James Harden at 37: take him at his late third-round price, and draft him for points, not assists

Draft James Harden where he is going, at the back end of the third round in 12-team leagues, in 9-category head-to-head. Do not reach above that, and do not let him slide past it. The arithmetic follows.

Format: 9-category head-to-head, 12 teams. Sample: one preseason game, 26 minutes. Minutes assumption: 32 a night, written down below so you can see where it breaks.

The number that surprises

24 points in 26 minutes. That is 0.92 points per minute, at age 37, in his first game of the preseason, starting next to another ball-handling guard in Donovan Mitchell.

The game was Cleveland 119, Orlando 105, on Sunday at Rocket Arena. It was Harden's preseason debut. He sat out Thursday's opener against Boston with right ankle soreness. He started with Mitchell, Peyton Watson, Evan Mobley and Jarrett Allen, and none of the five played in the fourth quarter. His 26 minutes were a starter's three quarters, not garbage time.

The rates

I took his line (24 points, 4 rebounds, 2 assists, 3 threes, 1 steal, 2 turnovers) and divided each number by 26 minutes:

• 0.92 points per minute
• 0.15 rebounds
• 0.08 assists
• 0.12 threes
• 0.04 steals
• 0.08 turnovers

The projection

Now the assumption: 32 minutes a night as a starter. Everything below hangs on it. Multiply the rates by 32 (I used the unrounded rates, then rounded each product to one decimal):

• 29.5 points, 4.9 rebounds, 2.5 assists, 3.7 threes, 1.2 steals, 2.5 turnovers

The range, because the minutes are a guess:

• At 28 minutes: 25.8 points, 2.2 assists
• At 34 minutes: 31.4 points, 2.6 assists

The scoring holds up in all three versions. The assists barely move, because the rate is the problem, not the minutes.

The honest knock

Two assists in 26 minutes, while Mitchell had six. If Harden's job in Cleveland is scorer first and table-setter second, do not draft him for dimes. At a late third-round price you are not paying for them, and that is why the price is fine.

What you are paying for is scoring volume with efficiency: 7 of 10 from the field, 3 of 5 from three, 7 of 9 from the line. One October game says little about his percentages for the season. Do not build your free-throw plan on 7 of 9.

Age and the ankle

He is 37 and he already missed a preseason game with ankle soreness. I am not projecting 82 games. I am projecting the per-game line when he plays. In head-to-head, a player who misses a week costs you that week, not the season, so the per-game line matters more than the games total.

Why the minutes are the whole column

A hypothetical guard who scores 0.50 points per minute shows the method. At 20 minutes a night he scores 10 points, and he is waiver-wire depth. At 34 minutes he scores 17, and he starts in 12-team leagues. Same player, different minutes, different roster decision.

The card

32 minutes, late third round, 9-category head-to-head. In points leagues he moves up, because the scoring carries more weight there. If you are punting assists in categories, he fits that build even better, since the category he is weakest in is the one you are giving away.

When to redo the math

Check back in two weeks. If his assists are still under four a game when the games count, the role is the role, and a late third-round price was still right for a scorer. If his minutes slip under 30, redo the projection. That is the deal on this desk: the assumption is written down, and you can see exactly when it breaks.