I like 82games.com some of their stats I like: - On Court/Off Court - "Clutch" ranking (4th quarter or overtime, less than 5 minutes left, neither team ahead by more than 5 points) - shooting performance by shot type and shot clock usage - Points per possession - Free Throw Shooting and Foul Drawing - Passing Stats Someone posted a really good stats site last year but I can't find the link...
I also like http://www.basketball-reference.com/ Its a fun site although I have learned that not only can you find some shocking stats but you can also make just about player look good with a little work.
Yeah basketball-reference is probably my favorite because I can look up various eras and I like their blog. 82games, hoopdata, and a couple of lesser known places.
I can't believe I've overlooked this site. I really like their interface. It's much easier to use than most stat sites.
I like hoopdata.com because it has stats on where a player/team shoots from on the floor. But I can't find stats on fast break points anywhere.
Another vote for basketball-reference. As for a specific statistic, I am rather fond of win shares per 48 minutes. It seems to be a decent predictor (though still not perfect) of player success with the context of their team. I also like Wins Produced, a statistic championed by David Berri at http://dberri.wordpress.com/. Once again, it's not perfect but it was an attempt to determine what values lead to wins via pure numerical analysis versus preconceived notions. Coincidentaly, that is why I do not care for PER. It is an attempt to derive a single number for values in a box score that everyone "knows" are valuable which makes it flawed from the very beginning.
Actually the PER stat is mainly meant to normalize statistical output on a per minute basis and to account for differences in pace.
Basketball reference and 82games give me most of what I need. I try to keep up with some of the more esoteric stuff like Pelton's when I can find it.
If you like per-48 minute stats, or effectiveness points (a summation of all stats into one stat), you should click around Hoopstats.com. The site is dying for lack of use and it's a shame. http://www.hoopsstats.com/basketbal...-trail-blazers/team/playerstats/11/25/1-3-eff
The question is how things are weighed in that attempt. PER's weights were setup based on preconceived notions of which statistics were worth the most. The two I mentioned above determined via number crunching which statistics were worth the most then weighed them accordingly. Certainly none of them are perfect and there are some really odd results for any statistic. I am just more likely to be swayed by a statistic that is derived scientifically. PER wasn't.
Dude you should know, most APBRmetrics people seem to be against Wins Produced. So that's not a very good argument. That includes Basketball-reference, 82 games, etc. Wins Produced measures team stats well, but has trouble separating individual success. APBR people on this matter: https://www.blogger.com/comment.g?blogID=31545676&postID=8497886239332362177
I didn't realize I was arguing about much other then my personal preference for what stats I look at. I don't have the time to go over all these arguement with a fine tooth comb to determine who is "correct", even though that whole notion of truth in statistics is pretty pointless. That being the case, I get to choose which expert I like more. I happen to like Berri more then I do the other members of the establishment since I tend to side with the mavricks and outsiders most times. Arguements like the above are why I steered away from statistics in college. I much prefered areas of math where if you made the logical steps in your proof correctly then you could be confident of the result. As for specific reasons I happen to like Berri's work I submit the following: Though it certainly values scoring it values rebounding nearly as much. This coincides with what I think so I dig that. It penalizes volume shooters who also score a ton of points more then the others. I like this notion as well. It actually counts fouls as a negative. I do not agree at all with some of things it says, such as teamates/coach/system not being a factors in production. That's not just counterintuitive, that's simply wrong. If I am a good small forward but sitting behind LeBron I would look absolutely terrible even though I might actually be good. Really, I look at all the evidence out there, discount nothing completely and see what I see. I just happen to have an order of preference that appears to be contrary to what a significant number of experts think. Fine by me since every revolution in thought was derided by the experts until it became obvious.
I only brought it up because your premise was incorrect, that other apbr stats are just some form of confirmation bias, or not as scientific, etc. One of the reasons I like Basketball-reference is because it uses various metrics to attempt to make it's point. Generally that's probably a good starting point for everyone. As mentioned above most advanced stats will have a high coefficient of correlation, that's why these fine details mean so much to these other apbr people.
