Malonga Went Off. Carrington Went Off. The Model Half-Saw It.

One game on the board Monday — Seattle hosting Chicago — and whatever you were expecting from this slate, you probably weren't expecting to be humbled quite this thoroughly. Injury news reshuffled the minutes distribution before tip, and then the box score proceeded to do what box scores do when lineups shift: it rewarded the players who were ready for expanded roles and quietly swallowed a few who weren't.

The standout performance came from Dominique Malonga, the Seattle big who dropped 26 points, 10 rebounds, 2 steals, and 3 blocks in 30 minutes. That is a complete game. That is the kind of line that makes you check the box score twice to make sure you read it right. But she wasn't alone at the top. Over on the Chicago side, DiJonai Carrington put together one of those quiet-until-it-isn't nights — 26 points, 6 rebounds, and 5 steals in 27 minutes at a salary that made every DFS lineup she anchored look extremely smart in hindsight. Two players, expanded opportunity, and two lines that defined the evening.

Across 18 graded players, the model finished with a mean absolute error of 10.5 DraftKings points. That's the honest one-number receipt. Some nights that number reflects a model that was broadly off; tonight it reflects a model that had two or three big misses pulling the average in a direction the overall picture doesn't quite deserve. Here's how it broke down.

The Top Scorers

• Dominique Malonga (SEA, $10,700) — Projected 39.5, delivered 52.0. Line: 30 min, 26 pts, 10 reb, 1 ast, 2 stl, 3 blk. An injury bumped her minutes, and the model caught the direction — it had her as a top projection on the slate. The gap between 39.5 and 52.0 is real, but the arrow was right, and that matters for lineup construction.
• DiJonai Carrington (CHI, $5,000) — Projected 19.7, delivered 45.0. Line: 27 min, 26 pts, 6 reb, 1 ast, 5 stl. An injury bumped her minutes too, and the model... underestimated what she'd do with them. That's a 25-point miss on a player at minimum salary. We'll get into this more below.
• Natisha Hiedeman (SEA, $9,200) — Projected 34.2, delivered 34.0. Line: 33 min, 18 pts, 2 reb, 9 ast. An injury bumped her minutes and the model was nearly perfect. Two-tenths of a point off. That's the kind of result that makes you feel good about the process even on a rough overall night.
• Kamilla Cardoso (CHI, $9,500) — Projected 44.3, delivered 29.8. Line: 27 min, 15 pts, 13 reb, 0 ast. An injury bumped her minutes, but the actual output didn't follow the projection. She rebounded well — 13 boards is a real contribution — but the scoring wasn't there, and the model's 44.3 projection looks ambitious against a 29.8 result.
• Azura Stevens (CHI, $8,900) — Projected 36.6, delivered 25.5. Line: 24 min, 9 pts, 8 reb, 2 ast, 2 blk. An injury bumped her minutes, and she was solid without being spectacular. An 11-point miss.
• Natasha Cloud (CHI, $8,200) — Projected 30.4, delivered 24.0. Line: 32 min, 7 pts, 2 reb, 8 ast, 2 stl. An injury bumped her minutes. The assists were there; the scoring wasn't. A 6.4-point miss — the smallest of the overestimates in this group.

Best Call

• Dominique Malonga (SEA, $10,700) — The model's best call on the night, and it earns the label honestly. A projection of 39.5 on a player who finishes at 52.0 isn't a perfect number, but it's a correct conviction. She was flagged as a top-tier play, the injury news reinforced the minutes picture, and she delivered at 4.9 DK points per $1,000 in salary. When you're right on the player, the ceiling sometimes just clears whatever ceiling you imagined.

Worst Calls — Owned Plainly

• Awa Fam (SEA, $7,900) — Projected 27.8, delivered 6.5. Line: 14 min, 0 pts, 0 reb, 2 ast, 2 stl. An injury bumped her minutes in the model's accounting, but she finished with just 14 minutes and a blank stat line where the points and rebounds were supposed to be. A 21.3-point miss. That one stings, and there's no dressing it up.
• Elizabeth Williams (CHI, $5,600) — Projected 27.4, delivered 6.8. Line: 18 min, 2 pts, 1 reb, 1 ast, 1 stl. An injury bumped her minutes, but she contributed almost nothing in 18 minutes of run. A 20.6-point miss. Two players with bumped minutes projections who gave us almost nothing — that's a pattern worth examining.
• Courtney Vandersloot (CHI, $7,700) — Projected 31.1, delivered 17.0. Line: 26 min, 13 pts, 0 reb, 3 ast. An injury bumped her minutes. The minutes were there — 26 is a reasonable night — but the production didn't follow. A 14.1-point miss, and the assist total of 3 is notably low for a player at her position and salary.

Tonight illustrated something that matters for how you use a projection system: the minutes bump is a real edge — Malonga and Hiedeman proved that — but minutes bumped does not automatically mean production bumped, and Fam and Williams are the cautionary side of that same coin. The model correctly identified which players were going to see more run. It was less reliable in projecting how efficiently each player would convert that run into fantasy points.

The Carrington situation is the most instructive data point of the night. A 25-point miss on a minimum-salary player who finished with a 45-point line — that's not just a projection error, that's a lineup-construction miss, the kind that flips GPP contests. The model bumped her minutes based on the injury news and still projected 19.7, which means it wasn't giving her enough credit for what she could do with an expanded role. That's a calibration note, and it's going into the work for tomorrow's slate.

The model is walk-forward validated — it beats naive baselines out of sample over time — and a 10.5 MAE on a one-game slate with five players receiving injury-driven minute bumps is not a number to panic over. But it's also not a number to wave away. The right response is what it always is: look at the misses honestly, figure out where the process broke down, and adjust. Fam and Williams in particular deserve a closer look at how the model allocates production among multiple players receiving bumped minutes on the same team. When the pie expands for several players at once, the model may be overallocating to everyone rather than identifying the primary beneficiary. That's the work.

Malonga was the story of the night. Carrington almost stole it. The model caught one and missed the other, and that gap — between a player who was already projected highly and one who was undervalued right up until tip — is exactly the kind of thing worth thinking about heading into the rest of the week.