Daily Bet Research Report — Thursday, September 3, 2026

Report generated at 2:40 PM ET on Thursday, September 3, 2026

Bankroll: $29.89 | Stake (20%): $5.98 Record: 46-60-3 | Hit rate: 43.4%

Record is all 109 logged bets (hit rate excludes pushes). Main-category only: 43-38-3, 53.1%. Bankroll pulled from Log sheet row 110, column H this run.

⚠️ Read this before the picks

Two things in the tracker data are more decision-relevant than anything on today's slate:

1. The bankroll is down 90.3% from its peak. The Log shows an all-time high of $309.11 and a current balance of $29.89. From $277.22 on Aug 10 to $29.89 on Sep 1 is 22 days.

2. The staking rule is the mechanism, not the bad luck. Main-bet hit rate is 53.1% (43-38-3) — roughly what you'd expect from picks with little or no true edge. Staking 20% of bankroll per day against a ~0% edge produces exactly this curve. Today's only qualifying pick has a full-Kelly fraction of 2.8% of bankroll; the house rule stakes 20%, about 7x Kelly. The Strategy tab's own "Drawdown awareness" rule says to favor higher-hit-rate plays when in a drawdown zone, and lists "rest day" as always-valid.

I've sized today's play per the 20% rule as instructed. Flagging that the rule and the results are related is not editorializing — it's the clearest read of the tracker, and this is a decision for you to make in the main chat, not for me to make by quietly adjusting a number.

Also, an honesty note on the edge figures below: I have real bet365 moneyline prices for five of six playable games (bet365 lines via Athlon previews, timestamped ~8:15 AM ET). The "Hit Rate" column is my own model estimate. Where my estimate diverges sharply from the market, I have shrunk it toward the market rather than claiming free money — see the line-drift note in Section 4.

Section 2: Schedule & Summary Analysis

Nine games on the slate. Three had already started before report time (2:40 PM ET) and are excluded per the 30-minute rule: SF @ PIT 12:35 PM, TOR @ CLE 1:10 PM, CWS @ HOU 2:10 PM. Six playable games below.

All times ET, taken directly from MLB.com's schedule page and cross-confirmed against MLB StatsAPI (UTC timestamps converted). No source time needed conversion from PT.

League	Game	Time (ET)	Bet Type	Pick	Est. Odds	EV	EV ★	Hit %	Hit ★
MLB	BOS @ BAL	7:15 PM	no play	— (BAL considered)	2.10	−$0.12	★	46.7%	★★
MLB	MIL @ CHC	7:15 PM	ML	Brewers ML	1.88	+$0.15	★★	54.5%	★★★
MLB	MIA @ KC	7:40 PM	no play	—	~1.83 (unconfirmed)	n/a	★	n/a	★
MLB	TB @ TEX	8:05 PM	no play	— (TB considered)	1.78	−$0.18	★	54.5%	★★★
MLB	ATH @ SEA	9:40 PM	no play	—	1.47 (SEA)	—	★	—	★
MLB	STL @ LAD	10:10 PM	no play	—	1.31 (LAD)	—	★	—	★

Why each no-play:

BOS @ BAL — Bennett (3.27) vs Young (3.38) is a coin-flip pitching matchup. Two estimation methods straddle the market (overall-record log5 gives BOS 50.3%, home/road-splits log5 gives BOS 56.3%; market no-vig is 54.7%). My uncertainty is wider than any edge I could claim. Pass.
MIA @ KC — The Athlon preview page for this game returned empty on two fetch attempts. I have only search-summary odds (MIA −116 to −128), which I did not open and will not treat as a verified price. Separately, Miami's road record (29-42, .408) and Kansas City's home record (36-32, .529) point the opposite direction from the market's Marlins favoritism, and I can't resolve that gap without the actual page. No play on unverified inputs.
TB @ TEX — Rays at 1.78 implies 56.2%; my estimate is 54.5%, so the offered price is worse than fair. Quantrill is also in a genuine hot stretch (2-1, 24 IP, 4 ER = 1.50 ERA over his last four, vs 3.01 season — a meaningful positive divergence) which cuts against the Rays. Negative edge both sides.
ATH @ SEA — Seattle at 1.47 decimal is below the 1.50 odds-band floor in the Strategy tab. Excluded by rule. The A's side at 2.71 has no case (54-86). Also: Kade Anderson has essentially no MLB track record (0-0, 4.50), so no confident projection exists either way.
STL @ LAD — Dodgers at 1.31 decimal is well below the 1.50 floor. Excluded by rule despite this being the day's cleanest template match (home favorite + large pitching edge + team quality). The price is the problem, not the matchup.

Section 3: My Recommendations

One qualifying play today. Five of six playable games failed on either the odds-band floor, negative edge, or unverifiable inputs. This is a thin slate that produced one marginal pick — not five.

#	Option	Lens	Est. Odds	Implied %	Hit Rate	Edge %	EV	Kelly Weight	Stake %	Stake $
1	Milwaukee Brewers ML @ CHC	Balanced	1.88	53.2%	54.5%	+1.3%	+$0.15	100%	20%	$5.98

Kelly Weight sums to 100% and Stake % sums to 20% trivially, as there is one qualifying pick. Stake $ = $5.98 = bankroll × 20%.

Sizing caveat, stated plainly: this pick's own full-Kelly fraction is 2.8% of bankroll ($0.84). The 20% budget is your fixed rule, not Kelly's answer. As the workflow doc notes, normalizing to a fixed 20% preserves relative weighting but discards Kelly's view on total exposure — with one pick there is no relative weighting left to preserve, so today the 20% is purely the fixed-fraction rule doing the work.

Rest day is a legitimate alternative today. A +1.3% edge is thin, and the Strategy tab explicitly lists rest day as valid when the slate is thin or after a losing stretch. Both conditions hold.

Section 4: Reasoning

Option 1 — Milwaukee Brewers ML at Chicago Cubs (1.88 / −114), 7:15 PM ET, Wrigley Field

1. The math.

Implied % from price: 1 / 1.88 = 53.2%
Market no-vig (MIL −114 / CHC −108 → 53.28% + 51.92% = 105.2%): 50.6%
My true probability estimate: 54.5%
Edge % = 54.5% − 53.2% = +1.3%

EV = (True% × profit if wins) − ((1 − True%) × stake) Profit if wins = $5.98 × 0.88 = $5.26 EV = (0.545 × $5.26) − (0.455 × $5.98) EV = $2.867 − $2.721 = +$0.15

Kelly: f = (0.545 × 1.88 − 1) / (1.88 − 1) = (1.0246 − 1) / 0.88 = 2.80% of bankroll.

How I got 54.5%, including the part where I talked myself down: Log5 on team quality puts Milwaukee at 52.6% (both via overall records with a 4% home adjustment for Chicago, and via home/road splits — Brewers road .588, Cubs home .563 — the two methods agree). Adding the starter ERA gap at a conservative 3-6% lifts that to 55.6-58.6%. But the no-vig market has this at 50.6%, meaning my model wants MIL around −145 while the book offers −114. That is roughly 31 cents better than model in our favor, which triggers the Strategy tab's line drift rule: a price that good against your model is usually the market knowing something you don't, not free money. So I shrank hard toward the market and landed at 54.5%. The remaining +1.3% is what survives that haircut. It is real but small, and I would not defend 58%.

2. The evidence.

Logan Henderson (MIL), last 5 starts — required recency read:

Date	Opp	Line
Aug 4	vs PIT	6.0 IP, 2 ER — W
Aug 10	@ SD	6.1 IP, 3 ER — L
Aug 16	@ LAD	7.0 IP, 1 ER — W
Aug 22	vs ATL	6.0 IP, 1 ER — W
Aug 28	vs TEX	7.0 IP, 1 ER — W

Last 5: 4-1, 2.23 ERA over 32.1 IP. Season: 9-2, 2.48 ERA, 0.83 WHIP. Divergence is −0.25, well inside the 1.5 threshold — no meaningful divergence to flag. Recent form confirms rather than contradicts the season line, which is the case I want before leaning on a season ERA.

Kevin Gausman (CHC), last 5 starts:

Date	Opp	Line
Aug 7	@ KC	7.0 IP, 2 ER — W
Aug 13	@ WSH	4.2 IP, 6 ER — L
Aug 18	vs CWS	4.1 IP, 3 ER — ND
Aug 24	@ ARI	7.0 IP, 0 ER — W
Aug 29	vs CIN	5.0 IP, 5 ER — W

Last 5: 3-1, 5.14 ERA over 28.0 IP. Season: 8-11, 4.52 ERA, 1.27 WHIP. Divergence is +0.62 — below the 1.5 flag threshold, but directionally it makes him worse than his season line, not better. The 3-1 record over that span is run support masking the performance: the Cubs scored enough to win games in which he allowed 5 and 6 earned runs. Athlon lists his 2026 run support at 6.8/game. Note also the volatility — 0 ER and 6 ER starts inside the same five-game window.

Starter ERA gap (season): 2.48 vs 4.52 = 2.04 runs. Gap on last-5: 2.23 vs 5.14 = 2.91 runs. That is the core of the pick.

Team quality — required last-10 for both sides: Brewers 6-4 in last 10 (87-53 overall, best record in baseball; 40-28 on the road). Cubs 4-6 in last 10 (78-62 overall; 40-31 at home). Both last-10 figures confirmed independently by MLB StatsAPI standings splits and the Athlon game-by-game list. Milwaukee has taken the first three of this four-game series, winning 9-4 and 9-5 on Sept 1-2 after being blown out 17-3 on Aug 31.

3. Risk factors.

The market disagrees with me. No-vig 50.6% vs my 54.5%. The book has both ERAs too and priced this near pick'em anyway. That is the single biggest risk and the reason this is a 1.3% edge rather than a 6% one.
Wrigley + Cubs offense variance. Chicago put up 17 runs on Aug 31 and 19 on Aug 23. This lineup can erase a pitching edge in one inning. Conforto (.438 over last 7), Kelly (5-13, 7 RBI last 3) and Crow-Armstrong (.348, 1.000 SLG last 5) are all hot per Athlon.
Henderson's workload. He averages under 6 IP per start, so Milwaukee's bullpen inherits roughly a third of this game. The pick is not purely "Henderson beats Gausman."
Template mismatch. The Strategy tab's preferred template is home favorite + pitching edge + team quality. Milwaukee is the road side here. Two of three criteria, not three.
Series fatigue / letdown spot. Fourth game of a four-game set that Milwaukee has already clinched, first week of September, with a large division lead. Lineup rest is plausible; I have not seen a confirmed lineup card.

4. Named source.

Milwaukee Brewers at Chicago Cubs MLB Prediction, Picks & Odds for Sept. 3, 2026 — Athlon Sports, published 12:41 PM ET Sept 3, 2026, credited to Cory Bonini. This is the page carrying the bet365 moneyline (MIL −114 / CHC −108, timestamped 8:15 AM ET), the O/U 8.5, both teams' last-10 game logs, and the Gausman run-support figure. Caveat I'm obliged to state: this page carries a disclosure that some or all of its content may be AI-generated. I therefore cross-checked its pitcher lines against MLB StatsAPI game logs and Baseball-Reference and they matched exactly (Henderson 9-2/2.48, last start Aug 28 vs TEX 7 IP/1 ER/1 H; Gausman 8-11/4.52). The odds themselves I could not independently verify against a second book — treat the 1.88 as indicative and re-check your actual price before booking.

Starter confirmation for this pick came from three sources I opened and read: MLB.com schedule, Baseball-Reference probable pitchers, and MLB StatsAPI. That satisfies the 2+ source requirement.

Sources

Pages and endpoints I actually opened and read for this report:

Not used as a source: the Athlon Marlins-Royals preview returned empty content on two attempts. Odds I saw for that game came only from search summaries, were not verified on a page I opened, and are marked unconfirmed in Section 2 accordingly.

Pending bets

None. The Log shows no open bets — every row through Bet ID 97 (Sept 1) is settled. Nothing for you to settle in the main chat from prior days.

Self-check (Step 3.5)

Check	Result
Odds-band floor (≥1.50 decimal)	Pass — sole pick at 1.88. SEA (1.47) and LAD (1.31) explicitly excluded on this rule.
Coors Field	N/A — Colorado is not on today's slate (Rockies hosted Baltimore through Sept 2; no COL game Sept 3).
Yankees tilt-fade	N/A — no Yankees game on today's slate.
2-source starter confirmation	Pass — Henderson and Gausman confirmed via MLB.com, Baseball-Reference, and MLB StatsAPI (3 sources, all opened).
Timezone	Pass — all times ET from MLB.com, cross-checked against StatsAPI UTC. Three games already started were excluded, not recommended.
Recency	Pass — last-5 starts given for both Henderson and Gausman; last-10 given for both Brewers and Cubs. Quantrill's divergence flagged in the TB@TEX no-play note.
No padding	Pass — one row, because one game cleared. Five no-plays documented with reasons.
Kelly sizing	Pass — weight 100%, stake 20%, $5.98 = bankroll × 20%.
Bankroll provenance	Pass — read from Log sheet row 110 col H this run ($29.89), not from a prior report header.
Section 4 completeness	Pass — math derived, evidence with recency, risks, named source appearing in Sources list.

← Back to all reports

Thursday, September 03, 2026