THE BOOK--Playing The Percentages In Baseball

He also said his fielding took a step back because of his injury.  His UZR, starting from 2006:
+12
+11
-10

So, fantastic call on his part.  The data reflects how he actually said his defense went. 

Unfortunately, the Tigers fans were blind to this change, as they continued to rate him one of the best-fielding CF in baseball.

I was also pleased with Granderson’s answers.  One thing that strikes me about Grandy is that he is very good at identifying his weaknesses and then correcting them. Two years ago, he talked about cutting down his strikeouts.  Last year, he talked about the need to hit better against left-handed pitchers.  Each time, he has made improvements.  It will be interesting to see what happens with his defense this year.

Lee

Unfortunately, the Tigers fans were blind to this change, as they continued to rate him one of the best-fielding CF in baseball.

So how much of the fans ratings are really current observations and how much is based on past observations and accumulated reputation?

If Granderson is typical of his peers

It has always seemed to me that Granderson is atypically intelligent.

It’s tough to say.  In some cases, like Junior, it takes a couple of years for reality to kick in.  In cases with some transience, like an injury, fans will basically not really recognize that impact.  If he says a Chuck Knoblauch, fans will respond quickly.

Asking a fan for his snapshot judgement of a player he’s seen play alot, and asking to limit your view of him of what you actually saw of him within the last 4-6 months is certainly asking alot.

If we were able to poll fans on a game-by-game basis (similar to those polls you see during election debates), we’d be in a better position to capture what they think they see. 

All I can say is that there is bias, and we have to be aware of the biases, regardless of whether that data is collected in subjective or objective fashion.

I agree that there are problematic biases in the Fan Ratings, which is one reason why combining them with objective measures, as well as being aware of what those biases might be and trying to correct or adjust for them, are good strategies.

On the other hand, what the Fans give us - presumably - is more like what the scouts give us when they say that a batter has quick hands and good plate discipline, or something like that.  If a player like Granderson has a “bad” year on defense, and has a UZR of, say, -10, yet he still has excellent defensive “skills,” how do we want to characterize his defense for that year?  Good or bad?  I am not sure I have the answer, other than it is helpful to know when a player has a bad year performance and numbers-wise, yet also has retained his good defensive “skills” as determine by some “scouting-like” entity.  Or vice versa.

In Grandy’s case, maybe he actually was no worse on defense in 2008 than he was in 06 or 07 and the fans recognized that.  Maybe he just didn’t get to a bunch of balls that he normally would have, for no particular reason, or did not make his usual number of spectacular catches (however many that might be), again, for no reason in particular, and maybe Grandy was aware of that (that he just didn’t make the plays he normally does).  Again, maybe it was just a random fluctuation with little or no predictive value.  If that is the case, then the Fans rating for 08 (good defense) is a lot better than the numbers (like UZR) or even the opinion of the player himself. I mean, if an offensive player has an unlucky year where lots of line drives are caught, nothing is falling in for a hit, and he has lots of near-home runs (etc.), he might not say, “I just had an unlucky year.  I am still a great hitter.” He might genuinely think that he sucked on offense and the numbers will reflect that sentiment.  That can happen on defense too.  That definitely happens with pitching.  Do you think that Cliff Lee is saying to himself, “Yeah, I pitched well this year, but not nearly as well as my numbers suggest.  I am definitely due for massive regression next year?”

Ditto mgl/6.

If you were getting fan ratings that said that Granderson’s reaction time, throwing ability, routes, and footwork were all excellent (or whatever rating they had given him in previous years), but that he didn’t seem to get to quite as many balls after his injury, then I would agree that you were getting useful information not available from current numerical analysis.  But if the fan consensus was that Granderson’s 2008 fielding was just as good in all respects as his previous years’ fielding when the player himself and the consensus of numerically based metrics said that he was not, then I don’t see how you can conclude that you have gotten anything useful from the fans ratings. 

Could the distribution of balls hit to Granderson in 2008 been that different from previous years to cause a 20 run (play?) swing in UZR.  Possible I guess, but very doubtful.  Plus, although it is very unusual to have a player’s assessment of his own performance, I would always trust that assessment over an outside observer’s.  On offense a player knows when he is seeing and striking a ball well and when he isn’t, and on defense he knows when he isn’t getting to balls that he has been capable of getting to in the past or when his throws are not as accurate or powerful.  What a player is incapable of knowing is when his skills have passed their peak and he will never be as good as he was in the past.

”...then I don’t see how you can conclude that you have gotten anything useful from the fans ratings.  “

The uncertainty level around the Fans’ opinions is less than the uncertainty level around a season’s worth of UZR.  If sUZR says that Andruw Jones is +110 runs in 5 years while bUZR says that Jones is +0, then can’t I use your same statement?

Here are Vernon Wells’ bUZR/150 from 2002 to 2008:
-11
-15
+11
+3
+8
+1
-22

If I use this as my sole example, can’t I say the same thing as you sentence?

What if we look at Jermaine Dye from 2002-08:
-12
-7
+13 (!!)
-1
-16
-17
-16

What are we to say of Dye in 2004?

In his rookie year, Zimmerman had a BUZR of +5.  A decent showing.  But the Fans were all over him, proclaiming him, in his first year, as almost as talented as the Rolen/Chavez/Beltre trio that dominated MLB until then.  What to make of it?  He was drafted very high.  He broke into MLB very very fast.  And scouts were comparing his fielding to Mike Schmidt.  All that means something, to some extent or other.  If it wasn’t for the Fans, you would have to regress his +5 DOWN to something like +2.  But, BECAUSE of the Fans, you really need to regress it UP to something like +7 or +9 (or whatever).  And the next year, he went +16 and the year after (last year) he was +9.

As I noted, there is bias in everything.  Just because you can’t account for every bias doesn’t mean that everything is not useful.

These are simply the exceptions and not the rule.  This applies to the Fans’ ratings as it does to UZR and every other sampling metric in existence.  Our job is to report on the level of uncertainty at various number of samples.

The Fans provide value, and given the choice between the rookie evaluations, as a group, of bUZR or sUZR or the Fans, I’ll choose the Fans every single time.

Tango - I wasn’t talking about the fans ratings in general, I was talking about the fans ratings for Granderson in the specific example that I described.

I would never conclude that the fans ratings in general don’t have any utility.  As your examples show they can be particularily helpful when there is little statistical information or when the data providers disagree.  And as MGL stated they also provide a natural buffer to year to year statistical fluctuations due to small sample size. 

My post was specifically about Granderson and the relative value of what Granderson said about his own fielding, what the numerical analysis said about his fielding and what the fans said about his fielding.

Sorry Peter, I read your post differently.

It’s hard to say why the Fans evaluate as they do: is it more “toolsy” or more results-based in that time frame.  So, you are correct that we don’t know if we’re getting anything useful out of the Fans, in this specific case.

If a player has an injury or series of injuries that are public knowledge it is quite posible that this knowledge makes the fans a little biased.

Even if it is only on a subconsious level I would belive there is a significant chance that the fans would cut a player some slack on bad plays and/or give him extra credit on good plays if he is batteling an injury and playing hurt. Espesially if it is a likable player that they percive as “taking one for the team”.