have to put them in context. Just like the best pitchers tend to have better W/L records, W/L record is far from the entire story. Same with defensive metrics. The best fielders are generally going to have better numbers than the worst ones, but they aren't good for a whole lot more than that. Now, for pitching metrics..we have a lot more detailed info we can go to. With defensive metrics, we don't, not right now anyways. Thus, for pitching, you can glance at W totals within the context of all the other numbers. For fielding metrics..you can't do it, so yes, you can look at them, but you should realize that there is a lot more to it that we just don't know yet. They are very shaky.

The statement of mine you quoted was in response to someone asking what else I would look at. There is nothing else to look at right now, so I said we should wait for a lot more data to be compiled. Nowhere did I say that we should ignore the current numbers completely or anything like that.

[KowboyKoop]

and before anyone says this, no I am not saying that defensive metrics are equivalent to W/L totals for pitchers..it's just a comparison.

[filihok]

[side note]
I think that hit-fx will do a lot for pitching and hitting stats as well.

Defense is important in those two aspects as well
[/side note]