So instead of an early season hitting slump he decided to have an early season fielding slump.

It’s weird, he’s looked pretty solid most of the year to me, but his ZR has been really erratic.  Could be an issue with the metric, not the player I guess.

That’s a purty graph. 
I forgot, when calculating the AL AVG 2b, are Robbie’s #s included?

I forgot, when calculating the AL AVG 2b, are Robbie’s #s included?

I really should remove them, but out of laziness I never set up my spreadsheet to do it.  It shouldn’t make a huge difference, but it’s probably worth doing at some point.

I’d love to see Jeter’s & Texeira’s graphs.

Couldn’t this have been presented more clearly?  Like in a pie chart?

I’d love to see Jeter’s & Texeira’s graphs.

Sure, refresh and you should see them now.

Couldn’t this have been presented more clearly?  Like in a pie chart?

ZR and pie charts don’t really work.  I could do a pie chart of fielded chances vs. missed chances but that wouldn’t show us the day to day fluctuation or the comparison to average.  I do have an idea for a pie chart that may work though, # of chances hit to each Yankee.

SS zone rating is biased against players dating Minka Kelly.

Is there anything to the appearance that Cano and Jeter’s graphs seem to negatively correlate with one and other?

Is there anything to the appearance that Cano and Jeter’s graphs seem to negatively correlate with one and other?

Since the zones don’t overlap, there shouldn’t be.

Here’s the zone breakdown for Yankee Stadium that STATS uses to calculate ZR.  The only difference between stadiums is foul territory and outfield wall distance.

“Since the zones don’t overlap, there shouldn’t be.”

Understood, but could it be the result of the defense shading a particular direction or depth?

Understood, but could it be the result of the defense shading a particular direction or depth?

That’s certainly possible.  The zones are fixed and don’t account for positioning, so if there’s a shift going on it could be a factor.  I haven’t noticed them playing a lot more shifts than normal, but that doesn’t mean they haven’t been.

How about putting up the total infield ZR as well to see if it’s been relatively consistent?  That might show it’s more positioning and less individual changes.  Although you might have to indicate when ARod came back as that was a fairly major change.

These graphs need more background color.  I can barely notice them on the page.

ZR and pie charts don’t really work.

I know.  I was kidding.  Somewhat.  If you can’t put it into a pie chart, it’s not worth analyzing.

Is it worth filtering the data so that only every day players are included or is that taken in to account already?  Basically, what I’m saying is it’s probably only right to judge every day players against every day players.  I’m not certain as to what the cut off would be, but around 85% of games played seems about right.

I suppose the argument against filtering for everyday players would be if you compiled everyone at 1st, 2nd and SS, the data would average itself out so that good and bad players with limited playing time end up around league average.  Would this be a correct assumption?

How about putting up the total infield ZR as well to see if it’s been relatively consistent?

Sure, I’ll post that in the next few days.

Is it worth filtering the data so that only every day players are included or is that taken in to account already?  Basically, what I’m saying is it’s probably only right to judge every day players against every day players.  I’m not certain as to what the cut off would be, but around 85% of games played seems about right.

Only problem is it’s a lot more manual work.  The spreadsheet I have now does all the work for me.  If I have to go in and prune out everyday vs. non-every day players it’ll take a non-trivial amount of time.  I’ve always assumed that backups should be close to average fielders since they generally can’t hit, but I could play around with it a little to see what the impact of removing them is.

I think part of the problem you’ll run into trying to determine who the “backup” is will be due to injuries.  E.g. if Jose Reyes only plays 50% of the time this year should you remove him?  Probably not.  And in some cases if a starter goes down midway through the season, the team will get a good player to replace the starter.  So now you get zero players on a team w/ sufficient time to make the cut, even if they had two “full time” players who were average or better at different times.

Maybe you feel differently SG, but I think if you go through the exercise you’ll either end up with a) removing so much playing time the sample size won’t be meaningful or b) setting the thresholds so low (to get enough data) that you’re just as well off including everyone.

Maybe you feel differently SG, but I think if you go through the exercise you’ll either end up with a) removing so much playing time the sample size won’t be meaningful or b) setting the thresholds so low (to get enough data) that you’re just as well off including everyone.

Actually, those are precisely my concerns.  Looking at 2B for example, only 8 AL 2B have played more than 73% of their team’s defensive innings at 2B, and only 10 have played more than 47%.

Wow, I didn’t realize so few players had logged that many games.  It certainly would be difficult to get a decent sample size in that case.  Is there anything to be said about the number of innings played versus quality of the zone rating?  I would imagine the number would be more representative, but as the season goes on, is it natural to assume a player’s zone rating will dip a bit, improve or drop off a cliff after a certain point like a pitcher on a pitch count?

Is there anything to be said about the number of innings played versus quality of the zone rating? 

I think what you’d see is that more innings correlates with higher ZR, because good defenders will play more.  Tangentially related to this, Chris Dial did some work on defensive replacement level in this article.

I would imagine the number would be more representative, but as the season goes on, is it natural to assume a player’s zone rating will dip a bit, improve or drop off a cliff after a certain point like a pitcher on a pitch count?

I don’t think there’s a fatigue factor that would manifest itself in a decline in zone rating as the season progresses, although I suppose it’s possible.  I think we should generally assume a player’s zone rating will regress towards their previously established level, although the fluctuations inherent in a single season could cause that to not hold true (easier chances, better positioning, better or worse health, etc.,)

Thanks, guys.  I’m still trying to get a handle on defensive metrics.