Out of curiosity I decided to check the results given by my values of NBA statistics that I described here:
How to compare NBA players comprehensively
All the statistical data I used I found on this site:
http://www.basketball-reference.com/
As always I ignored all the seasons prior to the season 1976-77 because after the NBA merged with the ABA the NBA players’ job became more difficult.
All the graphs are based on search results given by the awesome team-finder. For example the search results for the season 1976-77 were this:
http://bkref.com/tiny/CL4oT
For every regular season I summed all the numbers and divided them by the overall number of games, so I calculated per-game averages for an average team in every season.
The first graph shows the actual points-per game averages and the overall per-game value calculated my way:
Pretty good results, aren’t they? But in the last several years the OV/G is getting more and more higher than PTS/G. Why? A safe bet would be: “because teams attempt more and more 3-pointers on a good-enough percentage”. So let’s see a graph for 2PA, 3PA and FTA:
Attempts have to be see through the percentages:
So, the shooting percentages are more or less the same, but even tiny differences combined with a big number of shot attempts may become very important.
A combination of number of 2PA, 3PA and FTA with 2P%, 3P% and FT% (respectively), taking into account also team play and the fact that a 3P shot is more difficult than a 2P shot, is a value I called “scoring efficiency” different for each kind of shot (SCE_2P, SCE_3P and SCE_FT). The overall scoring efficiency (SCE_O) is the sum of SCE_2P, SCE_3P and SCE_FT. Let’s see how a graph for scoring efficiencies:
When I looked at the regular season 2018-19 I was surprised that SCE_2P was higher than SCE_3P. I asked myself a question: “Did 3P% got below the breaking point between 2P% and 3P%?” Well, it did not. The current 2P% is 0.520, so the breaking point is 3P% of 0.347 (2*0.520/3) and the current 3P% is 0.355 – still higher than the breaking point. So, what’s the catch?
After some time I realised that the problem lies in the reference point for my values for scoring efficiency for 2P and 3P shots (the reference is the same, but my formulas take into account that a made 3P shot is worth more than a made 2P shot). The reference point is the 2P% of 0.483 (average 2P% in the seasons from 1976-77 to 2016-17), but today’s 2P% is much higher (0.520). If the reference point were 0.520 (for all the seasons) the graph for scoring efficiencies would look like this:
In that case the SCE_3P in the season 2018-19 would be higher than SCE_2P (because 3P% is above the current breaking point between 2P% and 3P%), but the scoring efficiencies for 2P and 3P would be greatly distorted in MOST of the previous seasons.
So, the reference point of 2P% of 0.520 would be definitely worse than the reference point of 2P% of 0.483. But the crucial questions remained: “Are the current 3P shooters undervalued?” It turns out they are NOT!
To verify it I simply calculated a difference between scoring efficiencies calculated with the reference point of 2P% of 0.483 and scoring efficiencies calculated with the reference point of 2P% of 0.520 (all the differences would be negative because the reference point would be higher):
As you can see the overall difference would more or less the same in ALL seasons! Actually the overall value in the most boring era (late 1990s and early 2000s) would drop the LEAST. To verify this I checked how much would drop the per-game overall values of particular players in the season 2018-19 and at the top of the negative list there would be mostly 3-point shooters:
1. James Harden: -1.4
2. Paul George: -1.2
3. Kemba Walker: -1.1
4. Stephen Curry: -1.1
5. Russell Westbrook: -1.1
6. Donovan Mitchell: -1.1
7. LeBron James: -1.1
8. Bradley Beal: -1.1
9. Devin Booker: -1.1
10. Damian Lillard: -1.1
11. D'Angelo Russell: -1.0
12. Kawhi Leonard: -1.0
13. Kyrie Irving: -1.0
14. Joel Embiid: -1.0
15. Klay Thompson: -1.0
16. Blake Griffin: -1.0
17. Zach LaVine: -1.0
18. Anthony Davis: -1.0
19. CJ McCollum: -1.0
20. Kevin Durant: -1.0
21. Jrue Holiday: -0.9
22. John Wall: -0.9
23. Karl-Anthony Towns: -0.9
24. Giannis Antetokounmpo: -0.9
25. Buddy Hield: -0.9
26. Luka Doncic: -0.9
27. Nikola Vucevic: -0.9
28. Andrew Wiggins: -0.9
29. Victor Oladipo: -0.9
30. DeMar DeRozan: -0.9
Pheeew! It made me peaceful again – I don’t have to recalculate all the MVP and GOAT/draft values given in my other posts. It turns out that my scoring efficiency values are values that are truly RELATIVE (relative to the average 2P% in the season from 1976-77 to 2016-17), not absolute. And that’s the goal when you want to compare different NBA seasons.
No comments:
Post a Comment