http://probasketballtalk.nbcsports....leagues-leading-lineups-plus-tuesdays-recaps/ Just trying to add fuel to the Lin v Beverley fire - I kid, but the starting lineup seems to be doing just fine. I still have high hopes for Lin but it seems that the coach is right to start Bev. Lock if posted.
Well, I don't think that's the natural takeaway for the story. This is Bev+ rather than Lin-. Adds credence to the "fit" theory.
Would not mind at all trading you...and that's a no story. Our line-up will probably take a few more awards when it's all said and done.
This is why I hate statistics. Everyone thinks they understand statistics just because they took junior-high math. It's like people who think they understand grammar just because they know the alphabet song. These are literally the second and third sentences in the article: Does anyone know how Offensive Rating and Defensive Rating are calculated? Does anyone want to take a guess as to why you might not want to just subtract one from the other? Does anyone wonder why someone is taking an individual stat and applying it to "lineups"? Wouldn't it be nice to know the methodology involved? One might assume it's a straight average, but maybe some weighting was applied? Who knows, at this point. No detail is provided on how the numbers were actually calculated, and no work is shown. This is a hilarious misuse of Net Efficiency Rating, as well as its component parts. There is already a significant argument to be made that NER is not an appropriate calculation. Taking an individual player's ORtg and subtracting his DRtg is statistically akin to weighing a bucket full of oranges and subtracting the weight from a bucket full of apples to get the size of the bucket. Extending ORtg and DRtg from a single player evaluation into a lineup-by-lineup evaluation further takes those individual metrics and skews them completely out of context. Also, using a minimum of 150 minutes is laughable. Let me put it to you this way: 150 minutes is basically 3.125 games' worth of playing time. Does anyone here feel comfortable making assessments about individual players, let alone entire lineups, based on 3.125 games' worth of playing time? Think about how much noise there would be in that kind of data, let alone the limited amount of sample size in general. What lineups are usually played when? Don't small lineups usually play against small lineups? Doesn't that skew the comparison, in the sense that you are not evaluating the small lineup against its large counterpart, but you are evaluating the small lineup against other small lineups against the large lineup when it plays against other large lineups? Then, wouldn't you ask yourself, does it make sense to make this kind of comparison at all? Wouldn't you rather compare the two lineups against the same opponents in order to gauge their relative strengths? Look, there's statistics. And then there are fun numbers we throw out just to tell ourselves a story or to keep ourselves from surfing the internet from p*rn all day. These are clearly the latter. I thought baseball was bad enough with it's "Johnny never walked more than 4 batters on a Sunday when it was above 50 degrees in the Northern Hemisphere" inanity, but I see basketball is clearly catching up with numbers that are so obtuse that people can't or won't check them.
Nothing wrong with Lin. He looked good and comfortable off the bench. It's a defined role for him. He does well with it. He's hardly playing up to his full potential there at this point, lots of chaos happening on the court when he was healthy, much of it continues. Everything will settle into place soon enough. Lin will be a plus.
You are confusing the individual ORtg and DRtg stats that are given at basketball-reference.com with the lineup-based ORtg and DRtg stat. This Net Efficiency rating is not dealing with individual ORtg and DRtg stats. It is just taking the points scored per possession for the lineup and subtracting the points allowed per possession for the lineup. This is valid, mathematically, so long as the number of offensive and defensive possessions are the same. Your point about sample size and us not being able to draw conclusions from this is fair, though.
This is my favorite post ever on CF! You obviously get "it". You wouldn't happen to be a poker player, would you?
That just throws up red flags for this analysis. I'm not questioning the math, but something is wrong if the result is OKC has the worst starting 5 and only win cause of their bench? What's the explanation here?
I agree. This situation is no good for the Rockets and no good for Lin just trade him and be done with it. Why hang on to someone when they don't fit. Just let Lin go on his way.
Actually, you should watch the games for OKC. They play sooo sooo much better w/ Perkins on the bench. Their offense flows much better AND they play better Defense. Odd isn't it? Well, Reggie Jackson (MIP candidate), Westbrook (Westbrook moves to his primary position 2), Thabo (Defense), Durant, and Ibaka.
Has Lin started with that same line up before? Lin, DH, Parsons, Jones, Harden What about best line up disregarding who starts, as in who are the best 5 on at all times? I think Torocan did an analysis before but I'm wondering if there has been any updates to that.
I think this is the best lineup at all times. I don't think this is starters v. bench, it's starters v. starters + bench v. bench. Just so happens, that the starters post the best numbers. That's what the 150 minute qualifier is intended to do - a line-up that had played a minimum of 150 minutes. If this was simply starters v. other starters, then that qualifier isn't very meaningful.
