Sunday, September 20, 2026

Daily bet research report

Daily Bet Research Report — Sunday, September 20, 2026

Start: 11:10 ET | End: 11:16 ET

Bankroll: $60.23 | Stake (12.5%, one play): $7.53 Record: 3-1-0 (Main bets, main leg, since 9/14 reset) | Hit rate: 75.0% Slate: 29 games (15 MLB + 14 NFL) · 16 analysed · 22 sources · Playbooks: MLB v1, NFL v1 · Verified at 11:10 ET

Section 1: Verification

Verified against statsapi.mlb.com/api/v1/schedule (MLB) and sports.core.api.espn.com/.../events?dates=20260920 (NFL) at 11:10 ET — 15 MLB games, 14 NFL games.

Every probable pitcher cited below was read from the hydrate=probablePitcher response for 2026-09-20 specifically, and every price was pulled from the DraftKings feed keyed to that same event ID. No board from a prior day was used.

Section 2: Game Analysis

Sport Game Time (ET) DK Line Note Play?
MLB PHI @ NYM 1:10 PHI -174 / O-U 7.5 Mets start rookie Jonah Tong, who has no Savant xERA row — the deciding number is unavailable. No — no xERA
MLB KC @ PIT 1:35 KC -120 / O-U 8.5 Wacha ERA 3.34 vs xERA 3.90 (−0.561) is lucky, but the gap sits under the playbook's 1.00 flag. No — under threshold
MLB CHC @ CIN 1:40 — Cubs list no probable pitcher on statsapi; Hard Rule H3 fails outright. No — H3 violation
MLB ATH @ CLE 1:40 CLE -247 / O-U 7.0 Jack Perkins ERA 6.14 vs xERA 4.07 — a +2.066 gap, the largest on the board. #4
MLB BOS @ TB 1:40 TB -155 / O-U 8.0 Total already cut 8.5 → 8.0; Jax (−0.594) is a converted reliever, so this is a bullpen game I can't price. No — pen unknown
MLB ATL @ HOU 2:10 HOU -142 / O-U 8.0 Martín Pérez ERA 3.07 vs xERA 4.58 — a −1.507 gap, the most overperforming starter today. #1
MLB DET @ CHW 2:10 CHW -120 / O-U 8.0 Melton's 2.52 ERA sits 1.219 below his 3.74 xERA and the line hasn't moved a cent all morning. #2
MLB WSH @ STL 2:15 STL -148 / O-U 9.0 Quinn Mathews has no Savant xERA row, so half the matchup is unreadable. No — no xERA
MLB TOR @ TEX 2:35 TEX (deGrom) / O-U n/a Toronto lists no probable pitcher; H3 fails despite deGrom's +0.440 buy signal. No — H3 violation
MLB MIN @ LAA 4:07 MIN -121 / O-U 8.5 Kremer ERA 5.35 vs xERA 4.55 (+0.803) against Johnson at −0.178 — a ~0.98-run regression split. #3
MLB SEA @ COL 3:10 — Coors Field. Hard Rule H1 — automatic no-play. No — H1 Coors
MLB SF @ LAD 4:10 LAD -288 / O-U 8.0 Neither club lists a probable pitcher; H3 fails, and LAD is out of band regardless. No — H3 violation
NFL CAR @ ATL 1:00 ATL -1.5 / O-U 44.5 Indoor, so no wind read needed — but no dated QB designation was obtainable. No — H1 QB
NFL MIN @ CHI 1:00 CHI -5.5 / O-U 48.5 Spread moved 3.5 → 5.5 and the total jumped 45.5 → 48.5, a move that size implies QB news I can't confirm. No — H1 QB
NFL PHI @ TEN 1:00 PHI -7 / O-U 39.5 Spread ran 4.5 → 7 while the total fell 42.5 → 39.5 — classic starting-QB-out signature, unverified. No — H1 QB
NFL PIT @ NE 1:00 NE -5.5 / O-U 41.5 Line crossed the key number 5 (4.5 → 5.5) and the total dropped 2 points; no QB status available. No — H1 QB

Unanalysed: 13 games. Nine NFL games beyond the four above were left alone once the ESPN injuries feed failed — with NFL Hard Rule H1 unsatisfiable, further NFL work could not produce a play, so the time went to MLB instead. Three MLB games (MIA @ SD, NYY @ ARI, MIL @ BAL) were not priced; on the first two the xERA gaps were small or missing (Burnes has no Savant row), and Milwaukee's price behind Misiorowski was certain to sit outside the 1.50–2.10 band. One MLB game (SEA @ COL) was ruled out by H1 before any work.

Section 3: Today's Plays

# Play Sport Market Odds Confidence Calibrated Edge %
1 Houston Astros ML (vs ATL) MLB Moneyline 1.70 60% 57% −1.8%
2 Chicago White Sox ML (vs DET) MLB Moneyline 1.83 57% 53% −1.6%
3 Minnesota Twins ML (@ LAA) MLB Moneyline 1.82 57% 53% −2.0%
4 Athletics +1.5 (@ CLE) MLB Spread 1.89 55% 51% −1.9%

⚠ All four plays carry negative Edge. That is the honest headline of this card, not a rounding artefact. DraftKings' hold on these markets is roughly 4%, and the largest edge I can defend today is under 2% in the wrong direction — meaning the xERA regression gaps I found are real but not large enough to clear the juice. Ranked by Calibrated Confidence as required; #1 is the least bad rather than a bargain. A genuine no-play day was considered and is defensible here.

Section 4: Reasoning

#1 — Houston Astros ML @ 1.70 (Calibrated 57%, Edge −1.8%)

This rests on MLB playbook §2, the ERA − xERA gap, tagged [MODERATE]. Martín Pérez carries a 3.07 ERA against a 4.58 xERA — era_minus_xera_diff of −1.507, the most negative figure among all 26 starters I pulled today. Under the §2 sign convention a negative gap means he has been lucky and regression is coming, and 1.507 clears the playbook's 1.00 flag threshold (though not the 1.75 headline bar). Hunter Brown is the mirror image: 3.43 ERA against a 3.30 xERA, a +0.133 gap, meaning his results are real. The contact-quality spread between the two starters is about 1.28 runs of xERA in Houston's favour, while the market prices Atlanta's starter off a 3.07 ERA and Atlanta's 91-64 record.

Line movement cuts against me. Houston opened -136 and is now -142 — a 6-cent move toward my side, which per the pricing rules means the market already agrees and I am late to it. The risk that beats me is the bullpen: Houston used four relievers in last night's 6-3 loss (verified from the statsapi boxscore for gamePk 824137) against Atlanta's three, and this is the third consecutive day of the series for both pens — §3's documented decline breakpoint. That is a real disadvantage pointing the other way, and §2 explicitly forbids carrying a play on the xERA gap alone. Here it is paired only with a weak bullpen read that actually opposes it, which is why confidence sits at 60 rather than higher. Daikin Park is domed, so no weather input applies.

#2 — Chicago White Sox ML @ 1.83 (Calibrated 53%, Edge −1.6%)

Same §2 [MODERATE] mechanism, cleaner market context. Troy Melton's 2.52 ERA sits 1.219 runs below his 3.74 xERA — the second-most-overperforming starter on today's board — and Detroit is being priced off that shiny ERA. Davis Martin is also lucky (3.96 ERA, 4.64 xERA, −0.683) but by roughly half as much, leaving a net regression differential of about 0.54 runs in Chicago's favour.

The line has not moved one cent — Chicago opened -120 and remains -120, with the total pinned at 8.0. Under the movement rules that is the most encouraging pattern available: no sharp money has taken the other side, and nothing suggests I am chasing a move. Chicago won 3-1 last night and the clubs are near-identical on paper (79-76 vs 73-82), so the price is close to fair before any regression adjustment. The risk that beats me is that this argument is a difference between two soft numbers — both starters are lucky, so I am betting on the relative size of two regressions, which is a thinner claim than a single clear mispricing. I could not obtain wind for Rate Field (see limitations), though that matters for a total rather than this moneyline.

#3 — Minnesota Twins ML @ 1.82 (Calibrated 53%, Edge −2.0%)

The §2 [MODERATE] signal running the opposite direction — a buy rather than a fade. Dean Kremer's 5.35 ERA sits 0.803 runs above his 4.55 xERA, the positive-gap direction that per §2 means he has been unlucky and should regress better, while the public fades him on results. Ryan Johnson is mildly the other way at −0.178 (4.88 ERA, 5.06 xERA). The combined differential is about 0.98 runs in Minnesota's favour — just under the 1.00 flag threshold, which is why this ranks third rather than second despite the tidy direction.

Movement is negligible: Minnesota opened -122 and sits -121, with the total unmoved at 8.5. The risk that beats me is straightforward — Los Angeles won last night 6-5 and holds home field, and a 0.98-run xERA differential is precisely the magnitude §2 warns is at the edge of what this estimator can support. The playbook is explicit that xERA sits only roughly level with FIP for forward prediction, so a sub-1.00 gap is close to noise. Both bullpens worked a one-run game last night; I did not fetch per-reliever pitch counts for this matchup.

#4 — Athletics +1.5 @ 1.89 (Calibrated 51%, Edge −1.9%)

The most interesting number on the slate and the least trustworthy play. Jack Perkins carries a 6.14 ERA against a 4.07 xERA — era_minus_xera_diff of +2.066, past §2's 1.75 "headline-worthy" bar and the largest gap I found today in either direction. Cleveland is -247 in part because the market sees a 6.14 ERA; Statcast contact quality says he has pitched two full runs better than that. The structure fits §7's run-line guidance: taking +1.5 with a dog requires the game to stay close, which favours good pitching on both sides and a low total — and Gavin Williams is genuinely good (3.78 ERA, 3.77 xERA, essentially zero gap) while the total of 7.0 is the lowest on the entire slate. Both of §7's stated conditions are met.

Movement is the problem, and it is emphatic. Oakland drifted from +192 to +200 on the moneyline — 8 cents against, past the 5-cent threshold where the rules say sharp money disagrees and confidence must come down — and the +1.5 price drifted with it, from -118 to -112. The market is walking away from the Athletics in both markets simultaneously. The risk that beats me is simply that Oakland is 61-94, lost 12-6 here last night, and a 2.07-run xERA gap does not make Perkins good, only less bad than his ERA; §2 forbids the gap carrying a play alone, and while §7's structural conditions provide the second leg, that leg is [MODERATE] guidance rather than quantified evidence. This is ranked last deliberately.

Sources

Fetched this run:

Attempted and failed (detailed below): https://api.weather.gov/points/41.4962,-81.6852 · https://www.espn.com/nfl/injuries

Research limitations

The three counts: 16 games analysed of 29 bookable · 22 sources fetched successfully (26 attempted, 4 failed) · 7 games where a Tier-1 factor could not be obtained.

Source failures.

Tier-1 factors unavailable (7 games). Four MLB games lack a probable pitcher entirely on statsapi (CHC @ CIN, TOR @ TEX, SF @ LAD — the Cubs, Blue Jays and both San Francisco/LA starters), failing H3. Three more have a starter with no Baseball Savant xERA row: Jonah Tong (NYM), Quinn Mathews (STL) and Corbin Burnes (ARI). Tong and Mathews are rookies below the min=q qualification cut, which is the expected coverage gap; Burnes missing is unexpected for a pitcher of his profile and may indicate an injury-shortened season — I did not have a way to confirm that, and per §9 I did not substitute FIP.

Bullpen data is partial. I pulled pitcher-usage lists from last night's boxscores for ATH @ CLE and ATL @ HOU only, and those give me counts of relievers used, not innings or pitch counts. §3's checklist asks for pitch counts on the top 2-3 arms and days-since-off-day; I have neither. For the DET @ CHW and MIN @ LAA plays I have no bullpen read at all. This is the largest gap in today's work and it sits on the playbook's single most under-priced factor.

Stub playbooks: none used. NHL and NBA are out of season; no stub was relied on.

Calibration caveat, restated. The 3-point/4-point Calibrated deduction was measured on a mostly-prop sample and there is still very little history on game lines. Today's four plays all land within 2 points of break-even, so the ranking is sensitive to a calibration adjustment that is itself unvalidated. Treat the ordering as weakly held.

Evidence-ledger note. All four plays cite the ERA − xERA gap. Per §10 the 1.00 / 1.75 thresholds were invented, not derived, and the ledger asks whether plays citing the gap beat plays that don't — today adds four data points toward that count, which is worth tracking.

Cutoff: not hit. Research ran 11:10 to 11:16 ET, well inside the 11:50 deadline. No game inside 30 minutes of first pitch was considered; the earliest analysed game starts at 1:00 PM ET.

← Back to all reports
Betting Tracker · Generated 2026-09-21 11:19 · Personal use only