lessthanjake wrote:Lol, just read through the last few pages of this thread, and was surprised to find that my own posts in the PC board and the responses to them were discussed a bunch in this thread. My two cents is somewhere in the middle to be honest: I do actually think that there’s a lot of hostility on the PC board to anything that is perceived as pro-Jordan or anti-LeBron. I think I can reasonably say that I myself experienced quite a bit of hostility—including a boatload of sarcastic and mocking remarks—while arguing in favor of Jordan (though I won’t claim to be a delicate flower that doesn’t punch back). And those mocking remarks did start essentially immediately after I posted in the #1 thread saying I was voting for Jordan. Meanwhile, the number of voters in the project is small enough that the results aren’t really meaningful except to record what a very very small group of people think. That said, there’s also people posting a lot of interesting info in those threads and I do think I’ve learned a fair bit and been alerted to data/sources I was not previously aware of, so I think the exercise has value even if the results aren’t very meaningful and there’s a contingent that does make it unpleasant to argue views that they disagree with.
_________________
To answer the actual point of this thread:
I think the answer is that it’s not possible for there to be a super strong “data-driven argument” for Michael Jordan, if you define “data-driven” to just be about impact data, because the data that exists from his era is much more limited. I don’t think there can be an airtight “data-driven argument” regarding impact data in the era before play by play data.
That said, I do think that the argument for Michael Jordan as GOAT is supported by the limited data we have. Specifically, Jordan comes off quite well in RAPM data that exists for his career:
- Squared’s RAPM data includes RAPM for snippets of the 1984-1985, 1987-1988, 1990-1991, and 1995-1996 seasons. Jordan is #1 in all those seasons except for 1984-1985 (which was his rookie season, and he’s 6th, and 4th amongst star players). And in those three years that Jordan is #1, no one is all that close. Example:
https://squared2020.com/2021/09/11/1990-1991-nba-rapm/- GitHub RAPM data exists for Jordan’s last two seasons for the Bulls, using actual play-by-play data. It has Jordan 2nd in RAPM in the regular season and 2nd in RAPM in the playoffs in 1996-1997, and he’s 1st amongst actual star players in both. Meanwhile, it has Jordan as 11th in RAPM in the regular season (and more like 6th amongst stars) and 1st in RAPM in the playoffs in 1997-1998.
https://basketball-analytics.gitlab.io/rapm-data/- I just saw someone in this thread posted another source of 1997-1998 RAPM data, which has Jordan at #1 in RAPM that season, by a fairly significant margin:
https://docs.google.com/spreadsheets/d/1R_VgaKr980LMmReroItR94hzhuexf9wxM4l3cIw0QmY/edit#gid=0- Jordan was also #1 in RPM in 1996-1997 and #5 in 1997-1998, though that is a stat with a box-score component on top of the impact component, so it’s not a pure impact measure.
- JE derived a RAPM estimate for the entire 1990s, by using quarter-by-quarter box scores and minutes data. That estimate has Jordan as #1 in RAPM in the decade, and ahead by about 30% over the next highest person who actually had a lot of playtime in the decade.
The overall picture is that, in the limited RAPM data we have, Jordan actually looks incredibly good. It’s just single-seasons, single-playoffs, snippets of seasons, and a RAPM estimate, but everything we have has Jordan ranked really highly. Indeed, most RAPM data we have literally has Jordan #1 in the NBA. And, while it’s limited data, we could reasonably make a pretty strong inference that Jordan would probably look extremely good in RAPM if we had the full sample of data. After all, it is very unlikely that someone would look so consistently great in really noisy small samples if they weren’t way above everyone in larger samples. Indeed, even if you look at someone like LeBron in this era, he simply wouldn’t be #1 most of the time in smaller samples like this. So, if anything, the RAPM data probably allows for an inference that Jordan was more dominant than LeBron in RAPM in his era. But that’s just an inference, and not something we can have a super high degree of confidence in, given the inherent lack of full data.*
So, if we can’t have a super high degree of confidence in a RAPM-based argument because of lack of data, then the additional “data” we’d have to look at would have to be box-score-based data. And, of course, Jordan fares extremely well there. He’s #1 all time in regular season PER, #1 all time in regular season win shares per 48 minutes, #1 all time in playoff win shares per 48 minutes, #1 all time in playoff BPM, and is #2 all time in both regular season BPM and playoff PER behind only Jokic. And then we also have the fact that Jordan had the highest Game Score (basically PER) in every playoff series he ever played, except barely being behind in two (and one was because they blew the other team out so much that he didn’t even play all that many minutes but was #1 in the series in per-minute Game Score; and the other was being barely behind Shawn Kemp in the 1996 Finals that Jordan won Finals MVP in). This is actually really remarkable—a consistency of playoff box-score dominance that is just completely unrivaled.
So the data-driven case is basically that he’s dominant in box-score stats in regular season and playoffs, and the limited impact data we have also supports the case.
____________________
* There’s also various With-Or-Without-You measures—some that are more raw WOWY, while others regress for individual player impact across all the various lineups that existed. These measures are all a bit flawed and aren’t something that most would use much for present-day players when we have better measures. In any event, in the interest of completeness, in the various regressed measures, Jordan looks really good all-time (ranked #4 all time based on an average of the various measures). He’s not #1 in these, so in a sense it doesn’t support a GOAT case, but he is above all the other main GOAT candidates, so in that sense it does support his case. Jordan is a lot lower in raw WOWY (32nd in prime WOWY), so that’s one data point that doesn’t support his case—basically the only one.
