I forgot to mention this statement in my first post. Forgive me. I mentioned it in another post elsewhere and thought I also included it here when making my post.
Optimally you do want Artest to get the most touches on the ball. But the problem with that is, everyone in the building knows you want Ron to get the ball. Stopping every possession to let him iso isn't any better than what he was doing last year in Sacramento. The ideal would be to get some movement, either through penetration or through back screens and cuts, but we weren't getting either of those last night.
I don't think we want anyone dominating the ball with the second team. Certainly you look for ron, but you do that in the context of the wheeling motion high-post set that we were running better in the preseason--head and barry are both much more effective in this context, as it allows them to make smart cuts and shoot spot up jumpers. At the end of the shot clock, if that doesn't yield anything open, then you look for ron. The only problem with this is that last night, the rockets wern't getting into their quickly, if at all. Lowry and Conely are waaaaay quicker than luther and both pretty good defenders--if we can't get the ball up court and get the motion offense running quickly enough, we have two options, both of which we saw last night: A. Screw the offense, Iso artest B. Get into the offense too late, with no contingency strategy, and live with the consequences (the chuck hayes high-post spin move air ball, for example) The solution to this? Aaron Brooks. In case you didn't notice, he's fast. And has good handles. He can either get us into the offense quickly, or, failing that, use his speed and shooting to create offense for himself or others (unlike Luther "Kyle Lowry ate my Cookies" Head).
Problem i see is the sample will always be limited because while you can calculate what a player shoots over 82 course game. A line-up may only face the same opposing line-up for about 150 minutes a season if even. And I think that's too short of a time to make a statistically significant analysis. One player have a john starks type performance going 2-20 could lead you to a very wrong conclusion. In other words, outliers will have too much impact. Because of the limitation (5 on 5 with limited number of minutes) i don't think you can model this out. Statistics are great, but only when they are used to model things that have enough data. In this case, i think the data is the problem, not the methodology.
These stats reflect observations we naturally make when watching the game -- How does the score change when certain players are on the court together? Recording that information and having it available can only help you understand what works and what doesn't, so I'm not sure I'd call it a problem. Of course, I agree with you that you can't exclusively rely on it to form strong conclusions about players. You have to watch the game.
Wow... our best lineup by FAR was Brooks, Barry, Artest, Hayes, and Yao. +22 for the night. Didn't see that one coming.
I like that lineup. It's a nice combination of offense and defense, with good all around balance. Good penetration and shooting with Brooks, Barry, and Artest. Good post up game with Artest and Yao. Good defense with Artest, Hayes, and Yao. Good rebounding with Hayes and Yao in the game.
Not really. The starting group hasn't had many great starts to the 1st and 2nd halves. But that's to be expected, considering the poor shooting of the starters, collectively. It might be interesting to combine the game flows of all the games and see what those results look like.
