3/5 | Sunday Game Thread | NBA on ABC

[MrCheerios]

I don't use PER as an end all stat. I know it undervalues defense and I try to give context in player evaluations. I say Taj Gibson is better than Anderson in a lot of aspects, but I know Anderson is much more valuable next to Harden in that system and a better fit. And that blog says the average PER is off by .37 for the top 200 players, so formula would not be off significantly. You would have to shoot a metric f-ton to have it significantly affect your PER and that statement might apply to Westbrook more than anyone else.

And PER certainly seems like a more accurate and consistent summary than BPM alone. Covington didn't suddenly become a really good player over the two prior seasons in Philly. He's mostly the same player. And he's not better than Adams and Kanter. BPM isn't a starting point in player comparisons if it suggests that.

So basically you're going feels > reals here. OK then.

No, I'm saying there is no one stat that determines a players worth or contribution, whether BPM, PER, win shares, or RPM. There are several stats that show Adams and Kanter are better players than Covington. The one you're mainly using to denigrate OKC's supporting cast (BPM) says that Covington suddenly became a lot better than them this year. I think it's inherently flawed for player comparisons in that regard because it highly team dependent. I don't know why Covingotn's BPM jumped so much this year, but it probably has something to do with playing a lot of minutes with Embiid in the starting lineup, who has a very high BPM. And when Embiid didn't play, he played with Noel who is a very good defender and also has a positive BPM.

When you say things like Adams is so much worse than last year and no one can create offense for OKC besides Westbrook and VO part-time (ignoring Kanter), isn't that also feels over reals? You are arguing for some stats so strongly that you think Covington is better than your centers. Maybe that approach is flawed?

[bondom34]

I don't use PER as an end all stat. I know it undervalues defense and I try to give context in player evaluations. I say Taj Gibson is better than Anderson in a lot of aspects, but I know Anderson is much more valuable next to Harden in that system and a better fit. And that blog says the average PER is off by .37 for the top 200 players, so formula would not be off significantly. You would have to shoot a metric f-ton to have it significantly affect your PER and that statement might apply to Westbrook more than anyone else.

And PER certainly seems like a more accurate and consistent summary than BPM alone. Covington didn't suddenly become a really good player over the two prior seasons in Philly. He's mostly the same player. And he's not better than Adams and Kanter. BPM isn't a starting point in player comparisons if it suggests that.

So basically you're going feels > reals here. OK then.

No, I'm saying there is no one stat that determines a players worth or contribution, whether BPM, PER, win shares, or RPM. There are several stats that show Adams and Kanter are better players than Covington. The one you're mainly using to denigrate OKC's supporting cast (BPM) says that Covington suddenly became a lot better than them this year. I think it's inherently flawed for player comparisons in that regard because it highly team dependent. I don't know why Covingotn's BPM jumped so much this year, but it probably has something to do with playing a lot of minutes with Embiid in the starting lineup, who has a very high BPM. And when Embiid didn't play, he played with Noel who is a very good defender and also has a positive BPM.

When you say things like Adams is so much worse than last year and no one can create offense for OKC besides Westbrook and VO part-time (ignoring Kanter), isn't that also feels over reals? You are arguing for some stats so strongly that you think Covington is better than your centers. Maybe that approach is flawed?

Because nobody can create offense other than those 2? Kanter can't. He needs to be fed the ball. And I'm not even arguing for one stat. Try RPM. Try on/off. Try literally any measure of overall play, all you've said is "well Adams PER...". Sorry but when you look at the numbers for support they're not even ballpark close. Look at shooting, look at whatever you'd like.