RECAP · WNBA · 2026-07-21

Olivia Miles Went and Broke the Model (In the Best Way)

A two-game slate, 37 graded projections, a 5.9 MAE, and one night that reminded you why the injury-minutes edge exists — and why basketball still surprises everyone.

Arcline Analytics
00 · THE NIGHT

Two games. That's all you get on a night like this, and somehow it was plenty. Minnesota and Washington gave you a box score worth lingering over, while Golden State added a subplot that nobody fully saw coming. The injury news that trickled in before tip reshaped the minutes math in a hurry — and the players who absorbed those minutes, by and large, made good on the opportunity.

The theme of the night, if you had to name one: role players stepping into starring roles and a couple of stars quietly reminding you they don't need much of an invitation. Thirty-two points, eight rebounds, eight assists from a guard will do that. So will a 15-15 double-double from a center who just refuses to be invisible. It was, in the most honest sense, a good night of basketball — chaotic in the right ways, clarifying in others.

01 · OUR PROJECTIONS, GRADED

The honest one-number receipt first: a mean absolute error of 5.9 across 37 graded projections. That's a respectable night — not a perfect one, and we'll get to the misses — but the model found the right players in the right spots more often than not. Here's how it broke down, name by name.

The Top Scorers

  • Olivia Miles (MIN, $10,600) — Projected 44.2 DK points. Finished with 56. The line: 36 minutes, 32 points, 8 rebounds, 8 assists, 1 steal. An injury bumped her minutes heading into tonight, and she treated the extra opportunity like a closing argument. The model liked her; the basketball gods liked her more. A 11.8-point beat on a triple-double-adjacent night is the kind of outcome you frame.
  • Kiki Iriafen (WSH, $9,100) — Projected 29. Finished with 45.5. The line: 34 minutes, 23 points, 12 rebounds, 4 assists. No injury bump flagged here — this was pure outperformance, a 16.5-point gap that puts her at the top of the best-calls list and earns a full discussion below.
  • Kayla McBride (MIN, $9,700) — Projected 33.4. Finished with 42. The line: 32 minutes, 33 points, 2 rebounds, 2 steals. An injury had bumped her minutes, and she answered with the most efficient-looking line of the night — 33 actual points is a clean, no-fuss confirmation of what the model was pointing at.
  • Shakira Austin (WSH, $9,900) — Projected 32.5. Finished with 38.8. The line: 31 minutes, 15 points, 15 rebounds, 2 assists, 1 steal. A 15-15 double-double is the kind of stat line that wins you DFS lineups even when the points column looks modest. The model had her right, and the rebounding did the rest.
  • Kaitlyn Chen (GS, $5,300) — Projected 12.5. Finished with 34. The line: 24 minutes, 20 points, 2 rebounds, 6 assists, 1 steal. An injury bumped her minutes, and she turned 24 of them into the best value line of the slate — 6.42 DK points per $1,000 at that salary is tournament-winning territory. The gap between projection and outcome (21.5 points) is a miss in raw error terms, but the direction was right: she was flagged as a min-bumped name worth targeting.
  • Courtney Williams (MIN, $9,400) — Projected 35.9. Finished with 34. The line: 33 minutes, 21 points, 4 rebounds, 3 assists, 2 steals. An injury had bumped her minutes, and the projection was right on the ceiling. The actual scoring was lighter than hoped, but the minutes materialized and the floor held. A near-miss you can live with.

Best Calls

Three names the model earned tonight, in descending order of the outcome:

  1. Kiki Iriafen at $9,100 — the model liked her at 29 projected points; she put up 45.5. No injury bump, no special context — just a read on a player in a good spot. At $9,100 that's roughly 5.0 DK points per $1K realized. The 12-rebound, 4-assist complement to her 23 points is what pushed the DK total into another zip code. This is the call we feel best about tonight.
  2. Olivia Miles at $10,600 — projected 44.2, delivered 56. The injury-minutes edge was real here; the model caught the bump and priced the ceiling accordingly. She then exceeded even that elevated ceiling by nearly 12 points. At the top salary on the slate, she still delivered 5.28 DK points per $1K.
  3. Kayla McBride at $9,700 — projected 33.4, delivered 42. Another min-bumped player who cashed in. The model's projection was essentially confirmed by the raw points total (33 actual, 33.4 projected) — the DK overage came from efficiency bonuses and steals. Clean call.

Worst Calls

We own these plainly, because that's the deal.

  • Nia Coffey (MIN, $8,600) — Projected 25.9. Finished with 8.3. The line: 28 minutes, 3 points, 1 rebound, 1 assist, 1 steal. The minutes showed up — 28 of them, which is right in the window — but the production simply wasn't there. Three points on a night where teammates were scoring freely. A 17.6-point miss is the worst number on the board, and there's no clever framing for it. The model had her as a min-bumped beneficiary; the game had other ideas.
  • Gabby Williams (GS, $9,000) — Projected 28.5. Finished with 12. The line: 26 minutes, 8 points, 0 rebounds, 2 assists, 1 steal. Another min-bumped flag, another case where the minutes redistributed correctly and the stats didn't follow. Zero rebounds in 26 minutes for a player projected to contribute across the board is the part that stings most. A 16.5-point miss at $9,000 salary is the kind of outcome that costs lineups their GPP ceiling.
02 · WHAT WE LEARNED

A 5.9 MAE on a two-game slate with significant injury-driven minutes shuffling is a result the model can stand behind — the walk-forward validation holds up out of sample because the method is honest about what it knows and what it doesn't. Tonight reinforced both sides of that ledger.

The injury-minutes edge is real and worth pursuing. Four of the six top scorers tonight were flagged as min-bumped names before tip. Three of them — Miles, McBride, and Chen — outperformed projections. One (Williams) met expectations on minutes and missed on production. That's a three-for-four hit rate on the night's defining news edge, and it's the reason we surface those flags prominently in every column.

The honest lesson from Coffey and Gabby Williams is that minutes redistribution is a necessary condition for upside, not a sufficient one. When the minutes show up and the production doesn't, the model will miss — and a 17.6-point miss is significant. We'll carry that forward into how we weight stat-production consistency alongside the minutes signal, especially for role players on winning teams where usage can be volatile.

And then there's Kiki Iriafen, who needs no injury explanation and no special context — just a player in form, in the right matchup, doing her thing. Those are the projections that remind you the model's baseline read on talent and situation matters as much as the news edge. She was a best call before tip. She's the player of the night after it. Sometimes it's just that simple.

Back tomorrow with the next slate. We'll have the projections up when the injury reports settle.