As some may know, comprehensive play-by-play tracking began in 1997. This is why many stats say “since 1997”. Stuff like +/- with a player on vs off was not tracked back then, so impact metrics like RAPM can only start from 1997.

One reason this sucks is that it only captures the tail end of Jordan’s career, and certainly not his peak.

A researcher with the NBA known as Squared2020 has been conducting the monumental effort of recording play-by-play data from pre-1997 games by watching the broadcasts themselves.

The sample of games that he has completed tracking for includes 214/246 (87.0%) Bulls games in 1991, 1993 and 1996. That includes over 13,000 possessions with Jordan on the court.

Image that has the detailed results here, including some Grant/Rodman stuff, but a summary:

Jordan

• On (13250 poss.): 112.5 ORTG, 100.0 DRTG, +12.5 NET
• Off (2216 poss.): 100.0 ORTG, 101.4 DRG, -1.4 NET
• Difference: +12.5 ORTG, -1.4 DRTG, +13.9 NET

Pippen

• On (12956 poss.): 111.8 ORTG, 100.3 DRTG, +11.5 NET
• Off (2509 poss.): 105.4 ORTG, 100.0 DRTG, +5.4 NET
• Difference: +6.4 ORTG, +0.3 DRG, +6.1 NET

Jordan & Pippen

• MJ + Pippen (11452 poss.): 113.1 ORTG, 100.1 DRTG, +13.0 NET
• MJ + No Pippen (1798 poss.): 108.7 ORTG, 99.3 DRTG, +9.4 NET
• No MJ + Pippen (1504 poss.): 101.3 ORTG, 101.2 DRTG, +0.1 NET
• No MJ + No Pippen (712 poss.): 97.1 ORTG, 101.7 DRTG, -4.6 NET

Credit

Discussion

“Uh, so Jordan was good? I think we knew that?”

Fair, but having hard data is better than not, and allows us to contextualize things compared to modern stars.

Some other interesting points would be regarding Jordan vs. Pippen, and also the fact that neither of them actually seem to have a ton of defensive impact based on on/off. Do some people underrate Jordan’s offensive impact while maybe overstating his defensive impact? This data doesn’t tell us everything, but maybe something to consider.