Three games, first tip at 8:00 PM ET, and by the time the dust settled this slate had reminded you of something basketball does on a semi-regular basis: the player you least expected to be the story usually is. Fifty-two players graded out. The box scores showed a night defined by load redistribution — injuries quietly reshaping who touched the ball, who stayed on the floor, and who cashed. Minnesota ran three players into the top five in real DK points, which either means they had a great night or their opponent had a rough one, and honestly it might be both. Dallas leaned hard on Paige Bueckers and got a performance that looked exactly like what happens when a star gets a green light and uses every minute of it. And over in the Toronto game, a couple of quieter names outplayed their price tags in ways that made the value-hunting part of this exercise feel worthwhile. It was, in short, a good night to have done the homework.
Alanna Smith Broke Our Model and We're Fine With It
A 7.1-point MAE, a $6,300 big who nearly hit 39 DK points, and a bumped-minutes story that played out almost perfectly — here's the full honest receipt from August 12th.
The one-number receipt first: a mean absolute error of 7.1 points across 52 graded players. That's the whole honest story in a single figure — some nights it's tighter, some nights it's wider, and this one landed in a range we'll take. Now let's go name by name.
The Top Scorers
Tiffany Hayes (GS, $6,600) was the wildest line on the board. 17 minutes, 22 points, 6 rebounds, 5 assists, 3 steals, a block — 44.5 DK points. Our projection: 15.9. Yes, that's a miss of almost 29 points, and yes, it's going in the worst-call column in spirit even though she's listed here among the top scorers. We projected a modest night and she delivered a masterclass in a short burst. That's a 6.75 DK points-per-$1K return on a night where she was barely expected to crack double figures in projection. Sometimes the basketball gods are just laughing at your model and you have to tip your cap.
Then there's Paige Bueckers (DAL, $10,700). An injury bumped her minutes heading into the game, and the model caught it — 37.3 projected, 43.8 delivered. 34 minutes, 22 points, 3 rebounds, 8 assists, 3 steals. That's a 6.5-point miss on the right side of the ledger, at a salary where the upside justified the spend. We'll take it.
Napheesa Collier (MIN, $11,200) also came in with an injury-bumped projection — 40.7 projected, 42.8 actual. 33 minutes, 20 points, 9 rebounds, 2 assists, 2 steals, 3 blocks. A 2.1-point miss at the top of the salary range is about as clean a projection as you're going to see on a three-game slate. The bumped-minutes read held up.
Natasha Howard (MIN, $9,100) was another min_bumped player who delivered — 33.5 projected, 39.0 actual. 28 minutes, 14 points, 12 rebounds, 4 assists, 2 steals. The double-double was there, the value was real at that price point, and the injury news that inflated her role paid off.
Now about Alanna Smith (DAL, $6,300). We had her at 17.8. She scored 38.8 DK points. 33 minutes, 13 points, 11 rebounds, 4 assists, a steal, a block. That's a 21-point miss, the largest on the board by a significant margin, and it happened to be one of the best value performances of the night — 6.16 DK points per $1K from a player projected to be essentially a roster-filler. Her minutes were bumped by an injury and the model saw some upside, but not that much upside. She found a double-double, found assists, found blocks, and found a way to embarrass a projection that simply did not have her ceiling mapped correctly. Noted for future reference.
Kayla McBride (MIN, $9,600) closed out the top six — 31.3 projected, 34.8 actual. 32 minutes, 19 points, 1 rebound, 2 assists, 4 steals, a block. A clean 3.5-point miss, steals doing most of the work to push her over projection. Solid.
Best Calls
Laura Juskaite (TOR, $7,500) was the quiet value win of the night. We projected 20.2; she delivered 32.3. 32 minutes, 18 points, 7 rebounds, 3 steals. That's 4.31 DK points per $1K from a player priced like a role player who played like a contributor. The model liked her in this spot and she came through — a 12.1-point beat on a mid-salary target is exactly the kind of result that builds a tournament lineup.
Bueckers again, because she deserves the second mention: 6.5 points over a 37.3 projection with bumped minutes is the clearest example of the injury-edge read working as intended.
Isabelle Harrison (TOR, $6,200) rounded out the best calls — 22.0 projected, 28.0 actual. 30 minutes, 4 points, 10 rebounds, 5 assists, a steal, 2 blocks. The points weren't there but the everything-else was, and at $6,200 she returned 4.52 DK points per $1K. Floor plays at the bottom of a salary range doing their job.
Worst Calls
Own the misses. That's the deal.
Azura Stevens (CHI, $8,700) was projected at 27.9. She scored 7.8 DK points. 17 minutes, 4 points, 3 rebounds, 1 assist. Her minutes were flagged as bumped by an injury — and somehow that bump went the wrong direction entirely, or the situation changed post-projection. A 20.1-point miss on a player at nearly $9K is a costly construction error, and there's no way to dress it up. The injury news pointed one way; the game went another.
Courtney Williams (MIN, $9,000) — 30.4 projected, 17.3 actual. 30 minutes, 10 points, 1 rebound, 4 assists. She played the minutes; the production just wasn't there. No injury excuse on this one — the model simply overestimated her output in a game where Minnesota had plenty of mouths to feed. A 13.1-point miss at that price hurts.
Nyara Sabally (TOR, $8,400) — 26.2 projected, 14.8 actual. 17 minutes, 6 points, 7 rebounds, 1 assist. The minutes were the story — she was projected for more and didn't get them. A 11.4-point miss on a player priced in the upper-mid range is the kind of outcome that sinks a cash lineup. We had her wrong.
Three things worth carrying forward.
First: the bumped-minutes signal is real, but it's directional — not deterministic. Bueckers, Collier, Howard, and Smith all came through when their minutes were redistributed upward. Stevens did not. The lesson isn't to fade the signal; it's to recognize that the signal tells you where opportunity is being created, not that the opportunity will be seized. The model's job is to identify those spots. What happens inside the game is still basketball.
Second: the low-salary, high-upside tier is where slates get won or lost in tournaments. Alanna Smith at $6,300 returning 38.8 DK points is the kind of outcome that wins a GPP. Our projection of 17.8 meant most people missed her — us included, in terms of properly weighting her ceiling. The model had her as a bumped-minutes play; it just didn't have her ceiling high enough. That's a calibration note, not a disqualification of the approach.
Third: a 7.1 MAE across 52 players on a three-game slate is a number we can work with. The model is walk-forward validated — it beats the naive baselines out of sample — and nights like this, with a handful of real misses and a handful of genuinely good reads, are what that validation looks like in practice. Not magic. Honest edge, applied consistently, graded openly. That's the whole idea.
We'll be back tomorrow with the next slate. The misses are already logged.