Introduction: Or, Who Are the Top Ten?
The Consensus Top Ten (Really 11) in NBA History generally looks something like this:
1. Jordan
2. LeBron
3. Kareem
4. Magic
5. Bird
6. Russell
7. Wilt
8. Duncan
9. Kobe
10. Hakeem
11. Shaq
Obviously this is subjective -- you see some jostling toward the bottom, as someone always gets left off (typically one of Kobe, Shaq, and Hakeem, based on bias; occasionally it'll be Russell). Some people have different orders toward the top (Bill Simmons as of 2010 had Russell at #2, as well as West and Robertson above Shaq and Hakeem). But this is basically the list you tend to see.
This list is terrible, not necessarily because the order is bad (although it is...), but because it has no clear organizing philosophy. Jordan is on top, allegedly for his scoring and his Finals win percentage. But LeBron is there for, what, Finals appearances? Kareem's there for scoring and rings, Magic and Bird for playing in a bunch of legendary Finals together and "dominating the '80s," plus Bird's consecutive MVPs get brought up. Why isn't Bird ahead of Magic? Rings. But then why isn't Kareem ahead of LeBron? Well, Kareem had more help. But then why isn't Hakeem ahead of Shaq? Well, Shaq was more dominant. But then why isn't Wilt ahead of Bird? Well, he didn't win as many rings. But then why isn't Russell ahead of Jordan? Well, because he wasn't good at offense. And around and around we go.
All these arguments look more than anything like post-hoc rationalizations of an arbitrary list based on "common knowledge." Jordan is #1 because he's the GOAT, right? If you press me I can talk about 30.1 and 6-0, but let's be real, it's because he's the GOAT. Magic's ahead of Bird because he won more, but they're always back-to-back because, well, they played against each other! And how could you not have them together! And Kobe's low because he's mean, and he played with Shaq (unlike the other players on this list, who never had talent around them and CERTAINLY never played with anyone else on this list).
Now I, being a nice and friendly girl, would never dream of imposing a rigorous philosophy that might challenge this dogma, but I do want to give this list a spine! So in this article, I will attempt to create a single formula that will provide this exact same list but with a thin veneer of objectivity.
N.B. I'm sure there's a much more efficient mathematical method to generate a formula here but it's probably less fun.
The Parameters:
So what extremely objective metrics will I use? Nothing crazy; no Win Shares Per 48 or Field Goal Percentage or anything radical like that. No, the stats we're using for this are simple, wholesome, everyday stats: rings, career points per game, finals appearances, seasons played, rings per season, finals appearances versus another player on this list, three point percentage, finals win percentage, MVPs, and assist titles.
The Stat:
The Top Ten Players of All Time are determined by the following metric:
4 * rings + 4.5 * career PPG + 3 * Finals appearances + 25 * rings per season + 15 * Finals appearances vs another player on this list + 5 * 3 point percentage + 70 * Finals win percentage + 2 * seasons played + 60 * MVPs + 8 * assist titles
Perfectly sensible.
What? Why?
So here's how this goes.
- The rings are much less valued than other factors because there really isn't that strong a correlation between rings and the order of this list. In order from 1st to 11th, the number of rings goes 6, 4, 6, 5, 3, 11, 2, 5, 5, 2, 4. That's a correlation coefficient of -0.27 (which, since lower numbers is better, is basically a small positive correlation).
- Career PPG is largely here to boost Jordan and Wilt and keep Russell from running away with this (despite rings being a relatively undervalued metric, many of the metrics are Finals- or ring-based, and Russell has a huge lead over everyone there). It also keeps Duncan a little lower than he would otherwise be.
- Finals appearances actually correlate a little better with this list: -0.48, or a fairly strong correlation (again, negative numbers are good here). So this is just a good stat for generating our list in general.
- Rings per season I don't remember why I picked. This was iterative and I'm not trying to mess with the order now.
- Finals appearances vs. other players on this list is pretty arbitrary and funny but is also the easiest way to get Magic and Bird into the top half of the list (Magic in particular benefits from it a lot). There are other ways to do it without this stat, but this was the first way I used.
- Three point percentage is because Bird doesn't really have a strong argument for top 5 without it lol.
- Finals win % was the initial Jordan cheat code, to keep him in first while we figured out everyone else. As it turns out it ends up benefiting Duncan more than anyone (relative to ranking; obviously as a percentage it's Jordan and everyone else at 100%).
- Seasons played I think was for Kareem.
- The last two were annoying. Wilt was stuck at the end of the list and I had to bring in MVPs to get him above Duncan, Kobe, Hakeem, and possibly even Shaq. After a lot of nudging we got him in place, but then to bring Russell down from ahead of Jordan we needed to mess with 3 point percentage. But that ended up moving Bird over Magic, so to fine-tune I had to bring in assist titles (which also benefits LeBron and WILT!!) to push Magic into 4th.
So is this a good metric?
No lmfao. Just for fun, let's throw a bunch of other players into this metric and witness the horror that emerges. I'm going to add a bunch of other great players from history who aren't typically considered top 10 (11) of all time and see where they end up. I'll also add some modern players.
Our more inclusive list!
1. Jordan
2. LeBron
3. Kareem
4. Dr. J (including ABA)
5. Magic
6. Bird
7. Steph Curry
8. Russell
9. Kevin Durant
10. Wilt
11. Steve Nash
12. Duncan
13. Kobe
14. Dirk Nowitzki
15. Karl Malone
16. Moses Malone
17. Scottie Pippen
18. Kevin Garnett
19. John Stockton
20. Robert Horry
21. Dr. J (just NBA numbers and accolades)
22. Hakeem
23. Giannis Antetokounmpo
24. Bob Pettit
25. Bob Cousy
26. Shaq
There are definitely other guys in there ahead of Shaq (it's a fair bet that a lot of '60s Celtics are), but I got bored at 30 players (the other four, coming in after Shaq, are John Havlicek, Dennis Rodman, Jerry West, and George Mikan, in that order). It's a lot of data lol.
This is not actually evidence that the metric is bad -- what it's evidence of is that this approach to ranking players is very inconsistent and than a metric approximating it will give unpredictable results. And of course it does! The original list is not based on this metric lol. I don't actually think this metric does THAT BAD a job at measuring the thing it's measuring. Like I'm shocked at how good that list is lol. You get weird stuff like Horry (because we don't have a built-in factor for like "nth best player on a championship team" other than PPG and MVPs), but I'm not sure that guys like Pippen and Garnett SHOULDN'T be ahead of guys like Pettit and Dr. J. And I'm not sure there's a good argument for having Steph Curry all that much lower than, like, Russell or Wilt or Hakeem.
But that said, this metric IS bad. It's built from pieces that very roughly approximate how good a player is, but they're thrown together in really weird proportions to get the order we needed. For an example, let's look at what pieces actually make up a player's score.
Score Analysis: Magic Johnson
Magic's score in this metric is 624.75... That's not normalized or anything. I've described all the pieces that go into it, but here's how much each piece contributed to the final score:
Rings: 3.2% (i.e. 5 rings * 4 = 20 points / 625.75 points)
PPG: 14.0%
Finals appearances: 4.3%
Seasons played: 4.2%
Rings per season: 1.5%
Finals appearances vs another top 11 player: 9.6%
3 point percentage: 24.2%
Finals win %: 6.2%
MVPS: 28.8%
Assist titles: 3.8%
That's actually a fairly even breakdown -- 3 point percentage is high because he's a guard, and MVPs because he has so many (3 is a lot), but even the small percentages affect the results a lot -- Magic is only 1.6 points (0.25%) ahead of Bird, and 63 points (about 10%) behind Kareem. This is why fine-tuning the list took so long lol.
So?
So it's not a great metric but it produces the list we want. And if you're going to cling to a stupid list based on nothing, you might as well invent a consistent, objective basis for it, and this isn't any worse than your intuition lol. At the very least it's consistent.
So can LeBron pass Jordan?
Are you kidding?
No, can LeBron pass Jordan?
Well, a side effect of the tweaks I had to make is that Jordan is out in front by a mile (it's largely his MVPs, 3P%, and PPG, in that order), so... no! Your childhood hero is safe.
But what would it take for LeBron to pass Jordan?
Jesus, are you serious? Okay, fine. Let's assume LeBron's PPG and 3P% stays even. What would he need? Well, some MVPs would do it -- even one MVP brings him close enough to Jordan that it and an assist title puts him over the top. But that might be unlikely, given how old he is and how little he tries in the regular season. So what about rings?
Uh, he would need a lot of rings. Like if he made and won the Finals every year (and did nothing else), he would pass Jordan at TEN rings.
What the fuck?
It's MVPs, bro. I needed to crank them way up to get Jordan to #1 (and then crank up 3P% to get Kareem down to third lol).
So LeBron for MVP?
??? I guess lol. If LeBron wins one more MVP he is officially the greatest player of all time. But if he wins nine rings and DOESN'T win another MVP then he's #2 forever. Or until Steph passes him.
No comments:
Post a Comment