Bonner Went Off and the Model Caught Most of It

One game on the slate — Phoenix hosting Golden State — and it delivered. The box score reads like a night where two veterans decided they weren't going anywhere quietly. Alyssa Thomas put up 25 points, 9 rebounds, and 8 assists in 38 minutes, which is the kind of line that reminds you she is, quietly, one of the most complete players in this league. DeWanna Bonner went for 17 and 10 with three blocks off the bench — or wherever she was operating from — and completely scrambled what anyone expected from her at that salary.

Golden State got contributions scattered across the roster. Veronica Burton was steady and efficient in 34 minutes. Cecilia Zandalasini carved out 28.3 DraftKings points in just 23 minutes — an injury bumped her into a bigger role and she made the most of every possession. Janelle Salaun was in a similar spot, also picking up additional run because of injury, though her final line came in softer than expected. Kaila Charles was the quiet double-double — 10 and 10 in 29 minutes at a salary that nobody was probably thrilled about at lineup lock.

It was a single-game slate, which means every projection error gets magnified and every hit feels a little sweeter. Seventeen players graded. Let's get into it.

The honest one-number receipt first: a mean absolute error of 9.1 DraftKings points across 17 graded players. On a single-game slate where variance runs hot, that's a number worth sitting with — not a celebration, not a disaster, just the truth.

Where the Model Hit

Alyssa Thomas was the anchor call. We projected 38.4 DK points; she finished at 48.3 on a line of 38 minutes, 25 points, 9 rebounds, 8 assists, and a block. We were directionally right on the role and the ceiling — the actual line just ran a bit hotter than the number. At $11,500, she was the most expensive player on the slate, and she earned every dollar of it.

Veronica Burton was about as clean a projection as you'll get. We had her at 28.1; she finished at 28.8 on 34 minutes, 15 points, 3 rebounds, 5 assists, and a steal. That's basically a push, and in a model context, a push on a volatile asset is a win.

DeWanna Bonner was the best value call of the night, even if the model didn't fully see the ceiling coming. We projected 22.3; she finished at 44.0. The stat line — 17 points, 10 rebounds, 3 assists, a steal, and 3 blocks in 34 minutes — at $8,000 salary works out to 5.5 DK points per $1,000. The model correctly identified her as a high-floor player at that price. The triple-block performance took her somewhere we didn't project, but we were in the right neighborhood.

Where We Missed

No entries in the worst-calls list tonight, which is worth noting plainly: the model's misses were errors of magnitude, not direction. The players it liked, it liked for the right reasons. The gaps between projection and actuals — most visibly the 21.7-point miss on Bonner and the 9.9-point miss on Thomas — are real, and they account for the 9.1 MAE. We're not burying that.

The Injury-Bumped Players

Cecilia Zandalasini had her minutes bumped by an injury, and it showed: 28.3 DK points against a 17.4 projection on just 23 minutes. Sixteen points, three steals, and three assists in limited run — that's exactly the kind of elevated efficiency the model looks for when minutes redistribute. We got the direction right; the per-minute output exceeded what we modeled.

Janelle Salaun was also playing on injury-bumped minutes and came in at 22.5 DK points against our projection of 26.6. She had 13 points, 3 assists, and a steal in 24 minutes. The model was a bit ahead of where she actually landed — a miss in the right direction, but a miss.

Kaila Charles rounded out the top scorers at 32.0 DK points against our 14.6 projection. The double-double — 10 points, 10 rebounds, 2 steals in 29 minutes — at $5,800 was the kind of performance that wins GPP lineups. At 5.5 DK points per $1,000, she was the best value on the slate. We were too low. Own it.

A few things worth carrying forward.

First, the model is walk-forward validated — it beats naive baselines out of sample — and tonight's results are consistent with that. The directional calls held up. The floor players hit their floors or better. The ceiling misses, where they happened, were on players doing things (three blocks from a forward, a double-double from a value big) that are inherently hard to project with precision. That's not an excuse; it's context for what a 9.1 MAE means in practice.

Second, the injury-minutes-bump signal continues to be real. Zandalasini's line is evidence of it. When minutes redistribute late, the model flags the beneficiaries — and the upside of acting on that information is exactly what we saw tonight. The miss on Salaun is a reminder that not every bumped-minutes player converts at the same rate; role context and matchup still matter inside that signal.

Third, and maybe most importantly: DeWanna Bonner at $8,000 was a reasonable projection that turned into a great night because she did things she doesn't do every game. The model can't fully price in the 90th-percentile version of a player. What it can do is tell you when the price is right for the median version, and the median Bonner at $8,000 was already worth it. The rest was bonus.

We'll see the full grades when all results are confirmed. The slate gave us something real to learn from, which, honestly, is all you can ask.