It’s unnerving. Damn you Cirillo, you can’t even suck properly.

Hey, long time reader first time poster.

So, I will happily admit that Raul was terrible at defense, but I just don’t trust UZR to assign meaningful +- run differentials to defense performance. I think there are too many defensive factors that can’t be accounted for or measured (yet), let alone quantified in a meaningful sense, particularly since defense doesn’t have nearly the discrete milestones that offense has.

I have no problem conceding that UZR does a great job measuring relative defensive abilities between players, but I just don’t believe that Raul’s defense was seriously so bad as to negate almost all of his offensive production.

Was there some definitive study I missed that showed UZR’s indisputable power? (That’s not sarcastic, I’m genuinely asking.)

Don’t get me wrong, Raul was a terrible defender, but I just don’t quite buy that he was THAT bad.

Ive always liked Cirillo and felt he was a little over-hated, but I never thought he was good for us.

Glad he gets his due as an average-to-below-average player.

Okay, I should have clarified, and you’re right to put the burden of proof back on my end here, but the only point I wanted to make is that unless I am misunderstanding something about UZR (which is admittedly very possible) I don’t buy the accuracy of +- run ratings. With offensive metrics it makes sense to say that some event, on average, results in so many runs while some other event, on average, takes those runs away. However, I don’t see in UZR that kind of precision, largely because of the highly diminished discrete possible outcomes of a defensive event. On any plate appearance there’s usually ten-odd outcomes for the batter that can occur depending on the situation. It is possible to assign average run values to each event in each situation, etc. etc. On defense, there’s so many possible outcomes that assigning run values seems suspect. How many runs worse is taking a twelve step route to a ball than a ten? What about two ten step runs with slightly different angles of approach? What is the run value of the time difference between contact and the defender’s first step?

I am not disputing UZR’s ability to tell you whether or not someone sucks at defense, and I am not saying that Raul was a good defender, or that Cirillo sucked with the glove, my issue is only with taking UZR run ratings as precise values. In that sense, perhaps I posted in the wrong thread.

Also, @ DJW, that’s an excellent point. In response I would say that many rating systems tend to conflate along the trend. Take stock equity “buy” or “sell” recommendations, for example. By their very nature they a make a good stock look great and a bad-ish stock look terrible.