Making Fielders Accountable Through Stats

Did the Phillies really earn seven earned runs off Dontrelle Willis Monday night? The box score says so, but anyone who watched the game knows Dan Uggla was to blame for at least one of those runs. As the great statesman Benjamin Disraeli once said, "there are three kinds of lies: lies, damn lies and statistics."

Baseball fielding statistics have never been as popular as batting and pitching stats, probably because fielding stats aren't nearly as effective at painting a picture of a players true performance. First of all, counting up errors shows only half the picture. A particularly aggressive fielder might save a lot more runs with great plays than he costs with errors, though you may never know it by simply looking at the mistakes.

How about a keeping a new fielding stat for web gems called "gems"? If SportsCenter can pick ‘em then so can an official scorer. Fielding percentage is fine, but you could tell a lot more with a gems-to-errors ratio than with errors alone.

You can take this one step further. By keeping track of gems, you can evaluate a fielder based on runs saved, which is, after all, the name of the game on D. You might also begin to compare the number of runs allowed by errors versus runs saved with gems to derive a kind of fielders ERA. This would enable you to keep fielding stats separate from pitching stats.

I've always found it awkward that after an error a pitcher can give up a long home run like Billy Wagner did to Barry Bonds (number 711) but get a free pass on his record. Meanwhile, if a fielder makes a spectacular play the pitcher gets all the credit. Sure, if David Wright doesn't botch the previous play then Wagner gets out of the inning, but Wagner still served up the dinger. And if the other team scores another 10 runs after the error is it really accurate to say that none of them were earned? Look at it this way: unless it's the ninth inning, the pitcher is still going to have to get those hitters out in the next inning. I understand that getting three outs is what matters in an inning, but does that mean that an error only affects the inning in which it occurs? Not really. Let's say you make an error in the eighth which moves the lineup and allows Albert Pujols a chance to bat in the ninth that he wouldn't have had if an error wasn't made. Pujols might hit a home run that he never should have hit if the error wasn't made in the eighth. This may seem like nit-picking, but it really isn't, since an ERA is supposed to be an expression of how many runs a pitcher gives up over a nine inning game - never mind that pitchers just don't pitch nine inning games anymore. It just means that ERA really doesn't tell the story, especially for relief pitchers. The best stat out there for measuring pitchers is WHIP, or walks plus hits per innings pitched.

These days super-statisticians like Bill James do all sorts of nifty mental gymnastics trying to twist the classic fielding and pitching stats into something that will more accurately reflect the game as it happens on the field. In the end it doesn't quite work for fielding. On the pitching side, Sabermetricians call this DIPS, or defense independent pitching statistics. Among the more complicated efforts are dERA (defense independent ERA) or DICE (defense independent component ERA) or FIP (fielding independent pitching). Essentially they all try to do the same thing, which is measure the effectiveness of a pitcher for the things which are truly under his control. Admirable though the effort may be, it's a little like the astronomers before Galileo creating all sorts of mechanical adjustments to explain the motion of heavenly bodies because they couldn't accept that the Earth was not at the center of the Universe.

The simple truth is that current fielding stats are insufficient and the holy stat called ERA is flawed. Pitchers ERA was invented before fielding stats and the only real way to develop accurate fielding statistics is to disembody the errors from the pitcher's line. One way to do this would be to charge fielders with runs due to errors and continue to hold pitchers responsible for what happens after that.

If a pitcher has a terrible fielding team behind him, then the fielders would get those "unearned" runs on their fielding ERA. If a pitcher has a terrific fielding team behind him, you could compare his pitching stats to his fielders ERA to show how many runs the fielders saved behind him: runs allowed minus fielders runs saved. This would give a clearer picture in the box score of how the game was played. For example, when Barry Bonds hit his 711th home run, a two-run shot off Wagner, David Wright would be charged with one run but Wagner would be charged with the other.

As to reconciling this with the way stats have been tallied in the past, well, all I can say is that once upon a time a walk was counted as a hit but we got past that one, too.


Philly Baseball Insider Top Stories