My bad for writing it that way. I was trying to state that saying TD is better with the team is a bit dumb under the circumstances. Also, in case you didn't notice, rest of my post still refutes your argument.
I said earlier in the thread that I have not watched most of the games this season (you can guess why), which after the last few games may not be true any longer. I've managed to watch probably around half of the games completely, and portions of most of the rest. The reason I said it was to just make it clear that my views regarding how Lin was being used by the coaches was not based on watching all the games, so I didn't want to pretend like I had some sort of authoritative opinion on that subject. Of course, some people will use that and say that I therefore don't have the right to question the notion that the coaches were destroying Lin's confidence or to participate in a discussion on what the stats convey about the performance of the team or individual players. Oh well.
Can you repeat my argument for me, just so I'm clear we're on the same page on that. It seems like we're not.
I'm new here, but I can tell that you are a smart person. So it's great to get a compliment from you. My job is very much statistically orientated, both sports stats and various commercial applications of stats. I guess I have blinders on with +/-. It's quoted so often in so many different context and the worst is when it's used on ESPN by so-called NBA analysts or national writers. The sad truth about advance NBA stats is that there's very few people working on it publicly. Teams have much more advance stats as well as companies like Synergy but those aren't public. If I didn't have a 100+ hour a week job, I would seriously work on creating a new stat system to better measure individual performance in a single game.
You challenged "Try looking back at the games where Lin sits down. You will more often than not see the other team coming back." with incomplete data.
I have a hypothetical question, and I'm interested in your perspective (sorry this is a digression from the thread topic). Suppose there are two players (A and B), and you have to decide who is the better of the two. But you have to make a "best guess" based on incomplete data. Specifically, you know that both players put up very comparable, MVP-like numbers. For the sake of the discussion, let's say their box score stats, across the board, are identical. Then, you get one additional piece of information. Player A plays for a team that won 60 games, while player B plays for a team that won 20 games. Unfortunately, you don't get to know more than that. Is that last bit of information relevant? Most people, without the benefit of watching the teams play and knowing more about what happened in their respective seasons, would assume that the player who's team won 60 games must have had a much better year. Perhaps he was a better team leader, or he did a better job "making his teammates better", or he did things that don't show up in the boxscore. Is that reasonable because of the wide disparity in wins? Or would you be reluctant to consider team performance as part of the player evaluation even in this case?
OK. So when he says that when Lin sits down more of then that not the other team will come back, what does that have to do with whether Harden is playing or quality of shots or all the rest of the points you made? Is there some hidden caveat in his words that I missed?
I think I am approaching this as "if Lin sits down, TD will be playing" hence the additional points and the comparison, whereas you are approaching this from a pure data perspective... ... and my point is that you don't have the right data/stats to come to that conclusion.
Interesting point. Probably a better measure would be: TOV% = 100 * TO / ( %Ast*FGA + 0.44*FTA + TV) Instead of FGA, using %Ast*FGA will filter out assisted field goal attempts, for which there is very little chance of a turnover being committed.
OK. And I think I'm approaching it from what he actually said, whereas you're approaching it from what you think he should have said. Also, its not for me to come to a conclusion. He's making an unsubstantiated claim, which seems to me to be wrong or at least highly questionable. He or you or someone else can try to provide evidence to the contrary that actually supports what he says (Rockets tend to lose leads when Lin leaves the court). Burden of proof isn't on me.
I haven't really looked into +/- stats, but I think how well an player does +/- wise probably has to do more with substitution paterns than being a good or bad player. Does the other them starters or backups playing? Is it garbage time? For example at the beginning of the season Lin usually wasn't on the floor with Defino and TD who are #3 and #4 on our 3 pt made list. Also at a point Lin was taken out early so he wasn't getting as many minutes with Parsons and Harden who are #1 and #2 for 3pts made. Recently Lin has been sharing the floor more with Defino and TD so that could explain a little bit why his )+/- is improving. Lin's scoring his been up too, but who knows for sure without digging deeper into the stats. Let's pretend that Lin is on floor and we build a 7pt lead. He is +7. He goes to the bench and the lead jumps to 20(+13) pts inflated by good 3pt shooting. Lin comes back in and the other team goes on a run. They cut the lead to 12pt(-8). So Lin's +/- goes down to (-1). Say the game ends there and the Rockets win by 12. Lin's (-1) looks bad compared to the margin of victory(+12), but was it all his fault? We can't say without knowing who else was on the floor on both sides. From a pure numbers perspective it looks bad compared to someone else with say (+10) during the same game because they were on the floor for the initial (+7) with Lin and part(+3) of the (+13) run and was on the bench when the other team made its run(-8). So the other player had a (+10) does it mean they contributed more to the win than Lins(-1)? Maybe but we can't tell from straight +/-.
Aaaaand we come in full circle with you not understanding what context means. I give up. Let's enjoy the future Rockets games with our own perspective.
Why give up? If I'm wrong about what he was claiming, I'd like to know why I'm wrong. What do you think he really meant when he said: "Try looking back at the games where Lin sits down. You will more often than not see the other team coming back." Can you restate it for me so I'm not confused on the context?
I think all the points you make here are correct. You have to be *really* careful when using +/- stats to argue about how good a player is or isn't. Generally, its a bad idea to do that unless you're "adjusting" for various things and also considering other performance metrics, and your examples do a good job of illustrating why. However, if I say something like: "If you go back through all the games, you'll find the team tends to lose leads when player A is off the floor" that's a claim that is verifiable through +/- statistics. Similar, if I say "We tend to outscore teams when player A is on the floor" that's against verifiable through +/- statistics. So if one doesn't want to consider +/- stats, then they simply shouldn't make statements like that.
I see in the different view point. fight for what? Like the post name "Free Jeremy Lin"; I just want him free. JLin said "I don't worry about to prove anything to anyone." The bottom line is "Don't like him let him go." or "Set him free."
Lin can help build a lead every minute he's in there during the first 3 quarters and have the lead disappear everyone he sits which would give him a great +/- Then the rox could have a horrible 4th and get killed resulting in a negative +/- for the day. Or it could be a blowout and he could sit out the fourth and have the lead increase greatly lifting TDs +/- above Lins even though Lin was the one who got the team the huge lead going into the fourth forcing them to I've up and bringing in their bench players. This could also result in TD having a better +/- though Lin was more important when it mattered. Both these scenarios match the premise of the lead disappearing when he sits but still having a lower +/- than TD. That is why context is so important
that still won't explain why his +/- in majority of games early in the season is significantly negative
