What bothers me about these ratings is this: There is some positive correlation (r^2=.13) between predicted DER and the final rankings. Predicted DER is intended to be a measure of how easy the position is to play; the higher the predicted DER, the easier the balls in play are to field. It makes me wonder whether David is getting enough of the dependency on the degree of difficulty on BIP out of the equation in order to accurately assess the fielding skills of the SS - admittedly, not an easy task.

I am fully aware that (a) the positive relationship isn't *that* large and (b) the range of performance in ratio is relatively small, particularly in the middle of the group. But is it entirely coincidental that only three SS had it easier than Tulowitzki, and only one had it harder than Guzman (and only three had it harder than Jeter)? I don't know.

-- MWE

There is some positive correlation (r^2=.13) between predicted DER [for players] and the final rankings.

Not so much at a team level? Why would it be different there?

Am I the only one who gets a blank page at that link?

Am I the only one who gets a blank page at that link?

It worked for me.

This seals the deal in my opinion. Ryan Theriot is not a viable option for a major league team at SS. My impression was that his defense was solid (good range, weak arm) which compensated for the weak bat. Ronnie Cedeno is the superior option and I no longer think it's close.

Am I the only one who gets a blank page at that link?

Nope.

Fair enough, Mike. I still felt that it's worth linking. From what I understand, some folks feel that the expected DER can be skewed by a team's hitters hitting exceptionally poorly or well. I think that David is looking into other ways to come up with expected DER.

I'm not sure what the problem is, guys. What brower(s) are you using?

I think Twins have historically scored well on this metric, even when the Twin was Guzman. I think it's the turf.

I am fully aware that (a) the positive relationship isn't *that* large and (b) the range of performance in ratio is relatively small, particularly in the middle of the group. But is it entirely coincidental that only three SS had it easier than Tulowitzki, and only one had it harder than Guzman (and only three had it harder than Jeter)? I don't know.

I did a quick plot of Expected vs. (Expected - Actual) to get an idea of some degree of heterogeneity in the residuals. There does seem to be a bit of funnelling, that suggests a bit more residual variability at the higher Expected BIP's... I guess this is probably expected in the logistic regression though, because of the different numbers of Total BIP's (from a simple binomial model, the standard deviation of the fitted value is going to increase at something like sqrt(n p (1-p)) where p is the probability of getting the ball. One sneaky way to do this would be to get standard errors for each guy's "true" ability from the logistic regression and include error bars (and maybe even try to "standardize" things). When I normalized the deviations by sqrt(BIP) (which is probably not the right thing to do, but the first thing that I thought of), I get this ordering ( for (Actual - Exp)/(Exp*sqrt(BIP)):

Mike/#1: Predicted DER does not tell you only, or even mainly how difficult/easy a player's chances were. It tells you what proportion of all BIP that player should have turned into outs. So Tulo has a high DER because CO pitchers had a 46% GB rate (highest in NL), and Jeter is low because NY pitchers had a below-avg GB rate (43%). Degree of difficulty also plays a role, but looking just at DER you can't disaggregate it from the total number of balls hit into a player's general range.

To the extent there's a correlation, perhaps it shows that teams with GB pitchers have better shortstops, which would be interesting. But I'm not sure it tells us that.

Explorer.

"I had a chance to hear David Pinto speak last month and I hought that he did a great job explaining PMR to an intelligent layman like me."

Corrected: "I had a chance to hear David Pinto speak last month and I thought that he did a great job explaining PMR to an intelligent a layman like me I."

Last years PMR thread

Rich, thanks for the catch on the misspelling, but I think that we both got the grammar wrong.

me had a chances two here David Pinto speaks last month and me thoughted that she do a grate jobs explaining PMR too a intelligent laymans like i.

Run conversions of the PMR figures. Including comments and caveats.

Positioning should effect some of these rankings, right?
I was thinking about this. If we split the field into small zones ala ZR, and see if a fielder has a zone of maximum efficiency, where we choose a zone in which the fielder has had maximum chances, and how his DER is within that zone. Maybe expand it to 2 zones..

I'm using Apex Interweb 1.1, and it works fine.

On the other hand, that model ranks the Brewers D 12th in MLB, so it can't possibly be right.

"I had a chance to hear David Pinto speak last month and I hought that he did a great job explaining PMR to an intelligent layman like me."

Corrected: "I had a chance to hear David Pinto speak last month and I thought that he did a great job explaining PMR to a layman like I."

I was taught that since you'd say "...explaining PMR to me" that adding the desciptor "a layman like" doesn't change the "me" to "I".

That sentence definitely calls for a "me." Nothing worse than a grammar cop who's sloppy with grammar.

I would say, "I had a chance to hear David Pinto speak last month and thought he did a great job explaining PMR to a layman, like I am."

Or, "I had a chance to hear David Pinto speak last month and thought he did a great job explaining PMR to a lay audience."

Nevertheless, this grammar site suggests Gary was correct all along:

Question
"Certified teachers (like me) believe that merit pay is a bad idea."
"Certified teachers (like I) believe that merit pay is a bad idea."

Is "like I" appropriate ... in this construction?

Reply
Since you want an answer that addresses levels of correctness, "like me" would be more correct. ... In your sentence, the word "like" is acting as a preposition and "me" is the object of that preposition. It is not introducing a clause of any kind.

My bad.

You're missing the point. He meant to say "lame man like me."

On the other hand, that model ranks the Brewers D 12th in MLB, so it can't possibly be right.


It ranks the Brewers D at shortstop 12th in MLB.

Hey what do you know Al I had a chance to hear David Pinto speak last month and he did not do such a great job of ex plaining 1 thing he must look at sprit sheets all day and never watch baseball so I dont know what all the banana oil is about whether he explaind it better to him and them than me and you could it was all Jibrush to me.

It ranks the Brewers D at shortstop 12th in MLB.

Yeah, I caught on to that, which does not bode well, as SS might be their best position (unless it is LF or RF).

It ranks the Brewers D at shortstop 12th in MLB.

Yeah, I caught on to that, which does not bode well, as SS might be their best position (unless it is LF or RF).
Page 1 of 1 pages

And the Brewers think JJ Hardy is gold-glove material too. That might be a big problem in of itself. I say JJ moves to 2nd in 3 seasons.

The language is kinda vague, but I think watching the Brewers play defense does qualify as torture under the GGC.

12th out of 30 isn't bad.

-- MWE

12th out of 30 isn't bad.

No, Hardy's fine, but that's only one position. And Hardy's numbers are skewed because Rickie is an error hog.

Yeah, if I read Pinto correctly, the Brew Crew are 27th out of 30 at the team level. I don't know how they look using other measures.

When talking defense, I think an easy thing to do in any thread is to post "Derek" Jeter" the way Norm Mcdonald used to insert "Frank Stallone" into his weekend updates.

I would say, "I thought he did a great job of explaining PMR to a layman such as myself." It's always much easier to simply switch to more pompous phrasing than to do the work of getting the grammar and syntax right.

Where is Levski?

Levski, there's a very good reason ZR and UZR don't account for IF popups and line drives.