Posted by burnsy483 on 12/6/2012 7:17:00 PM (view original):
Posted by inkdskn on 12/6/2012 7:08:00 PM (view original):
OK.
Do you think w-l has more value than allowing a person to guess how a pitcher performed, e.g., in 'high-leverage' situations?
That's the only point that I think has been made by pro-w/l posts here. I don't put much value in that because the guess is likely to be incorrect for a myriad of reasons. Do you see more in it than that?
I'm not sure I entirely understand your question. But yes, I think W-L record has more value than allowing anyone to guess how a pitcher preformed.
Your examples thus far have shown how, in theory, a guy with a higher ERA and better w/l could reasonably be called a better pitcher than someone else with a better ERA, but worse w/l. However, their stats don't inherently reflect that--the scenarios require explicit knowledge of how both guys threw in different game situations all year (in other words, the guy with the better ERA and worse w/l could easily have been better 'when it mattered,' as you know).
Could the w/l of the two pitchers be consistent with the scenarios you describe? Sure, the guy who threw better 'when it mattered' may well get more wins. However, could the exact opposite also be true? Yes. W/L doesn't actually tell you any of the hypotheticals you've brought up.
Your posts suggest that a reasonable person can
guess that a pitcher with a better w/l record 'pitched to the situation' better than someone else who had a worse w/l but better ERA. And while that's true, that guess is at least as likely to be wrong than correct.
Please explain what value you see in w/l beyond that. The fact w-l is a function of so many things completely beyond a pitcher's control (managerial attitude, bullpen, offense), renders using w/l as a useful metric to measure individual performance useless in my eyes, particularly when w/l shows higher correlation to them than to effective pitching.