Author Topic: C's Third best offense from Nylon Calculus  (Read 1501 times)

0 Members and 1 Guest are viewing this topic.

Re: C's Third best offense from Nylon Calculus
« Reply #15 on: August 12, 2017, 05:50:00 PM »

Offline KGs Knee

  • Satch Sanders
  • *********
  • Posts: 9518
  • Tommy Points: 1055
Quote
Boston and Minnesota might be surprises to many people being this high. One aspect is that the Player Type ‘Gods’ is binary, so that Gordon Hayward counts as much as Steph Curry, even though not all Basketball Gods are actually equal.

This seems like a pretty big deficiency in the model.

Actually he's averaging the results of three different models.

If two of those models consider Hayward the equivalent of Lebron, the whole result is broken, wouldn't you agree?

The "player types" models are basically junk. If the methodology doesn't convince you, the results should. Do you think that Minnesota will have the best offense in the NBA next year? Do you think that we will essentially be tied with Golden State? That's what the player type model predicts.



I agree that any statistical model that puts Hayward on par with Durant/LeBron/ect. is a bad model.   Nobody should be putting much, if any, stock in such a model.

That said, I actually do think Boston has a good chance of being one of the top 5 offenses in the league this season. Our three best players seem to be a very good fit and have skills that complement each other quite nicely. How close we are to GSW is questionable, but I won't be surprised if the answer is "a lot closer than you'd have thought".

Re: C's Third best offense from Nylon Calculus
« Reply #16 on: August 13, 2017, 12:23:26 PM »

Offline ThePaintedArea

  • Gordon Hayward
  • Posts: 707
  • Tommy Points: 104
Quote
Boston and Minnesota might be surprises to many people being this high. One aspect is that the Player Type ‘Gods’ is binary, so that Gordon Hayward counts as much as Steph Curry, even though not all Basketball Gods are actually equal.

This seems like a pretty big deficiency in the model.

Actually he's averaging the results of three different models.

If two of those models consider Hayward the equivalent of Lebron, the whole result is broken, wouldn't you agree?

Honestly I don't understand the methodology. The mathematics is beyond me and I don't have an opinion one way or the other. It looks like you're asking the right question.

The "player types" models are basically junk. If the methodology doesn't convince you, the results should. Do you think that Minnesota will have the best offense in the NBA next year? Do you think that we will essentially be tied with Golden State? That's what the player type model predicts.

Again, I'm pretty sure I don't know enough to have an opinion. I guess we can check back in April.

I do think that Boston's offense will be pretty spectacular, and even better in 2018-19.

Minnesota was already 10th in offense last year (there's a puzzle: Thibodeau's team was wretched defensively...), and they've added Jimmy Butler.

http://www.espn.com/nba/hollinger/teamstats/_/order/true

Brahme's ten player types and the whole question of lineup balance and fit, on the other hand, is extremely interesting, especially nowadays; we seem to be undergoing a seismic shift in how the game is played and who gets on a roster and on the floor.

Somewhat lost in the Celtics' roster shakeup this offseason is the radical shift away from bigs and toward wings and swings who can create off the dribble. The journeyman Aron Baynes is basically the second big, and a key stat to watch this coming season will be his minutes. Brad Stevens played two bigs predominantly last year - but Danny Ainge has made that tactic problematic.  This is a big deal.