One of the frequently used arguments used against Lebron, which I think has merit, is that he played in a historically weak division (Magic did too), so his stats might be inflated.
Are there are stats or modifiers than can be used to level out according to competition?
Is there a statmodifier for level of competition?
Moderator: Doctor MJ
Is there a statmodifier for level of competition?
-
- Bench Warmer
- Posts: 1,404
- And1: 401
- Joined: Aug 05, 2016
Re: Is there a statmodifier for level of competition?
-
- Senior
- Posts: 666
- And1: 721
- Joined: May 03, 2015
Re: Is there a statmodifier for level of competition?
90sgoat wrote:One of the frequently used arguments used against Lebron, which I think has merit, is that he played in a historically weak division (Magic did too), so his stats might be inflated.
Are there are stats or modifiers than can be used to level out according to competition?
On basketball reference you can see player splits from '84-present. Included in these splits are the player's box scores against each team. You can easily use these splits to isolate a players box scores against teams that meet certain requirements. For instance, you can quickly find Lebron's efficiency against teams that are among the top half defensively.
A few weeks ago I determined the "added points" [ PTS - 1.07*(FGA+0.44*FTA+TOV) ] for Magic, Lebron, Jordan, Shaq, and Duncan against the top half of defenses and the bottom half. I then took the difference to see who how resilient (as scorers) the aforementioned players were against tough defenses. Here's the results if you're interested:
Lebron
Spoiler:
Shaq
Spoiler:
Jordan
Spoiler:
Magic
Spoiler:
Duncan
Spoiler:
You'll notice that Magic was actually the most resilient and regularly scored better against strong defenses than weak ones. Duncan closely follows Magic. Jordan and Shaq are pretty similar and Lebron is arguable the least resilient with the fewest seasons of small or no drop off.
Re: Is there a statmodifier for level of competition?
-
- Bench Warmer
- Posts: 1,404
- And1: 401
- Joined: Aug 05, 2016
Re: Is there a statmodifier for level of competition?
Blackmill wrote:You'll notice that Magic was actually the most resilient and regularly scored better against strong defenses than weak ones. Duncan closely follows Magic. Jordan and Shaq are pretty similar and Lebron is arguable the least resilient with the fewest seasons of small or no drop off.
Great stuff man, you should type this up more and make a blog post or Bleacher Report somewhere. Definitely a valuable stat to have.
Where do you draw your stats from? I suppose it's not manually, do you use software?
Re: Is there a statmodifier for level of competition?
-
- Senior
- Posts: 666
- And1: 721
- Joined: May 03, 2015
Re: Is there a statmodifier for level of competition?
90sgoat wrote:Great stuff man, you should type this up more and make a blog post or Bleacher Report somewhere. Definitely a valuable stat to have.
Where do you draw your stats from? I suppose it's not manually, do you use software?
I actually do it manually. I copy and paste from bballref into excel and enter whatever formula defines the stat I want. I probably won't do any posts about this stuff though. I'm not that into most box-score derived stats. There's just too many caveats. For instance, in game 3 of the '91 finals, Jordan shot very poorly, but on five of his misses he collapsed the defense such that his teammates had an overwhelming advantage on the boards and easily grabbed the offensive rebound. In particular, Jordan rebounded two of his own missed shots. Using the "points added" stats, Jordan's game 3 would show up very poorly, but actually many of his misses weren't harmful. Box-score derived stats have a place, I think, but I'd rather pursue alternatives that are less likely to significantly break down when the implicit assumptions go awry.
Return to Statistical Analysis