Posted by simrud:
Would you drop the +/- crap, it means nothing.
Let's take a closer look, shall we?
+/- are a product of bad ang good teams. People on bad teams, who play major minutes will have bad +/-'s cause guess what, there teams get outscored cause they loose.
Wrong. +/- measures how much better the team plays with a player on the court, vs. off of it. Suppose there's a good player, Smith, on a bad team. Suppose that whenever Smith is on the court, the team gets outscored by 10, and when he's off, the team gets outscored by 20. That comes out to -10 - (-20) = +10. Now suppose there's a good player named Jones on a good team. Whenever Jones is on the court, his team outscores the opposition by 20, and whenever he sits, his team outscores the opposition by 10. That's 20 - 10 = +10. So Smith and Jones get the same ranking from +/-, even though one is on a bad team and one is on a good team. It doesn't matter how their teams do overall.
To take an 11 game at like 10 min a game sample and use stast is just plain ignorant, such a small sample space is not telling of anything.
You'd be right if I were using Taylor's +/- thus far to estimate what his +/- will look like in 100+ games as a Knick. But I'm not using his +/- stats to predict how well he'll do in the future; I'm only using them to evaluate how he's done so far. So this is not a question of estimating, but straightforward measuring.
For example, suppose we take a sample of 3 posters on this forum, and we find that their average weight is 200 lbs. If we use this small sample to estimate that the average weight of every poster to this forum is about 200 lbs, we'll be in trouble, because the sample size is so small that the variance is going to be incredibly high. But if we just want to ask what the average weight of the 3 sampled posters is, obviously, the sample size will not matter. We just go ahead and do the calculations, find the average is 200, and that's that.
That's what I've done in this thread. gunsnewing claimed Taylor has been terrible on the court so far as a Knick, but I observed that his +/- as a Knick so far is a quite good +9.9. 100+ games from now, Taylor's +/- very well could be something average like +1.6, but that is not what we're talking about. We are talking specifically about the games he's played so far. In those games, his +/- is +9.9, and the sample size does not matter-- we are not estimating, we are just measuring.
If you want to talk about bad use of samples, look no further than gunsnewing complaining about a couple of bad 4th quarters to come to the conclusion that Taylor's production so far as a Knick has been terrible. That's only looking at roughly 10 minutes of court time out of about 110. That is bad use of sample statistics. What I have been doing is taking all 110 of those minutes into account, so my claims do not suffer from poor sample sizes.
It is unbeleavable how he gets away with pretty much treatin NY fans as a bunch of morons.
You weren't talking about yourself here, were you?
help treat disease with your spare computing power : http://www.worldcommunitygrid.org/
