看板 MLB 關於我們 聯絡資訊
※ 引述《abc12812 ()》之銘言: : sport, like the world, is messy. But the bottom line is that WAR works. In : 2012 the correlation between Baseball Prospectus' WAR and team victories was : 0.86 (where 1.0 would have meant a perfect correlation). The correlation : between batting average and victories was 0.27. Teams with more WAR win more : games. Teams with better batting averages don't. Sure. WAR is a better statistic than the batting average. I do not think that many people disagree. You can check the correlation between victories and the differences of the runs forced and the runs allowed. It would give you a higher correlation. Does that mean R+RBI is a better statistic for offensive players? Furthermore, WPA (win probability added) would give you a perfect correlation to victories. Would you consider it? There is no doubt that WAR is a good answer. But is it THE answer? Could be, but little support from your long arguments. : at a certain point, the question in any debate against science is: What are : you really fighting and why? You should ask this to yourself, and answer it seriously. -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 140.109.74.113
GuyPearce:context 02/20 12:59
searoar:wut 02/20 13:02
searoar:runs scored minus allowed has higher correlation?? 02/20 13:06
searoar:even if it's true, wut is that to say about R+RBI?? 02/20 13:08
searoar:WPA has a perfect correlation? have you done teh math? 02/20 13:09
tlchen:For the same season (2012), Corr(RS-RA, W)=0.947 02/20 13:37
tlchen:You wont believe that Corr(R+RBI, W)is about the same? 02/20 13:39
tlchen:I said for offensive player, so I roughly divide each 02/20 13:40
tlchen:run to the player scores, and that the player drives in 02/20 13:40
tlchen:Now, for the WPA, do you know the definition? Even if 02/20 13:41
tlchen:you do not know, how about taking the total WPA of any 02/20 13:42
tlchen:team to see if W=81+WPA? If you still do not believe 02/20 13:43
tlchen:that correlation=1, you can check with any calculator 02/20 13:44
tlchen:Sorry, I did not mean to say Corr(R+RBI,W), I meant 02/20 13:47
tlchen:Corr(R+RBI,RS) is nearly 1, so I use it to evaluate the 02/20 13:48
tlchen:offensive performance 02/20 13:48
tlchen:I was not going to argue that R+RBI is better. I want 02/20 13:49
tlchen:to say that it is better by the logic of the author 02/20 13:50
searoar:http://tinyurl.com/a37ngwx not the nats but the giants 02/20 13:50
tlchen:are you sure that you added both batting and pitching? 02/20 13:54
searoar:R=RS and RBI≒R I am not sure wut u goona proof here 02/20 13:56
searoar:OK, my bad, so u know the def of WPA 02/20 13:58
searoar:but don't know wut it designed to answer? 02/20 13:59
tlchen:I don't know what makes you think so. Anyway, whether I 02/20 14:20
tlchen:I know or not is irrelvant. I just follow the author's 02/20 14:22
tlchen:logic to present WPA as a superior statistic 02/20 14:22
tlchen:For R&RBI, as I mentioned above, it would also be a 02/20 14:24
tlchen:better statistic from his logic. 02/20 14:25
Parazicecum:my dick is so big 02/20 20:36
star01:mom, i am here 02/20 22:18