Sixteen games on a Friday slate is a lot of surface area, and the run environments were noisy in both directions. The moneyline board came back 0-0 — no graded sides — so the night lived and died on totals, where we went 1-2. That is a losing night. The one win was clean; the two losses were not close. The under in TB @ TEX ended at 13 combined runs on an 8-run line, and the under in TOR @ KC finished at 11 on a 9.5 line. Neither game was decided by a bad-hop single. Both were just high-scoring baseball.

The one that went our way — ATL @ PHI over 6.5 — ended 5-2 for 7 total runs, which is about as clean a low-total win as you get. The problem with totals is that one blowout covers a lot of ground, and Friday gave us two of them on the wrong side of our leans.

No money was down on Friday. The card was empty — nothing cleared our staking threshold, so there is no official record to grade for the day. What we published were leans, tracked edges with no units attached. Here is how those leans settled.

The Leans

• ATL @ PHI Over 6.5 — WIN. Final: ATL 5, PHI 2. Seven total runs, just enough. The lean was right.
• TB @ TEX Under 8 — LOSS. Final: TB 7, TEX 6. Thirteen runs in a game lined at 8 is a rough night. The under had no chance after the first few innings.
• TOR @ KC Under 9.5 — LOSS. Final: TOR 9, KC 2. Eleven runs, and the lopsidedness of it — a 9-2 final — means the scoring was concentrated rather than a back-and-forth shootout. Either way, it cleared the line.

Leans went 1-2. Again, no units were at risk. The board reflects that accurately.

We scored 494 players on the DFS slate Friday, with 242 matched against our projections. The range of outcomes was wide enough to give you a headache and a reason to keep building rosters at the same time.

Top Scorers

• Brayan Rocchio (Cleveland Guardians) — 34.0 DK pts, 2-4, 2 HR, 5 RBI. Our projection: 5.37. Rocchio was a significant outperformer; the model did not see this coming at anywhere near this scale.
• Rafael Devers (San Francisco Giants) — 34.0 DK pts, 3-4, 1 HR, 5 RBI. Our projection: 7.66. Salary: $4,800. Devers tied for the slate lead in DK points and outperformed his projection by more than 26 points — at a salary that made him widely accessible.
• Jared Jones (Pittsburgh Pirates) — 33.1 DK pts, 6.0 IP, 9 K, 0 ER. Our projection: 14.09. Jones was a genuine model hit. The projection pointed toward him and he delivered nearly 20 points over it.
• TJ Rumfield (Colorado Rockies) — 29.0 DK pts, 3-4, 1 HR, 1 RBI. Our projection: 8.1. Salary: $4,400. Another low-salary outperformer. The model underestimated him significantly, but the salary made him easy to fit.
• Francisco Alvarez (New York Mets) — 28.0 DK pts, 2-4, 2 HR, 2 RBI. Our projection: 6.31. Salary: $3,200. Two home runs at the lowest salary on this list. The model projection does not explain how he got there.

Where the Projections Hit

Three players stood out as the model's best calls on Friday, relative to what they actually produced.

• Rafael Devers — projected 7.66, scored 34.0. The model liked him enough to surface him at a reasonable salary, and the production validated the direction even if it could not anticipate the magnitude.
• Jared Jones — projected 14.09, scored 33.1. This is the cleaner of the two pitcher calls. Jones was clearly identified as a strong arm to build around, and a six-inning, nine-strikeout shutout backed that up.
• TJ Rumfield — projected 8.1, scored 29.0. The model caught the right player even if it undershot the ceiling. At $4,400, the miss on magnitude is forgivable.

Where the Projections Missed

The worst calls on Friday were all starting pitchers, and they were bad in the way that hurts the most in DFS — early exits with large ERA damage.

• Logan Gilbert (Seattle Mariners) — projected 28.96, scored 10.5. Line: 4.7 IP, 10 K, 7 ER. Gilbert generated the strikeouts but gave up seven earned runs, and the model clearly did not price in that risk. A near-29-point projection settling at 10.5 is a significant miss.
• Logan Allen (Cleveland Guardians) — projected 14.09, scored -7.5. Line: 2.3 IP, 1 K, 5 ER. Negative DraftKings points. Two and a third innings. There is not much to analyze here beyond the outcome — the start was over before it started.
• Keider Montero (Detroit Tigers) — projected 13.76, scored -0.4. Line: 4.0 IP, 4 K, 6 ER. Another pitcher who gave up six earned runs and cost lineups points. Three of the day's worst performances belonged to starting pitchers, which is not a coincidence — Friday was hard on starters across the board.

The totals leans were 1-2, and both losses came from games that were not remotely competitive with the line. That happens. Thirteen runs in a game with an 8-run total is not a calibration failure — it is variance doing what variance does. The more useful pattern is that all three of Friday's worst DFS calls were starting pitchers, and the common thread was run suppression failing badly. When pitchers collapse early, the damage in DFS is compressive: you lose points, the game script changes, and the remaining lineup value often evaporates too.

Going forward, the pitcher projection process is the thing worth reviewing when we see a day like this. A near-29-point projection for Logan Gilbert that ends at 10.5 is a data point worth sitting with, not an excuse to abandon the model, but a reminder that pitcher upside projections carry meaningful downside risk that may not be fully priced in on high-variance starts. We will keep watching that spread between projection and outcome on the pitching side. The hitter side Friday was largely fine — the best calls were real, the misses were magnitudes, not directions.