Pete Crow-Armstrong Went Supernova and the Unders Held Up

Seventeen games on a Saturday in late August, and the slate basically sorted itself into two categories: pitchers who came to work and offenses that either erupted or went completely dark. No in-between. No quiet 4-3 affairs. The run environments were polarized, which is exactly the kind of slate where totals betting either looks genius or gets humbled — and Saturday was mostly the former.

We tracked four totals edges on the board. Three of them cashed. The one that didn't — we'll get to it, and it's worth talking about honestly. On the moneyline side, nothing was staked and nothing was graded. The totals record for the day: 3-1. That's a good Saturday.

No money was officially on the card Saturday. Nothing cleared our threshold for a staked play, so the ledger stays clean — no units won, no units lost. That's not a cop-out; that's the discipline. The leans, though, were live on the board and they're worth grading because that's how we build toward the plays that do make the card.

The Leans

• BOS @ NYY — Lean: Under 8. Final: BOS 6, NYY 0. Six total runs, a shutout in the Bronx. WIN. Jake Bennett gave Boston six innings of nothing, and the Yankees offense obliged by producing exactly that. The under never felt threatened after the third inning.
• MIA @ WSH — Lean: Over 7.5. Final: WSH 5, MIA 4. Nine total runs, cleared the number with room. WIN. A back-and-forth ballgame in Washington that did exactly what an over needs to do — stayed messy and productive the whole way through.
• BAL @ OAK — Lean: Over 9.5. Final: OAK 3, BAL 5. Eight total runs, short of the number. LOSS. This one simply didn't get there. Both offenses found ways to strand runners and the pitching held. We had the over, the game had other ideas. Own it.
• ARI @ SF — Lean: Over 7.5. Final: SF 7, ARI 2. Nine total runs, cleared the number comfortably. WIN. The Giants did the heavy lifting on their side of it, and Merrill Kelly was sharp for Arizona even in a loss.

Three wins, one loss on the tracked board. Not staked, but tracked — and the tracking is what builds confidence in the plays that do eventually make the card.

Okay. Let's talk about the Chicago Cubs, because whatever happened in their game Saturday deserves its own paragraph before we get into the numbers.

Top Scorers

• Pete Crow-Armstrong, Cubs — 53.0 DraftKings points. The line: 4-for-6, 3 home runs, 6 RBI. Our projection had him at 9.89 points. He scored 53. That's not a miss, that's a different sport. When a player goes 4-for-6 with three home runs in a single game, no model on earth is projecting that — you're praying for it to land in your lineup. If you had him, you had a very good Saturday. If you didn't, you watched it happen from the outside like the rest of us.
• Carson Kelly, Cubs — 31.0 DraftKings points. Line: 2-for-4, 1 home run, 4 RBI. Our projection: 7.14. Another Cubs hitter, another significant outperformance. There was something happening in that game.
• Merrill Kelly, Arizona Diamondbacks — 30.1 DraftKings points. Line: 7.0 IP, 7 strikeouts, 0 earned runs. No projection on our end for Merrill Kelly. A gem of a start in a losing effort — that's the most thankless line in baseball, and he put up a 30-point DFS day anyway.
• Cristopher Sanchez, Philadelphia Phillies — 29.9 DraftKings points. Line: 7.0 IP, 8 strikeouts, 0 earned runs. Our projection: 22.65. This one we actually saw coming. Sanchez was projected for a strong outing and delivered — 22.65 projected, 29.9 actual. That's a hit.
• Jake Bennett, Boston Red Sox — 29.7 DraftKings points. Line: 6.0 IP, 7 strikeouts, 0 earned runs. No projection on our end. Bennett's performance was the engine behind that BOS-NYY under cashing so cleanly.

Where We Hit

• Pete Crow-Armstrong — Projected 9.89, scored 53. We had him flagged. Nobody projects a 53-point day, but flagging the right players for the right spots is still a win for the process.
• Carson Kelly — Projected 7.14, scored 31. Same story. The Cubs lineup was the place to be and the model pointed there.
• Alex Bregman, Cubs — Projected 8.68, scored 28.0. Line: 3-for-5, 1 home run, 3 RBI. Three Cubs hitters in the best-calls column. The model liked the Cubs' offensive environment Saturday and the Cubs' offense rewarded that read.

Where We Missed

• Bennett Sousa, Houston Astros — Projected 13.52, scored -2.3. Line: 0.7 IP, 0 strikeouts, 1 earned run. Two-thirds of an inning. Minus points. The model liked his spot and he didn't make it out of the first. That's as bad as it gets for a pitcher projection and we're not going to dress it up.
• Alec Gamboa, Red Sox — Projected 13.27, scored 4.4. Line: 1.3 IP, 1 strikeout, 0 earned runs. The zero earned runs is a mercy. 1.3 innings is not what we needed from a 13-point projection.
• Mitch Bratt, Diamondbacks — Projected 12.11, scored 4.8. Line: 4.3 IP, 2 strikeouts, 3 earned runs. Didn't make it through five innings and gave up three. The projection was wrong. Saturday's best-calls list was dominated by hitters; the worst-calls list was all pitchers. Make of that what you will going into next week.

The honest takeaway from Saturday is that the totals model had a good day and the pitcher projection model had a rough one, and both of those things are true at the same time. Three leans went green, the card stayed empty because nothing cleared the staking bar, and a Cubs outfielder went nuclear in a way that no projection system is designed to anticipate.

The pitcher misses are worth watching. Sousa, Gamboa, and Bratt all had spots the model liked — bulk innings, favorable matchups, something in the data pointed toward productivity. All three gave back the ball faster than expected. That's a pattern worth noting heading into the final month of the regular season, when bullpen usage gets unpredictable and starter leashes get shorter. We'll be recalibrating the early-exit risk component of our pitcher projections for next week.

On the totals side: the 3-1 day reinforces that the model is finding real edges in run environment, and the discipline of keeping nothing on the card when the staking threshold isn't met is exactly the right call on a day where our best read was a lean, not a conviction. That distinction matters. We'll be back tomorrow with the Sunday card.