Spurs: the problem is the bench, not the starters

Update, November 14: this feature was first published on October 25, after three games and a 1-2 record. It has been rewritten with the first thirteen games (11-2). The original conclusions are recalled each time.

For his third season, Victor Wembanyama has a new starting five around him. Mitch Johnson decided to open games with his five best players, even if it meant three guards: De’Aaron Fox, Stephon Castle and Dylan Harper, alongside Devin Vassell and Wembanyama. After two losses to the Thunder (104-116) and the Rockets (110-114), San Antonio was worried. Then the Spurs won eleven straight, the longest active streak in the NBA. Before them, only the 2017-18 Celtics had strung together ten wins after starting 0-2.

To understand what changed and what didn’t, we looked closely at the minutes: by rebuilding the five players on the floor at every substitution, we know what each group actually produced.

• +250point differential with Wembanyama on the floor, in 381 minutes
• −78with him on the bench, in 248 minutes
• +86for the three-guard lineup, in 113 minutes together

It all still comes down to when Wembanyama sits

After three games, the two numbers cancelled out: +32 with him, −32 without. Thirteen games later, the number with him has exploded, and the one without him has gotten worse. Per 48 minutes, the Spurs win by 31 points with Wembanyama and lose by 15 without him. These are two different teams: the first is the best in the league, the second would be near the bottom.

With Wembanyama+250point differential in 381 min1004 scored, 754 allowed

Without Wembanyama−78point differential in 248 min549 scored, 627 allowed

The player-by-player picture hasn’t changed shape. All five starters are well in the black, and so are Julian Champagnie and Keldon Johnson, because they play a lot with Wembanyama. At the very bottom are Luke Kornet, the backup center (−51 in 195 minutes), and Carter Bryant (−51 in 137 minutes). This still isn’t a case against Kornet. It’s about the spot he fills: the giant’s, when the giant is resting.

On-court point differential, by player (13 games)

1. Victor Wembanyamastarter · 381 min+250
2. Dylan Harperstarter · 408 min+206
3. De'Aaron Foxstarter · 372 min+181
4. Devin Vassellstarter · 325 min+134
5. Stephon Castlestarter · 375 min+130
6. Julian Champagniebench · 338 min+86
7. Keldon Johnsonbench · 314 min+63
8. Harrison Barnesbench · 159 min+4
9. Tarris Reed Jr.bench · 54 min−30
10. Jordan McLaughlinbench · 66 min−33
11. Luke Kornetbench · 195 min−51
12. Carter Bryantbench · 137 min−51

The three-guard lineup: from “neither culprit nor savior” to wrecking ball

At the end of October, the starting five was +2 in 28 minutes, and we wrote that it was neither the culprit nor the savior. It is now +86 in 113 minutes. NBA.com noticed too, and put the Spurs at the top of its power rankings: that lineup gives up just 85.8 points per 100 possessions, the best defense in the league among lineups that have played at least 50 minutes together.

The shooting, our only concern, is holding up: the three guards are 56 of 164 from three (34.1%). Not brilliant, but enough, because everything else works. Harper is averaging 17.3 points, second on the team behind Wembanyama.

The most-used lineups

1. Wembanyama · Fox · Vassell · Castle · Harperthe starting five · 112.9 min+86
2. Barnes · K. Johnson · Bryant · Reed · McLaughlinwithout Wembanyama · 32.2 min−7
3. Wembanyama · Fox · Castle · Harper · ChampagnieChampagnie in for Vassell · 26.7 min+28
4. Vassell · Castle · Harper · Champagnie · Kornetwithout Wembanyama · 26 min−8
5. Wembanyama · Vassell · Castle · Harper · ChampagnieChampagnie in for Fox · 23.9 min+17
6. Wembanyama · Fox · Harper · Champagnie · K. Johnson2 starters on the bench · 20.4 min+36

The only two negative groups on the list are lineups without Wembanyama. One of them, Barnes, Johnson, Bryant, Reed and McLaughlin, mostly plays late in games that are already decided. The other, with three starters and Kornet, plays minutes that matter.

Fox: the probation is over, and he passed

On October 25, I wrote that I wasn’t convinced by De’Aaron Fox next to Wembanyama, and I gave him ten games, “not one more”. Thirteen games have gone by. Here’s what the numbers say, and they don’t help my case.

Fox scores less (13.6 points), but he has never been this efficient: 49.6% from the field, 13 of 40 from three (32.5%), 62.1% True Shooting, and 5.8 assists a game. Above all, he is Wembanyama’s best partner on the roster.

Fox and the three guards, depending on who's alongside

1. The three guards together139.9 min+111
2. Fox with Wembanyama287.6 min+228
3. Fox without Wembanyama84.5 min−47
4. Wembanyama without Fox93.5 min+22

When Wembanyama and Fox play together, the Spurs are +228 in 288 minutes. When Wembanyama plays without Fox, they’re only +22 in 94 minutes. In other words, almost all of Wembanyama’s positive differential has come with Fox next to him. There’s one number left on my side: without Wembanyama, Fox is −47 in 85 minutes. But those are the minutes of the whole bench, and we already knew they’re the team’s worst.

So I’ll admit it: on the court, Fox is the right point guard next to Wembanyama. I’m still free not to like him.

What we would do

• Stagger Wembanyama’s minutes, still. He plays 29 minutes a game: every minute without him costs a lot, every minute with him earns more than half a point.
• Find a backup center who protects the rim, or go smaller with Champagnie and Keldon Johnson when Kornet is struggling.
• Leave the starting five alone. We said it at 1-2 on principle; we say it at 11-2 with conviction.

Method: the five players on the floor were rebuilt from the substitutions in ESPN’s play-by-play (the maj:cinqs script). The per-player differentials obtained this way match the official plus/minus within 4 points. Sample: 13 games, from Oct. 20 to Nov. 13. Original version (October 25, 3 games): +32 with Wembanyama, −32 without him, +2 for the starting five.