Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 11/03/84 (WLS Mods); site fisher.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!princeton!astrovax!fisher!david
From: david@fisher.UUCP (David Rubin)
Newsgroups: net.sport.baseball
Subject: Re: New Statistic?
Message-ID: <667@fisher.UUCP>
Date: Tue, 11-Jun-85 20:21:00 EDT
Article-I.D.: fisher.667
Posted: Tue Jun 11 20:21:00 1985
Date-Received: Wed, 12-Jun-85 20:16:34 EDT
References: <3900007@csd2.UUCP> <3900008@csd2.UUCP>
Organization: Princeton University Department of Statistics
Lines: 121


Continuing the discussion of Isaac's proposal for "runs saved"...

>> ..........  First, it is situational.  Unless someone can demonstrate
>> the existance of "clutch" fielding, it would appear to me that a great
>> fielding play would occur at both critical and non-critical junctures
>> of a game with equal likelihood.  Thus the statistic would to a large
>> degree measure random factors (who got the ball hit at them at the
>> critical juncture) rather than skill, which is what we are presumably
>> trying to measure.

> I think clutch fielding is actually a lot easier to demonstrate than
> clutch hitting. In a situation where the game is on the line, fielders
> will clearly try to make plays that they wouldn't ordinarily.....
> ............................................... Now, imagine:
> man on third, score tied, bottom of the eighth inning. What are the
> chances that the right fielder will dive for the ball ....?

Your argument satisfies the first part of my first objection: it is likely
there is such a thing as "clutch" fielding, as it is obvious that some
players will field differently with the game on the line (unlike
hitting, where it is unlikely that players hold back).  However, it is
does not follow that "runs saved opportunites" will balance out over
the course of a season, even for players playing in similar
environments.  For example, it is well documented that two pitchers
pitching for the same team can receive well over a run's difference in
average run support over the course of an ENTIRE season.  As runs
saved will be of a lesser order of magnitude than runs scored, it may
suffer the same problem (random luck) as pitcher won-loss records.
Indeed, as the runs saved will be scattered over many positions, I'd
be very surprised if it didn't.

>> Second, it is heavily dependent upon the team the player plays on,
>> especially the pitching staff.  It favors players who play on teams
>> with poor pitching, as they will have more fielding opportunites and
>> more of those opportunities will occur with runners on base and in
>> scoring position.  It would favor first basemen on teams with
>> predominately right-handed staffs, for example.  It would favor
>> infielders on teams with a few sinkerballers.  It would favor center
>> fielders who play in big ballparks. Etc.

> There certainly is some randomness about who gets the ball hit
> to him in the crucial situation, but I think it would even out
> to some extent over the course of the season, at least for each
> individual position ...........................................

See my comment above.  To elaborate, as the average number of runs
saved per player per 4-5 games will be less than the average number
of runs scored for a pitcher per start (this is becuase there are more
runs scored than runs saved, and fewer starters to divvy up the former
than there are fielders to divvy up the latter), most players will 
have fewer runs saved over the course of a season than a starter will 
have runs scored in support of him over the course of the season.  We
know the latter don't even out, so the former will not, either.  This 
effect is RANDOM; it is independent of the objection to environmental
BIASES that followed.

> ............................  What you say about the pitching
> staff affecting the number of opportunities a fielder gets for
> runs saved is also true, although I could point out that this
> is true of errors as well.......................................

It is, but "errors" are misleading, too.  I'm certainly not defending
the old defensive statistics (except for careful use of putouts and
assists).

>> Third, it is judgmental..........................................
>> [S]ince there will supposedly be relatively few runs saved awarded, such
>> [scorers'] biasedness will be tremendously magnified.

> This is true in part. However, I think you're mistaken about how many
> runs saved would be awarded. It's my impression that two or three in
> a game would not be uncommon. Let's try an experiment along the lines
> of watching three or four games and seeing how many occur.

Fair enough.  I'll make a note of "runs saved" for the next couple of
weeks.  Other readers are encouraged to do the same, noting date,
teams playing, "runs saved" and by whom.  If you send me the results,
I'll gladly share them with all.  Remember, only credit outstanding
plays.

>> Fourth, it will encourage hot-dogging.  As the league leaders will be
>> rewarded financially for their accomplishment, there will be a
>> financial incentive to make plays look hard........................

> I don't think this is true at all. I just don't buy the idea that a
> fielder will, for example, try to make a diving catch and risk missing
> the ball altogether instead of making the play as well as he can and
> risking not getting a run saved. This is especially true in the
> situations where doing so might allow runs to score, which is
> presumably when a run saved might be awarded.

You are probably right for teams which are in contention and in the
game at hand; I am probably right for teams out of contention or out
of the game.

>> Fifth (this is the most critical, but also the most opinionated), I
>> believe it will punish true defensive excellence.  We already see this
>> in fielding percentage, where a Hernandez is charged an error on a
>> ball where a Garvey would never have been able to even attempt a play.

> This I disagree with the most. The whole idea behind this statistic
> would be to recognize the fielder who makes the game saving plays,
> as opposed to the one who has the range of a carrot but never drops
> the ball once he catches it. I would be willing to bet a six month
> vacation for two in Bora Bora that Hernandez would have more runs
> saved in a season than Garvey.

It's judged by the same guys who judge errors, and their going to
charge Hernandez for "runs allowed" on plays Garvey couldn't even
try, and award Garvey "runs saved" on plays that would be routine for
Hernandez.

(To Garvey fans:
I'm not beating up on Garvey because he's a BAD fielder (he's about
average for a first baseman, I think), but because he is the
beneficiery of the inadequacy of the fielding percentage statistic,
and therefore the example par excellence of statistics ill-chosen.)

					David Rubin
			{allegra|astrovax|princeton}!fisher!david