Ciezki Came From Nowhere and Took the Whole Slate

One game on the slate — Toronto versus Phoenix — and it delivered. This wasn't a quiet Saturday night box score you scroll past. The injury news that came in before tip reshaped nearly every meaningful minute on both rosters, and the players who absorbed that playing time made the most of it in ways that will be fun to talk about for a while.

On the Toronto side, Isabelle Harrison turned in the kind of night you circle: 16 points, 10 rebounds, the block, the steal, 27 minutes, a complete performance. Kiki Rice ran the offense with quiet efficiency — 14 and 5 assists over 28 minutes. Phoenix countered with chaos in the best possible sense. Alyssa Thomas finished with 12 assists and looked like the engine of everything Phoenix was trying to do, even if the scoring wasn't there. And then there was Shay Ciezki, who cost $3,100 and proceeded to go out and post 15 points, 5 rebounds, 5 assists, a steal, and two blocks in 26 minutes. The night belonged to her.

The honest one-number receipt first: a mean absolute error of 8.7 DraftKings points across 16 graded players. On a slate this condensed — one game, significant injury-driven lineup flux on both sides — that's a number we're not proud of, but we'll own it plainly. Here's the full ledger on the top scorers.

• Isabelle Harrison (TOR, $9,100) — Projected 28.5, scored 36.0. Line: 27 min, 16 pts, 10 reb, 2 ast, 1 stl, 1 blk. An injury bumped her minutes, the model flagged it, and she delivered. The double-double with a block on top is exactly the ceiling scenario we were projecting into. Best call of the night, full stop.
• Shay Ciezki (PHX, $3,100) — Projected 7.1, scored 34.8. Line: 26 min, 15 pts, 5 reb, 5 ast, 1 stl, 2 blk. An injury bumped her minutes as well, but our model — and frankly anyone's model — had her at 7.1 because there was essentially no sample to work with at that salary. She went out and posted a stat line that would be impressive at three times the price. A 27.7-point miss. We'll say it plainly: we didn't project this. No one did. That's what a $3,100 ceiling game looks like when it lands.
• Alyssa Thomas (PHX, $12,200) — Projected 41.7, scored 30.5. Line: 22 min, 4 pts, 6 reb, 12 ast, 1 stl. An injury bumped her minutes into the projection, but the scoring didn't follow. The 12 assists are real — she controlled the game — but 4 points at $12,200 is the kind of mismatch that hurts DFS lineups anchored to her. We were high. The miss matters.
• Noemie Brochant (PHX, $6,300) — Projected 14.3, scored 28.5. Line: 25 min, 15 pts, 2 reb, 2 ast, 4 stl. An injury bumped her minutes, and four steals in 25 minutes is the kind of defensive explosion that's genuinely hard to model in advance. Significant outperformance at a salary that made the upside real.
• Kiki Rice (TOR, $9,600) — Projected 25.5, scored 25.0. Line: 28 min, 14 pts, 4 reb, 5 ast, 1 stl. Essentially a push. An injury bumped her minutes, the model priced in the opportunity, and she cashed almost exactly to projection. Quiet, steady, exactly what you want from a cornerstone salary.
• Sami Whitcomb (PHX, $5,500) — Projected 13.8, scored 24.8. Line: 16 min, 14 pts, 3 reb, 2 ast, 1 stl. An injury bumped her minutes. She made them count — 14 points in 16 minutes is efficient by any measure. An 11-point outperformance at a mid-range salary.

Best call: Harrison, cleanly. Worst call: Thomas on the miss, and Ciezki on the scale of the gap — though a 27-point swing on a $3,100 player says more about the volatility of the position than the model's blind spot.

The lesson from August 29th isn't complicated, but it's worth saying clearly: on a one-game slate, the injury news doesn't just matter — it's most of the model. Every player in our top-scorer table had their minutes bumped by an injury to a teammate. Every single one. The players who absorbed that opportunity split into two groups: the ones the model could project into (Harrison, Rice, Whitcomb — all within a reasonable range) and the ones whose pre-injury sample was so thin that any projection was essentially a guess (Ciezki at $3,100, Brochant's defensive explosion).

That's the honest takeaway. Walk-forward validation out of sample beats the naive baseline because it's systematically updating on information — and injury-driven minute redistribution is the clearest version of that edge. But there's a ceiling to what any model can do with a player who had almost no prior sample at this level of responsibility. The Ciezki line was a DFS miracle. You can position for the possibility of that outcome; you can't project it.

What you can project — and what the model did reasonably well — is identifying which established players would see meaningful usage bumps and roughly how productive they'd be. Harrison at 28.5 going for 36 is a good night. Rice at 25.5 going for 25 is a good night. The Thomas miss at the top salary is the one to carry forward: even with a minutes bump, a player's role can shift in ways that suppress scoring while preserving peripheral stats. Twelve assists is a real basketball accomplishment. It's a complicated DFS outcome. We'll remember that distinction.