LunaticWolf
Well-Known Member
God, you are just using On-Off court. Use the simple rating, combines Production and On-Off. On-Off by itself is absolutely meaningless.
Not quite. Look again.
God, you are just using On-Off court. Use the simple rating, combines Production and On-Off. On-Off by itself is absolutely meaningless.
This is the system Houston GM Daryl Morey uses to evaluate players.
Hardly meaningless, and accurate more times than not. By isolating it from production you can really identify defensive liabilities or chemistry problems. Clearly Jefferson is a problem - we are getting blown out by an average of 21 points/48 minutes with him on the floor vs one of our centers. There is no way to sugar coat that, but I welcome you to try.
Now if you want to get more technical you can use adjusted +/- which accounts for quality of players they face, but the trend is usually the same as we see Jefferson at a gruesome -30!! This is the system Houston GM Daryl Morey uses to evaluate players.
https://basketballvalue.com/teamplayers.php?year=2010-2011&team=UTA
On-Off relies too much on what other players are doing on the court. The reason why their On-Off is terrible is because they are being replaced by players who are currently performing better than they are.
If I had two LeBrons, and I played each 24 minutes, assuming they performed exactly the same, their On-Off would be 0.
Likewise, if I had two Treton Hassels, their On-Off would be zero.
Now lets say I had a Lebron and a CJ Miles, CJ's on-off court is now terrible, because LeBron is amazing.
If I had a CJ and a Treton Hassel, CJ's On-off is now just as good as Lebron's was.
Are you telling me you see nothing wrong with that?
Lets also say we have Marc Blount playing center when James is on the court, and Mark Madsen playing center when CJ is on the court.
Blount's On-Off matches Lebrons while Madsens matches CJs.
Clearly Blount is 10x better than Madsen.
No system is precise, and these numbers are on a very low sample size of minutes. The magnitudes aren't very meaningful. The relative values are a bit more meaningful.Well these perfect sub patterns without crossover do not happen in reality and what are their minutes? If James and Blount are playing 80% of the time it means Madsen and CJ are playing 20% and I'm not overly concerned with low usage bench players. I'll be the first to admit Fesenko isn't as good as his +18 would indicate, but I do believe such a wide margin this year and last year shows our weakness of lacking a real center.
OK, but this doesn't exist.On-Off relies too much on what other players are doing on the court. The reason why their On-Off is terrible is because they are being replaced by players who are currently performing better than they are.
If I had two LeBrons, and I played each 24 minutes, assuming they performed exactly the same, their On-Off would be 0.
Then use on-court +/-, too (or instead), which would still tell you to play more of Miles & the centers and less of Bell & Jefferson. I don't think that it's a coincidence that Raja and AJ talk a good talk and are getting more minutes than what might be optimal, but Fes & Miles probably don't (and Elson is new to the team) and are probably playing less. (Another big factor, though, is Miles huge swings from being abysmal to amazing.)Now lets say I had a Lebron and a CJ Miles, CJ's on-off court is now terrible, because LeBron is amazing.
If I had a CJ and a Treton Hassel, CJ's On-off is now just as good as Lebron's was.