Andruw Jones isn't nearly the player that Clemente was (and this is also what WAR says, Clemente leads Jones 94 to 62, so it's not even close). Clemente was a much better hitter (and had a longer career, even given, you know). Clemente did a much better job getting on base, even though he played in an offensive environment that was friendlier to pitchers than Jones did.
However, according to the data that we have, Andruw Jones was the greatest defensive outfielder of all time. He's 20th all-time in the defensive component of WAR (behind Brooks and a bunch of short stops, basically). Clemente is 160th in the defensive component of WAR.
Basically, the defensive component of WAR looks at every defensive thing that a player did on the field, asks how likely other players would have been to make the same play, and looks at, historically, what has happened if the play wouldn't have been made. So if player A makes a play that a run-of-the-mill fielder would have made only 80% of the time, and if a ball hit to this part of the ball park has historically led to 0.7 runs scoring, player A gets credit for (0.2 * 0.7) = 0.14 runs saved. And then you do this for each defensive play that he made, and add them all up. There are limitations to the defensive part of WAR (which I mentioned in a previous post in this thread). We now have very finely grained defensive data: we know exactly where each ball was hit, how hard it was hit, etc. We use that data when calculating modern players' WAR. We use less fine-grained data when calculating older players' WAR (because they didn't keep track of this information back then). There is, consequently, more room for error in older players' fielding records. Clemente may have been better than the data indicates - we just don't have any evidence that he was. And no, subjective impressions from watching him play don't count as evidence. Lots of people who saw Jeter play thought that he was a great defensive player, even though we have very detailed records of the plays that he made, and they were not, in fact, very impressive. He just looked impressive making mostly routine plays.
We've got lots of really bright people working really hard to figure out how to measure baseball performance. Let's help them. If you think that there's something wrong with WAR (and I do, as noted up-thread), let's figure out what's wrong with it, and how to fix it, so that it can become more accurate.
|