Stop Fighting Probabilities: Category Win % Strategy

Through the All-Star break, Rudy Gobert is shooting 70.5% from the field and 49.4% from the free throw line. A manager who drafted Gobert already made an implicit decision about free throw percentage, whether they realized it or not.

The better framework treats each category as a probability to calculate. Pythagorean win expectation, borrowed from baseball sabermetrics, estimates the probability of winning a category based on how a team's output compares to the league average. The formula uses an exponent that varies by category: higher for low-variance categories like points, lower for high-variance categories like blocks.

The shape of the probability curve matters more than any individual number. Small changes near 50% produce large swings in win probability. The same improvement at 90% or 10% produces almost nothing.

Where marginal value lives

A roster projects to 75% win probability in steals and 48% in threes. Moving steals from 75% to 82% adds about 0.07 expected wins. Moving threes from 48% to 58% adds about 0.10. The player who helps the category closer to 50% delivers more total value, even if the raw stat line looks worse.

Every selection after the first few rounds should be evaluated by how much it shifts categories between 35% and 65%. Categories above 80% or below 20% deserve less weight. Adding blocks at 85% is diminishing returns. Trying to fix free throw percentage at 15% is a lost cause.

Category variance

Among the top 150 fantasy-relevant players this season, blocks have a standard deviation of 0.45 against a mean of 0.6 (CV: 0.75). Wembanyama at 2.7 blocks per game swings a team's output by four standard deviations. No other category offers that leverage from one player.

Points have a standard deviation of 5.1 against a mean of 17.7 (CV: 0.29). One elite scorer shifts the probability less, in relative terms, than one elite defender.

A manager at 40% in blocks can reach 65% with Evan Mobley (2.0 BPG). That same manager at 40% in points would need multiple upgrades for the same shift. High-variance categories are easier to dominate with one move, and easier to lose without one.

Category correlation

Giannis Antetokounmpo this season: 28.0 points, 10.0 rebounds, 65.1% FG, 64.7% FT. Drafting Giannis improves three categories and damages a fourth. Stephen Curry: 27.2 points, 4.5 threes, 45.3% FG, 93.4% FT. Different bundle, different trade-off. Of the top 40 scorers this season, 11 shoot below 47% from the field. The trade-off between volume and efficiency is structural.

The probability framework handles this. If adding Giannis pushes FG% from 52% to 90% win probability but drops FT% from 55% to 25%, that is a 38-point gain offsetting a 30-point loss. And at 25% FT, further damage costs almost nothing. Adding Alperen Sengun (70.0% FT) to a team at 15% in free throw percentage moves it to maybe 13%. The categories Sengun helps (9.4 rebounds, 1.0 steals, 1.0 blocks, 50.8% FG) are the ones that matter.

This is what punting actually is: a category's win probability has dropped low enough that further investment yields near-zero return.

Draft-day application

After each pick, recalculate projected output in all nine categories against the league average. Estimate win probability using the pythagorean formula. Sort by proximity to 50%.

The player who produces the largest net gain in total expected wins across all nine categories is the correct pick, regardless of generic rankings. Cason Wallace's 2.0 steals per game make him a mid-round value on most boards. For a team at 45% in steals, he might be the highest-impact player available. For a team at 80%, he offers diminishing returns.

Z-scores, static rankings, and generic draft strategies leave value on the table because they ignore roster context. Category win probability gives that context a number.

"Scottie Barnes (51914221392) (cropped)" by Erik Drost is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/.