well yeah, you can actually cull your malcontents.

in all these other threads, these guys believe they've reached mythical (i.e. unbannable) status.

no i meant sports.

lol no, the stuff that actually generates good bball discussion are stuff that are close enough to argue

for every post that actually made someone think more about sports, thremp made 300 trolling posts that were argumentative and unreadable - it's not hard to see how victor would be a more informative poster

i think victor will eventually be brought back

Disagree. The stuff that is legitimately close is usually boring and useless. Nobody has the expertise or tools to make meaningful distinctions at that level of resolution, so you end up with argument after argument where the sides are basically drawn by entrenched biases, homer and otherwise.

Arguments that make a large departure from expectation, while by no means being universally good, at least offer the potential for interesting discussion.

That only matters, though, if you actually give a **** about that sort of thing, rather than a continuing focus on who's right and who's wrong, judged against silly strawmen by silly over-confident posters.

Dunno about basketball, but this is definitely true for baseball.

Great post, thanks Clark

.

There is no such thing as a tiny sample of RAPM though. The thing is it does mean a lot.

I don't want people to misinterpret this, but about ridge regression in general: (RAPM) it's not my job to explain higher level math to people. Some people understand it. if you don't, and you make an effort to learn, kudos. But if you don't get it, arguing against it is very ignorant, since it's important to understand what you're actually arguing.

Basically, if you want to say it's a small sample, maybe learn about ridge regression first, since it pretty much addresses those concerns.

draft threads tho

Scottish thread tho

hockey threads

someone just sent me this, ainge of ellie's head, lol awesome

http://www.youtube.com/watch?v=JDKea...layer_embedded

wangz-sharks thread is a dumpster fire

i don't really know **** about advanced basketball metrics or higher level statistics in general but 15 minutes of googling (including skimming sill's paper on rapm which included a bunch of stuff i didn't understand) leads me to believe that apm is a metric with absurd standard errors for 1 year sample sizes. and that rapm does mitigate this problem somewhat but does not turn it into an amazing super stat that will never have a sample size problem again.

if i am correct (and since i don't really know much about this particular subject there's a vg chance i'm not) it would make me think that bringing up sample size issues for the rapm of a guy who has played 82 games at 12 minutes a game might not be as ludicrous as you suggest even if it is better than apm.

I haven't been reading the playoff threads because I have been too busy watching people golf, but I the regular season hockey thread was probably my favorite to participate/read, especially if you decide to ignore goofyballer liveblogging DVRed games

agree

I just spent some time catching up on it. Yes, you are correct, it sucks.

playoff threads are usually aids in every sport

Except for the fact his RAPM is so high, much higher than the pretty low SE, means that even if he's not historically great, at worst, he's pretty ****ing good.

Meaning, you can't dismiss his numbers - you may argue he's not actually the SECOND best defender in the league (I wouldn't mind that position) but at worst he's top 10, basically.

While true it did give me quite a few laughs.

Mavs/Lakers has been fun.

I've read Sill's paper on RAPM when I was thinking about submitting an article to the JQAS. It sounds like it significantly reduces the standard error of estimating someone's true talent in a given discipline. So it sounds like Bobbo's position is reasonable.

Here is that paper if anyone wants to read it, I'm doing it now

http://www.sloansportsconference.com...erWithLogo.pdf