Daily Bet Research Report — Friday, September 4, 2026

Report generated at 4:23 PM ET on Friday, September 4, 2026

Bankroll: $23.91 | Stake (20%): $4.78 Record: 46-61-3 | Hit rate: 43.0% (46 of 107 decided; 41.8% including pushes)

Bottom line: no qualifying plays today. The full slate was reviewed with confirmed starters and live pricing. Every game failed on a documented rule or on negative edge once recent form was applied. Details in Sections 3 and 4.

Drawdown note: bankroll peaked at $54.33 on 8/30 and sits at $23.91 — a 56% drawdown over five days. Per the Strategy tab's drawdown rule, the bar for max-EV variance plays is raised today, not lowered.

Section 2: Schedule & Summary Analysis

All times ET, confirmed against MLB.com's official schedule (which publishes in ET) and cross-checked on RotoWire. Est. Odds are the best available decimal price across the four books RotoWire lists, favorite side.

MLB is the only active league on the board today. NFL opens next week; NHL/NBA/CBB are out of season.

Section 3: My Recommendations

No qualifying plays today — full slate reviewed, no edge cleared the bar.

This was a reviewed slate, not an unreviewable one. To be explicit about what was actually verified:

• All 16 games identified, start times confirmed in ET from MLB.com's official schedule.
• Every probable starter cross-confirmed on two independently opened sources (MLB.com probable-pitchers page and RotoWire daily lineups), with matching season lines.
• Live moneyline pricing pulled for all 16 games across four books.
• Last-5-start game logs pulled from the MLB Stats API for all seven pitchers central to the candidate plays.
• Last-10 and home/away splits pulled for all 30 teams.

Nothing cleared positive edge after that work. No Kelly Weight, Stake %, or Stake $ is allocated — $0 of the $4.78 daily budget is deployed today. Rest day.

Three games (ATL @ PHI, CHC @ MIA, AZ @ HOU) did show the underdog as the genuine value side, but RotoWire publishes only the favorite's price, and the one attempt at a two-sided board (covers.com) returned stale data listing entirely different pitchers — Gausman, Skubal, Alcantara — for today's matchups. That contradicted both verified sources, so it was discarded rather than used. I will not quote a dog price I did not actually see. If your book shows ATL, MIA, or AZ at a real number, those are the three worth pricing yourself — see §4.

Section 4: Reasoning

With no Section 3 recommendations, this section documents the four candidates that came closest and the specific math that disqualified each. Every pitcher line below is from a game log I pulled this run, not a season aggregate.

Candidate A — CLE ML (G2), 1.73 — the day's best template match, killed by weather

The math. Implied % at 1.73 = 57.97%. Base log5 from season records (CLE .500, DET .460) = 54.0%, plus ~3.5% home field = 57.5%. Adjustments: pitching edge +5.5%, DET's weakness vs LHP +2.5%, DET's recent form +1.5%; minus CLE's poor home record −3.0%, doubleheader/bullpen state −2.0%, rookie-unknown uncertainty −1.5% → 60.5% pre-weather. Rain-delay and postponement risk discounts that to ~58.0%. EV = (0.58 × $3.47) − (0.42 × $4.78) = $2.01 − $2.01 = +$0.00. Edge ≈ 0%. Not a play.

The evidence. This matched the template on paper: home favorite, pitching edge, team-quality edge. Foster Griffin (LHP) is 15-4, 3.21 ERA on the season; over his last 5 starts he is 3-1 with a 3.96 ERA (25.0 IP, 11 ER) — modestly worse than his season line but a 0.75 divergence, under the flag threshold. Detroit counters with Andrew Sears, who has two career MLB appearances — 8/23 (4.0 IP, 5 ER) and 8/29 (2.2 IP, 0 ER), 6.75 ERA in 6.2 innings. A last-5 cannot be stated for him because it does not exist, and that 2.2-inning outing suggests a short leash or a bullpen game. Team quality: CLE 70-70, last 10 6-4; DET 64-75, last 10 3-7. Detroit is 17-26 vs left-handed pitching and Griffin is a lefty.

Risk factors. The disqualifier. MLB.com's probable-pitchers page labels the 2:10 PM opener "delayed", RotoWire's Progressive Field weather shows 25% rain, and the Stats API still reports all 16 games as "Scheduled" with zero in progress — meaning G1 has not completed. A nightcap whose opener is sitting in a rain delay carries unknown start time, unknown bullpen usage on both sides, and live postponement risk. That uncertainty is worth more than the ~2.5% edge I had. Separately, Cleveland is 33-38 at home — a real counterweight to the home-field assumption — and is on an L2 streak.

Source. MLB.com Probable Pitchers, Sept 4 2026 — carries the "delayed" tag on G1 and both starters' season lines.

Candidate B — CWS ML, 1.87 — killed by the recency check

The math. Implied % at 1.87 = 53.5%. Season log5 (CWS .521, MIN .479) + home field = 57.7%. But the recency read reverses the pitching edge, worth about −5% → 52.7%. Edge = −0.8%. EV = (0.527 × $4.16) − (0.473 × $4.78) = $2.19 − $2.26 = −$0.07. Negative. Not a play.

The evidence. On season lines this looked like the best price on the board: Erick Fedde 4.16 ERA vs Zebby Matthews 4.58, White Sox 41-27 at home, Twins 29-39 on the road, near even money. Recent form flips it. Matthews' last 5 starts: 5-0, 2.90 ERA (31.0 IP, 10 ER) against a 4.58 season mark — a 1.68 divergence, which clears the meaningful-divergence flag, in his favor. His most recent start, 8/30, was 8.1 IP and 1 ER against these same White Sox. Fedde's last 5: 4.58 ERA (19.2 IP, 10 ER) versus a 4.16 season line — drifting the wrong way, though his last three (13.2 IP, 2 ER) have been sharp. Team form: CWS 73-67, last 10 5-5; MIN 67-73, last 10 4-6.

Risk factors. This is the pick I would have made on season aggregates alone, and it would have been wrong to. Flagged here as the clearest example of why the recency rule exists.

Source. MLB Stats API game log, Zebby Matthews 2026 — start-by-start log including the 8/30 outing vs Chicago.

Candidate C — PIT ML, 1.52 — killed by the starter's collapse

The math. Implied % at 1.52 = 66.0%. Blending season log5 with home/away splits puts Pittsburgh around 66% before pitching. Jared Jones' recent form is worth roughly −4% → 62.0%. Edge = −4.0%. EV = (0.62 × $2.46) − (0.38 × $4.78) = $1.53 − $1.82 = −$0.29. Not a play.

The evidence. The setup was tempting: the Angels are 53-87, 1-9 in their last 10, and 23-43 on the road, which is about as bad as a road team gets. But Pittsburgh is only 69-72 (last 10 6-4) and is asking Jared Jones, who over his last 5 starts is 0-5 with a 7.40 ERA (20.2 IP, 17 ER) against a 4.83 season figure. That is a 2.57 divergence and the largest negative flag on the slate. Ryan Johnson (5.47) is worse on the season, but not by enough to survive Jones pitching this badly at a price demanding 66%.

Risk factors. A 1-9 opponent is exactly the spot where a stale price looks like value; the market has not mispriced Los Angeles here, it has correctly priced Pittsburgh's rotation problem.

Source. MLB Stats API game log, Jared Jones 2026 — the 0-5 / 7.40 stretch from 8/4 through 8/28.

Candidate D — TB ML @ TEX — passed on the line-drift rule, not on the numbers

The math. Not computed to a stake, because the rule fires before pricing. Tampa Bay is 83-57 and starting Nick Martinez (13-4, 2.99); Texas is 70-71 starting Trevor Williams (3.93 ERA, 23 K — a limited workload suggesting a recent return). My model has Tampa Bay as a clear favorite. The market has Texas favored at −110 to −115 across all four books.

The evidence. That is a gap of well over 5 cents between my model and the market, in the direction that would make Tampa Bay look like free money. The Strategy tab's line-drift rule is explicit: when the price is more than 5 cents better than the model in our favor, treat it as a sharp-money signal on the other side and reduce or pass. The most likely explanations — a Martinez workload limit, a lineup absence, or something in the Williams return I could not confirm — all argue for standing down. Team form: TB last 10 6-4, TEX last 10 5-5, Texas 39-30 at home.

Risk factors. This may well be the one that wins tonight. Passing a play the model likes because the market disagrees is the rule working as designed, and it is worth logging either way.

Source. RotoWire MLB Daily Lineups — four-book line showing Texas favored, plus both confirmed starters.

Step 3.5 Self-Check

• Odds-band floor: Pass. ATH @ SEA (1.44) and WSH @ LAD (1.35) explicitly excluded as sub-1.50. No recommended pick exists to violate the floor.
• Coors Field: Pass. STL @ COL marked no play across the board.
• Yankees tilt-fade: Pass. Tilt-fade is listed as currently active on the Strategy tab; NYY ML excluded at Petco and not recommended anywhere.
• 2-source starter confirmation: Pass. All 16 games' starters confirmed on MLB.com and RotoWire with matching season lines.
• Timezone check: Pass. All times ET from MLB.com (publishes ET natively); RotoWire agrees. The only game past its start time (2:10 PM G1) is excluded.
• Recency check: Pass. Last-5 logs pulled for Griffin, Sears, Fedde, Matthews, Drohan, Lowder and Jones; last-10 and home/away for all teams. Two meaningful divergences flagged (Matthews +1.68 favorable, Jones −2.57 adverse). Sears flagged as having no 5-start history.
• No padding check: Pass — zero rows, zero padding.
• Kelly sizing check: N/A. No qualifying plays, so no budget allocated; noted explicitly in Section 3.
• Bankroll provenance: Pass. $23.91 read from the Log sheet column H, row 99 (Bet ID 98, dated 2026-09-03) this run — not from a prior report header.
• Section 4 completeness: Pass. Each candidate shows implied %, true estimate, edge, worked EV, and a named source appearing in the Sources list.

Pending Bets

None. The Log shows no Open or unsettled rows; the last entry (Bet ID 98, 9/3, Main/SGP, $5.98) is settled as a Loss. Nothing to settle.

Sources

• MLB.com Official Schedule — September 4, 2026
• MLB.com Probable Pitchers — September 4, 2026
• RotoWire MLB Daily Lineups
• MLB Stats API — 2026 standings with last-10 and home/away splits
• MLB Stats API — schedule/status for 2026-09-04
• MLB Stats API — game log, Foster Griffin
• MLB Stats API — game log, Andrew Sears
• MLB Stats API — game log, Erick Fedde
• MLB Stats API — game log, Zebby Matthews
• MLB Stats API — game log, Shane Drohan
• MLB Stats API — game log, Rhett Lowder
• MLB Stats API — game log, Jared Jones

Opened but rejected: covers.com/sports/mlb/matchups returned stale content listing pitchers not on today's slate (Gausman, Skubal, Alcantara, McClanahan). Not used for any figure in this report.