Chasing Seven

The 49 number ones

The 644 can only cover one season at a time, because of the advanced analytics it uses. In order to see where Jokić sits in history, we need every season, so we rebuilt our model from Basketball Reference alone for every season from 1977–78 to the present (all the seasons with complete box-score data).¹ We call that version GRAVITY Historical, and it's the same historical model you see in our other articles whenever we need to compare the present to the past. A season's number one is simply the player with the highest impact per minute among players who played at least 1,500 minutes. All of the number ones are derived from the GRAVITY model, and not by any vote or league consensus.

To give you an idea of what the impact numbers mean: an impact of 0 is an average NBA minute, +1.3 is usually a top-ten player, and +2.0 is a top-six player. Every number one on the list had an impact of at least +2.2. Table 1 is the whole list. The last column shows where each number one ranked the following year. Looking at that list, the league kind of sorts itself into eras. There have been 16 number ones, and five of them account for 30 of the 49 seasons: nine times for Jordan, seven for LeBron, six for Jokić, and Bird and Shaq with four each.²

The runs

One season at the top could be luck, but a long run of them can't be. We have six players who have held the top spot for 3+ years in a row.

Each of these runs tells a different story. Jordan's run was the longest and the most dominant: he beat the second-best player by 0.61 on average. LeBron's run won the most MVPs, with four. Jokić's run is statistically “the best.” His six seasons have an average impact score of +3.36, above Jordan's +3.17 and LeBron's +3.19. Jokić's worst season at number one, +3.24, is higher than the best seasons Larry Bird, Shaq, or Julius Erving ever had. Jordan had one edge over Jokić: margin. Jokić led by just 0.10 and 0.15 in the last two seasons, with Shai Gilgeous-Alexander right behind him both times. Being the best in our model and being the most dominant are two different things, and Jokić can only claim one of them.

The voters

The MVP went to our number one in 24 of 49 seasons, and the typical winner ranked second in our data. In most years the winner stays within our top three, and the voters probably also weigh team success. However, some of the misses have been huge. Moses Malone won in 1978–79 while 14th in our rankings, and Steve Nash won in 2004–05 while ranked 12th.

Jordan was our number one in nine seasons and only won the MVP in four of those. Magic Johnson took three of the other five, and Charles Barkley and Karl Malone also took one each. Jokić has lost three of his six: to Joel Embiid in 2022–23 and to SGA in the last two seasons.³ Part of that is voter fatigue. Once a player has a few MVPs on the shelf, voters start looking for a new story, and Jokić already has three. By the standards of the last 50 years, that's completely normal. Our model's best player loses the vote roughly half the time, so if you think Jokić has been robbed, you're kind of wrong.

Why The 644 has it closer

Everything you read above comes from GRAVITY Historical. Our actual ranking, The 644, gives Jokić a rating of 97.8, 4.9 points ahead of SGA in second at 92.9.⁵ That's a tighter race than the historical model shows, for two reasons, and only one of them is a flaw. The flaw is ours: the model limits every input at four standard deviations above average, and Jokić's 2025–26 Box Plus/Minus sat at 4.66. The cap takes about 3 points off his rating.⁴ The other reason is fair. The model took away 3.4 points because of his playoff BPM (8.0 in a six-game series loss to Minnesota), which was well below his regular-season 14.2. Playoff performance is a real metric, and it should count. Cleaning the Glass, which filters out garbage time, shows the playoff dip clearly: his points per 100 shot attempts fell from 133.5 to 111.5 in the series, and he shot only 20% from three.⁷

The seventh year

Does a number one usually keep his spot? History gives three different answers, depending on which number ones you look at. Looking at all of them, our number one has kept that spot the following season 30 times out of 46. Looking only at the ones whose margin was 0.15 or less, like Jokić's is right now, the rate drops to 7 out of 11. Looking only at the ones who were 30 or older, it was just 3 out of 7, and Jokić is 31 this season.

The rivals matter more than the base rates do, so let's project each number one candidate from his 2025–26 season. Table 3 lays it out. If Jokić repeats his season, his gravity score rises to 100.7. If SGA repeats his, he scores 95.4. To reach a score of 100, SGA would need to improve his impact by 0.4, which is exactly the gap between him and Jokić last season. In other words, he would have to play like Jokić did. Victor Wembanyama jumped by 1.2 last season to reach +2.7. Another jump of 0.8 on top of that would get him to 98.6, still short of a repeating Jokić. The last row is the aging risk. A typical star loses 0.29 of impact between 30 and 31. If that happens to Jokić, his score next year is 97.4, which puts him ahead of an SGA who repeats and behind an SGA who improves.

We think Jokić has a 55% chance of getting his seventh year in a row as the GRAVITY number one. Our external prediction is that Jokić will have the best BPM in the league, which ironically he's also done six years in a row. The BPM leader has repeated 34 of 52 times, and 15 of 21 times once the leader had three or more titles in a row. We think this one has about a 60% chance. The betting market says his MVP chances are 17%, or +500 on DraftKings. We think it's really a 25% chance, considering we think he's more likely than not to be the best player in the league.⁶

External predictions are graded by April 30, 2027 (BPM) and June 30, 2027 (MVP). Registered October 17, 2026, before the October 18 freeze. Scored by Brier score in the Forecast Ledger.

Method and sources

GRAVITY Historical (GRAVITY-RS in the code): 0.533 z_BPM + 0.133 z_WS/48 + 0.333 S', with the specification's small-sample shrink, minutes-weighted moments over players with 200+ minutes, clipped at ± 4. Code: papers/common/gravity_rs.py, papers/jokic/history.py, papers/jokic/analysis.py.

1. Basketball Reference, league advanced and per-100 tables, 1977–78 to 2025–26. https://www.basketball-reference.com/leagues/
2. Basketball Reference, NBA MVP winners. https://www.basketball-reference.com/awards/mvp.html
3. Basketball Reference, 2025–26 awards voting. https://www.basketball-reference.com/awards/awards_2026.html
4. The Gravity Report, The GRAVITY Specification, v3. https://thegravityreport.com/notes/spec/
5. The Gravity Report, The 644, July 2026 edition (CSV). https://thegravityreport.com/the644/the644-2026-07.csv
6. VegasInsider, MVP odds (DraftKings, September 28, 2026). https://www.vegasinsider.com/nba/odds/mvp/
7. Cleaning the Glass, Nikola Jokić (garbage time filtered), accessed September 29, 2026. https://cleaningtheglass.com/stats/player/1883

Cite this article

Freedman, F. (2026). "Chasing Seven." The Gravity Report, GRAVITY Deep Dive No. 5.
https://thegravityreport.com/notes/jokic/