One Game, Six Stories, and a Bumpy Night for the Model

One game. That's what July 15 gave us — Indiana Fever against the Golden State Valkyries, tip at 8:00 PM ET, nowhere to hide. On a single-game slate, every projection is load-bearing. There's no spreading the variance across eight matchups, no quietly burying a miss in the footnotes. You grade every name in the box score, and the box score grades you right back.

Indiana's stars — Caitlin Clark, Aliyah Boston, and Kelsey Mitchell — were all on the floor, all productive by any reasonable basketball standard. Clark ran the offense and distributed the ball. Boston worked the glass and scored efficiently. Mitchell did what Mitchell does: got buckets. The Fever did their thing.

On the Golden State side, the story was shaped by minutes redistribution. An injury bumped playing time to Gabby Williams, Kayla Thornton, and Veronica Burton — three players whose roles expanded in ways the pre-game projections had to scramble to reflect. That's where the interesting stuff happened. And that's where a one-game slate gets complicated fast.

The honest one-number receipt for the night: a mean absolute error of 5.3 DraftKings points across 19 graded players. On a single-game slate, where variance is concentrated and there's no regression to a larger sample, that number stings a little. We'll own it. Here's the full picture, name by name.

The Indiana Stars

Kelsey Mitchell ($10,600) — We projected 32.3. She landed at 29.8 DK points: 35 minutes, 20 points, 1 rebound, 4 assists, 1 steal. A 2.5-point miss, which is about as close as this night got for us. Mitchell played big minutes, scored, and facilitated. The projection read the role correctly. This one we'll take.

Aliyah Boston ($11,200) — We projected 36.6. She finished at 28.8 DK points: 25 minutes, 15 points, 7 rebounds, 1 assist, 2 steals. A 7.8-point miss, and the minutes are the culprit — 25 is a softer load than what we were anticipating for a player at her salary and role. The production rate was there. The volume wasn't. That's a model miss and we'll say so plainly.

Caitlin Clark ($11,600) — We projected 37.0. She finished at 28.3 DK points: 26 minutes, 13 points, 3 rebounds, 6 assists, 1 steal, 1 block. An 8.7-point miss, the biggest on the Indiana side. The assist total was solid — six dimes is a good Clark night by distribution — but 13 points in 26 minutes is a lower-efficiency output than we built into the number, and the minutes were again lighter than projected. At $11,600, that's a tough outcome for anyone who rostered her in a GPP.

The Golden State Valkyries (Min-Bumped)

This is where the injury edge gets real — and where the grading gets nuanced. An injury bumped minutes for three Valkyries, which our model flagged. The question is always whether the model captures the new ceiling accurately.

Gabby Williams ($8,600) — We projected 27.9. She delivered 27.8 DK points: 27 minutes, 16 points, 3 rebounds, 1 assist, 3 steals, 1 block. A 0.1-point miss. That is, for lack of a better word, absurd. The model caught the minutes bump, priced the role correctly, and Williams delivered almost exactly what we asked for. That's the injury-edge working the way it's supposed to.

Kayla Thornton ($6,900) — We projected 19.4. She finished at 23.0 DK points: 31 minutes, 11 points, 8 rebounds, 1 assist. A 3.6-point beat — she outperformed the projection. Thornton played a full 31 minutes and vacuumed the offensive glass in a way that pushed her past our number. At $6,900, the value-per-dollar was excellent. This is the kind of min-bumped name that wins GPPs.

Veronica Burton ($9,000) — We projected 31.0. She finished at 20.8 DK points: 30 minutes, 4 points, 5 rebounds, 4 assists, 3 steals. A 10.2-point miss — the worst call of the night. The minutes came through. The steal total (3) was right in our wheelhouse. But four points on 30 minutes of floor time is brutal scoring efficiency, and the model didn't account for that kind of offensive quiet night. We caught the opportunity; the execution didn't cooperate. That's the honest version of this one.

A few things worth carrying forward.

First: single-game slates punish projection error without mercy. A 5.3 MAE across 19 players is the full truth of the night — no hiding behind a large slate, no counterbalancing hits softening the blow of the misses. The walk-forward validation that backs this model holds up at the season level, but individual nights are individual nights, and this one had some lumps.

Second: the min-bumped signal is real, and Gabby Williams proved it as cleanly as it gets — a 0.1-point miss on a role that expanded due to injury news. That's the live-information edge functioning correctly. The lesson isn't to abandon that signal; it's to pair it with honest ceiling-and-floor thinking rather than assuming the opportunity always converts to points. Veronica Burton had the minutes. She didn't have the scoring night. Both things are true.

Third: at the top of the salary range, minutes variance is the variable that breaks projections more than anything else. Both Aliyah Boston and Caitlin Clark came in lighter on floor time than we projected. When $11,200 and $11,600 players are playing 25 and 26 minutes respectively, the DK output is going to trail a projection built for something more. That's worth watching as we look ahead to how Indiana manages rotation load over the rest of the summer.

Kayla Thornton, at $6,900, beating the projection by 3.6 points with 31 minutes of work? That's the reminder that the best place to absorb high-end misses is at the value tier. Build your GPP lineups accordingly.