- WINBryce Eldridge · SF vs NYM · Hits Over 0.5 · final 1
- VOIDJazz Chisholm Jr. · NYY vs SD · Hits + Runs + RBIs Over 0.5
Results & Receipts
Track the official paper-card record, open exposure, settlement status, and model-performance receipts in one place.
Official product-card record. Raw model-performance results are tracked separately, further down this page.
Settlement status
Live todayCards settle on official finals only. When a slate produces no qualified card, it is a no-play day — no exposure is created and nothing is graded.
Bank Builder — settled cards
Open Bank Builder →Lane A advanced — awaiting next qualified card. An awaiting lane is not a pending settlement and is not counted in the record.
MLB model performance
Latest grading · yesterdayRaw model-performance ledger — every published MLB projection graded against official box scores. This is not part of the 19-14 product-card record above, and the two are never combined.
How settlement works
- Product cards settle on official finals only — never on in-progress or web-snippet scores.
- A pending card is not a loss. It stays pending until the official result is in.
- Pushes and voids are handled separately and excluded from the win/loss denominator.
- Raw MLB model-performance grading is a separate ledger and never changes the product-card record.
- No paper exposure is created without an approved card — no-play days create nothing.
Explore the record
Dates are the day the card was published and graded as a cohort — not the day an individual leg settled. A card whose legs cross midnight stays in one day. Settled cards run 2026-08-17 to 2026-09-05.
Whole published slips, graded once each under the policy frozen at publication. A slip wins only if every leg does — one winning leg is not a winning parlay.
| Sport | Record | Decisive | Pending | Period |
|---|---|---|---|---|
| Premier League | 20.0% 1-4 | 5 | — | all settled history |
| MLB | 17.9% 7-32 | 39 | 5 | all settled history |
| Mixed sport | no settled cards yet | 0 | — | all settled history |
| NFL | no settled cards yet | 0 | — | all settled history |
| UFC | 0.0% 0-2 | 2 | 4 | all settled history |
By sport and risk tier
| Sport | low | medium | high | longshot | All tiers |
|---|---|---|---|---|---|
| Premier League | — | — | 1-4 | ||
| MLB | 7-32 | ||||
| Mixed sport | — | — | — | — | — |
| NFL | — | — | — | — | — |
| UFC | 0-2 |
A trailing +n counts cards that have not settled; they are in no rate above. A dash is an empty cell — no card of that tier settled in this period, which is not a loss. Mixed-sport cards are their own row and are never counted inside a single sport.
Day by day
all settled history · all sports · 18 of 20 days had a settled card. Column height is the number of DECIDED cards that day, so a one-card day cannot look like a trend. A day with nothing on it has no rate and is drawn as a gap, never as a zero.
The numbers behind these charts
| Date | Day | Decided | Pending | Day rate | Cumulative |
|---|---|---|---|---|---|
| 2026-08-17 | 0-3 | 3 | — | 0.0% | 0.0% · 0-3 |
| 2026-08-18 | 0-1 | 1 | 1 | 0.0% | 0.0% · 0-4 |
| 2026-08-19 | 1-3 | 4 | — | 25.0% | 12.5% · 1-7 |
| 2026-08-20 | no card | 0 | — | — | 12.5% · 1-7 |
| 2026-08-21 | no card | 0 | — | — | 12.5% · 1-7 |
| 2026-08-22 | 0-4 | 4 | 2 | 0.0% | 8.3% · 1-11 |
| 2026-08-23 | 0-4 | 4 | — | 0.0% | 6.3% · 1-15 |
| 2026-08-24 | 0-2 | 2 | — | 0.0% | 5.6% · 1-17 |
| 2026-08-25 | 1-1 | 2 | — | 50.0% | 10.0% · 2-18 |
| 2026-08-26 | 0-2 | 2 | — | 0.0% | 9.1% · 2-20 |
| 2026-08-27 | 1-1 | 2 | 1 | 50.0% | 12.5% · 3-21 |
| 2026-08-28 | 0-2 | 2 | — | 0.0% | 11.5% · 3-23 |
| 2026-08-29 | 1-2 | 3 | — | 33.3% | 13.8% · 4-25 |
| 2026-08-30 | 1-1 | 2 | 1 | 50.0% | 16.1% · 5-26 |
| 2026-08-31 | 0-2 | 2 | — | 0.0% | 15.2% · 5-28 |
| 2026-09-01 | 0-2 | 2 | — | 0.0% | 14.3% · 5-30 |
| 2026-09-02 | 0-3 | 3 | — | 0.0% | 13.2% · 5-33 |
| 2026-09-03 | 1-1 | 2 | — | 50.0% | 15.0% · 6-34 |
| 2026-09-04 | 1-1 | 2 | — | 50.0% | 16.7% · 7-35 |
| 2026-09-05 | 1-3 | 4 | 4 | 25.0% | 17.4% · 8-38 |
A rising line is not evidence the model learned. These are whole published cards over a short period, their legs are not independent of each other, and no period here is long enough to support a claim about the future.
The 55 cards behind this record
| Date | Sport | Tier | Legs | Price | Result | Slip |
|---|---|---|---|---|---|---|
| 2026-09-05 | mlb | medium | 2 | 2.86 | win | slip_2026-09-05_conservative_ca54196217e8 |
| 2026-09-05 | mlb | high | 3 | 6.98 | pending | slip_2026-09-05_balanced_e0d9e48237f7 |
| 2026-09-05 | mlb | longshot | 3 | 7.64 | loss | slip_2026-09-05_balanced_5b027bb92333 |
| 2026-09-05 | ufc | low | 2 | 1.90 | pending | ufc-low-2026-09-05 |
| 2026-09-05 | ufc | medium | 2 | 2.10 | pending | ufc-medium-2026-09-05 |
| 2026-09-05 | ufc | high | 4 | 4.84 | pending | ufc-high-2026-09-05 |
| 2026-09-05 | epl | medium | 2 | 3.12 | loss | epl-medium-2026-09-05 |
| 2026-09-05 | epl | high | 3 | 6.33 | loss | epl-high-2026-09-05 |
| 2026-09-04 | mlb | medium | 2 | 2.92 | win | slip_2026-09-04_conservative_7c390b35b2d0 |
| 2026-09-04 | mlb | high | 3 | 4.54 | loss | slip_2026-09-04_balanced_2bee02c34d6f |
| 2026-09-03 | mlb | medium | 2 | 3.71 | win | slip_2026-09-03_conservative_f900ffd5dd71 |
| 2026-09-03 | mlb | longshot | 5 | 30.13 | loss | slip_2026-09-03_aggressive_bb7c938c16cb |
| 2026-09-02 | mlb | medium | 2 | 2.76 | loss | slip_2026-09-02_conservative_f87cbfaf4ea5 |
| 2026-09-02 | mlb | high | 3 | 6.69 | loss | slip_2026-09-02_balanced_dfb953195e67 |
| 2026-09-02 | mlb | longshot | 5 | 34.12 | loss | slip_2026-09-02_aggressive_b755d45b5390 |
| 2026-09-01 | mlb | medium | 2 | 2.72 | loss | slip_2026-09-01_conservative_09997f5f3542 |
| 2026-09-01 | mlb | longshot | 5 | 39.16 | loss | slip_2026-09-01_aggressive_36b88c13c50b |
| 2026-08-31 | mlb | medium | 2 | 2.69 | loss | slip_2026-08-31_conservative_c279195bb053 |
| 2026-08-31 | mlb | longshot | 5 | 35.93 | loss | slip_2026-08-31_aggressive_e2a526bf6e8b |
| 2026-08-30 | mlb | medium | 2 | 2.27 | pending | slip_2026-08-30_conservative_1984dceacedd |
| 2026-08-30 | mlb | longshot | 5 | 45.48 | loss | slip_2026-08-30_aggressive_ad582ff3988d |
| 2026-08-30 | epl | medium | 2 | 2.53 | win | epl-medium-2026-08-30 |
| 2026-08-29 | mlb | medium | 2 | 3.18 | win | slip_2026-08-29_conservative_eccb703ffb25 |
| 2026-08-29 | mlb | longshot | 5 | 27.29 | loss | slip_2026-08-29_aggressive_732b9f46d3f3 |
| 2026-08-29 | epl | medium | 2 | 2.44 | loss | epl-medium-2026-08-29 |
| 2026-08-28 | mlb | medium | 2 | 3.04 | loss | slip_2026-08-28_conservative_ee70eec67f76 |
| 2026-08-28 | mlb | longshot | 5 | 30.71 | loss | slip_2026-08-28_aggressive_910e0e45535a |
| 2026-08-27 | mlb | medium | 2 | 2.65 | loss | opt_2026-08-27_public_low_mlb_52b1319da1f3 |
| 2026-08-27 | mlb | high | 3 | 4.22 | pending | opt_2026-08-27_public_medium_mlb_e97372044e32 |
| 2026-08-27 | mlb | longshot | 5 | 13.83 | win | opt_2026-08-27_public_longshot_mlb_42ab1283f8ba |
| 2026-08-26 | mlb | medium | 2 | 3.10 | loss | slip_2026-08-26_conservative_26a76e6a2db2 |
| 2026-08-26 | mlb | high | 2 | 4.91 | loss | slip_2026-08-26_conservative_08cde1627b1d |
| 2026-08-25 | mlb | medium | 2 | 3.35 | win | slip_2026-08-25_conservative_baf50ff03ad2 |
| 2026-08-25 | mlb | longshot | 5 | 42.39 | loss | slip_2026-08-25_aggressive_6ef8e36de0cb |
| 2026-08-24 | mlb | medium | 2 | 2.62 | loss | slip_2026-08-24_conservative_4f8f00680e4a |
| 2026-08-24 | mlb | high | 3 | 5.46 | loss | slip_2026-08-24_balanced_aef097a3d015 |
| 2026-08-23 | mlb | medium | 2 | 2.32 | loss | opt_2026-08-23_public_low_mlb_5b7bd33ca28d |
| 2026-08-23 | mlb | high | 4 | 6.60 | loss | opt_2026-08-23_public_medium_mlb_cf2e587083f0 |
| 2026-08-23 | mlb | longshot | 5 | 11.47 | loss | opt_2026-08-23_public_longshot_mlb_e20cb76efdad |
| 2026-08-23 | epl | medium | 2 | 2.40 | loss | epl-medium-2026-08-23 |
| 2026-08-22 | mlb | medium | 2 | 2.66 | pending | slip_2026-08-22_conservative_f12f43ed4ecf |
| 2026-08-22 | mlb | high | 3 | 6.70 | loss | slip_2026-08-22_balanced_e58186663874 |
| 2026-08-22 | mlb | longshot | 5 | 27.48 | loss | slip_2026-08-22_aggressive_79b182e795f8 |
| 2026-08-22 | ufc | medium | 2 | 3.35 | loss | ufc-medium-2026-08-22 |
| 2026-08-22 | ufc | high | 3 | 6.24 | loss | ufc-high-2026-08-22 |
| 2026-08-22 | ufc | longshot | 2 | 15.65 | pending | ufc-longshot-2026-08-22 |
| 2026-08-19 | mlb | low | 2 | 1.97 | loss | opt_2026-08-19_public_low_mlb_209431d5ae10 |
| 2026-08-19 | mlb | medium | 2 | 2.26 | win | opt_2026-08-19_public_low_mlb_df57240200b9 |
| 2026-08-19 | mlb | high | 3 | 4.53 | loss | opt_2026-08-19_public_medium_mlb_6f33fcc0fb0c |
| 2026-08-19 | mlb | longshot | 5 | 10.67 | loss | opt_2026-08-19_public_high_mlb_17db1ad1db48 |
Showing the 50 most recent of 55. Narrow the range to see the rest — the totals above count every one of them.
- Model picks — every published per-sport model read vs the official outcome; graded from official results only.
- Suggested parlays — the Lab’s own daily cards, by sport and risk tier; a card grades only when every leg has an official result.
- Saved slips — cards readers built and kept (this page’s per-date sections below); a different population from the Lab’s suggestions.
- Paper products — the Bank Builder / Moonshot bankroll journeys, settled card by card in their own ledger.
Every row we generated
These counts start from what the model actually produced, not from what happened to be graded. A row that never resolved, a slate we withheld, and a date we never produced a slate for each keep their place in this list — removing any of them would quietly improve every number beside it.
2026-09-06
In progressGames are still being played. Rows resolve as they finish; nothing is counted early.
- Generated
- 615
- Win
- 0
- Loss
- 0
- Void
- 0
- Pending
- 553
- Unavailable
- 0
- No play
- 62
No result on this slate has been decided yet, so there is no rate to show for it. Every generated row is accounted for.
2026-09-05
CompleteEvery generated row reached a final state.
- Generated
- 673
- Win
- 297
- Loss
- 264
- Void
- 46
- Pending
- 0
- Unavailable
- 0
- No play
- 66
Decisive record 52.94% (297 of 561 decided results on the 2026-09-05 slate; 95% interval 48.80%–57.04%). Every generated row is accounted for.
- 607 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
2026-09-04
CompleteEvery generated row reached a final state.
- Generated
- 689
- Win
- 300
- Loss
- 284
- Void
- 34
- Pending
- 0
- Unavailable
- 0
- No play
- 71
Decisive record 51.37% (300 of 584 decided results on the 2026-09-04 slate; 95% interval 47.32%–55.40%). Every generated row is accounted for.
- 618 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
2026-09-03
CompleteEvery generated row reached a final state.
- Generated
- 358
- Win
- 162
- Loss
- 127
- Void
- 25
- Pending
- 0
- Unavailable
- 0
- No play
- 44
Decisive record 56.06% (162 of 289 decided results on the 2026-09-03 slate; 95% interval 50.29%–61.66%). Every generated row is accounted for.
- 314 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
2026-09-02
CompleteEvery generated row reached a final state.
- Generated
- 581
- Win
- 273
- Loss
- 221
- Void
- 43
- Pending
- 0
- Unavailable
- 0
- No play
- 44
Decisive record 55.26% (273 of 494 decided results on the 2026-09-02 slate; 95% interval 50.85%–59.59%). Every generated row is accounted for.
- 537 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
2026-09-01
CompleteEvery generated row reached a final state.
- Generated
- 681
- Win
- 311
- Loss
- 260
- Void
- 37
- Pending
- 0
- Unavailable
- 2
- No play
- 71
Decisive record 54.47% (311 of 571 decided results on the 2026-09-01 slate; 95% interval 50.36%–58.51%). Every generated row is accounted for.
- 2 row(s) were generated and gradable but produced no stat. They are recovered from the comparison report because the ledger does not record them.
- 608 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
2026-08-31
CompleteEvery generated row reached a final state.
- Generated
- 533
- Win
- 249
- Loss
- 218
- Void
- 23
- Pending
- 0
- Unavailable
- 0
- No play
- 43
Decisive record 53.32% (249 of 467 decided results on the 2026-08-31 slate; 95% interval 48.79%–57.80%). Every generated row is accounted for.
- 490 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
2026-08-30
CompleteEvery generated row reached a final state.
- Generated
- 328
- Win
- 135
- Loss
- 133
- Void
- 25
- Pending
- 0
- Unavailable
- 0
- No play
- 35
Decisive record 50.37% (135 of 268 decided results on the 2026-08-30 slate; 95% interval 44.42%–56.31%). Every generated row is accounted for.
- 293 settled row(s) predate canonical event lineage and are labelled legacy rather than retro-stamped.
What each outcome means
- Win
- The official box score resolved above or below the line, on the side we took.
- Loss
- The official box score resolved against the side we took.
- Void
- The result landed exactly on the line, or the player never came to bat. Not a loss.
- Pending
- The event has not produced a gradable result yet. Not a loss.
- Unavailable
- The game finished but this row never produced a gradable stat. Not a loss.
- No play
- We generated no directional call here, so there is nothing to grade.
- Withheld
- Settlement was withheld because the data failed an integrity check. No outcome was published.
Parlays by risk level
One flat unit per card, 68 graded days (2026-05-25 → 2026-09-05). Return on investment is the column that matters: a hit rate cannot be read without the price it was paid at.
| Tier | Price band | W–L | Hit rate | Paper ROI |
|---|---|---|---|---|
| Low risk | −200 to +100 | 137–213 | 39.1% | -3.1% |
| Medium risk | +100 to +300 | 63–291 | 17.8% | -5.6% |
| High risk | +300 to +600 | 45–309 | 12.7% | +8.3% |
| Longshot | > +600 | 24–333 | 6.7% | -7.1% |
| All tiers | 269–1146 | -1.9% |
Every tier is negative. Published because the record is the point — a card shown without it is a claim, and this stream has not earned one.
Paper only · separate ledger · never part of the settled product record or the bankroll
Deeper transparency & model audit
The sportsbook market remains the better probability on the settled record. Every hit-rate figure below should be read against that baseline — a projection clearing 50% is not the same as out-predicting the market. The raw model has averaged 9.1pp more confidence than the settled outcomes justified.
Model Projection Accuracy
Individual picks are graded leg-by-leg against the line on settled slates. This is the cleaner read on the model — parlays are higher variance (see below).
Leg-level accuracy across settled slates since 2026-05-27. Pushes and voids are excluded from the denominator. Not a profit or guarantee claim.
On settled slates, heavy favorites have hit 67.5% while plus-money props hit 41.5% — that's why plus-money lives in High Risk & Longshot, not the conservative Bank Builder.
We track every settled pick — including losing days — and weight markets from the results, not from hype. Sample-floored and shrunk toward 50% so a short hot streak can't inflate a number. Not a profit or guarantee claim.
Parlays are stricter: every leg must hit for the card to cash, so card hit rate is naturally lower than individual projection accuracy. The published cards you saw are tracked below for full transparency.
Settled slate: Sep 5
Public tracking from 2026-05-27. Slips are saved before games and graded after — hit rate counts only finished slips; pending and pushes are shown separately.
Published cards are the ones shown to users on Suggested Parlays before games. The second number is every card the model generated, tracked internally for transparency. Only settled outcomes count toward hit rate — pending games are not counted yet, and pushes are listed separately.
Sep 5 breakdowns
published cards — by risk and by sport mixThe breakdowns below count the published cards shown on Suggested Parlays for this slate. The all-generated-cards record (every card the model produced, tracked internally) is the second card at the top.
Sections are derived from each settled official slip's combined odds and leg count using the same strict gate the Suggested mode applies pregame. Slips that don't fall in any section's window are tracked but not shown.
Mixed = published cross-sport (NBA + MLB) cards. When no Mixed cards settled on this slate, the row reads zero.
Low Risk2–3 legs · under +3006 slips
- WINBryan Torres · STL vs COL · Hits Under 1.5 · final 1
- WINAlec Burleson · STL vs COL · Hits Under 1.5 · final 1
- WINKaelen Culpepper · MIN vs CWS · Hits Over 0.5 · final 2
- WINTrevor Story · BOS vs BAL · Hits Over 0.5 · final 3
- WINBen Rice · NYY vs SD · Hits Over 0.5 · final 3
- WINRiley Greene · DET vs CLE · Hits + Runs + RBIs Over 0.5 · final 11
- WINLawrence Butler · ATH vs SEA · Hits Over 0.5 · final 1
- WINLiam Hicks · TB vs TEX · Hits Over 0.5 · final 2
- WINDrake Baldwin · ATL vs PHI · Hits Over 0.5 · final 2
- LOSSJonah Heim · ATH vs SEA · Hits + Runs + RBIs Over 0.5 · final 0
Medium Risk3–4 legs · +300 to +5996 slips
- WINMichael Massey · KC vs TOR · Hits Over 0.5 · final 1
- WINBryce Harper · PHI vs ATL · Hits Over 0.5 · final 2
- LOSSOzzie Albies · ATL vs PHI · Hits Over 0.5 · final 0
- WINGleyber Torres · DET vs CLE · Hits Over 0.5 · final 2
- LOSSBrandon Nimmo · TEX vs TB · Hits Over 0.5 · final 0
- WINDylan Beavers · BAL vs BOS · Hits Over 0.5 · final 1
- LOSSSamad Taylor · SD vs NYY · Hits Over 0.5 · final 0
- LOSSChandler Simpson · TB vs TEX · Hits Over 0.5 · final 0
- WINCody Bellinger · NYY vs SD · Hits Over 0.5 · final 1
- LOSSNolan Gorman · STL vs COL · Hits Under 0.5 · final 1
- WINJunior Caminero · TB vs TEX · Hits Over 0.5 · final 2
- WINJavier Sanoja · MIA vs CHC · Hits Over 0.5 · final 1
- WINJeff McNeil · ATH vs SEA · Hits Over 0.5 · final 1
- LOSSCoby Mayo · BAL vs BOS · Hits Over 0.5 · final 0
- LOSSEdmundo Sosa · PHI vs ATL · Hits Over 0.5 · final 0
- WINJosh Bell · MIN vs CWS · Hits Over 0.5 · final 2
- WINCal Raleigh · SEA vs ATH · Hits Over 0.5 · final 3
- LOSSFreddie Freeman · LAD vs WSH · Hits Over 0.5 · final 0
- LOSSTy France · SD vs NYY · Hits Over 0.5 · final 0
- WINLuke Keaschall · MIN vs CWS · Hits Over 0.5 · final 2
- LOSSDavid Fry · CLE vs DET · Hits Under 0.5 · final 1
High Risk4–5 legs · +600 to +9996 slips
- WINGleyber Torres · DET vs CLE · Hits Over 0.5 · final 2
- WINLawrence Butler · ATH vs SEA · Hits Over 0.5 · final 1
- WINLiam Hicks · TB vs TEX · Hits Over 0.5 · final 2
- WINTrevor Story · BOS vs BAL · Hits Over 0.5 · final 3
- LOSSJonah Heim · ATH vs SEA · Hits Over 0.5 · final 0
- WINKaelen Culpepper · MIN vs CWS · Hits Over 0.5 · final 2
- WINBryan Torres · STL vs COL · Hits Under 1.5 · final 1
- WINDylan Beavers · BAL vs BOS · Hits Over 0.5 · final 1
- WINAlejandro Kirk · TOR vs KC · Hits Under 1.5 · final 1
- LOSSAlex Bregman · CHC vs MIA · Hits Over 0.5 · final 0
- WINRyan Kreidler · MIN vs CWS · Hits Under 0.5 · final 0
- WINJake Burger · TEX vs TB · Hits Over 0.5 · final 1
- WINMichael Massey · KC vs TOR · Hits Over 0.5 · final 1
- WINBryce Harper · PHI vs ATL · Hits Over 0.5 · final 2
- LOSSOzzie Albies · ATL vs PHI · Hits Over 0.5 · final 0
- WINJunior Caminero · TB vs TEX · Hits Over 0.5 · final 2
- WINBen Rice · NYY vs SD · Hits Over 0.5 · final 3
- LOSSBrandon Nimmo · TEX vs TB · Hits Over 0.5 · final 0
- LOSSSamad Taylor · SD vs NYY · Hits Over 0.5 · final 0
- LOSSChandler Simpson · TB vs TEX · Hits Over 0.5 · final 0
- LOSSNolan Gorman · STL vs COL · Hits Under 0.5 · final 1
- WINJeff McNeil · ATH vs SEA · Hits Over 0.5 · final 1
- LOSSCoby Mayo · BAL vs BOS · Hits Over 0.5 · final 0
- LOSSEdmundo Sosa · PHI vs ATL · Hits Over 0.5 · final 0
- WINJosh Bell · MIN vs CWS · Hits Over 0.5 · final 2
- WINCody Bellinger · NYY vs SD · Hits Over 0.5 · final 1
Longshot5–6 legs · +1000 and up6 slips
- LOSSJonah Heim · ATH vs SEA · Hits Over 0.5 · final 0
- WINKaelen Culpepper · MIN vs CWS · Hits Over 0.5 · final 2
- WINTrevor Story · BOS vs BAL · Hits Over 0.5 · final 3
- WINBryan Torres · STL vs COL · Hits Under 1.5 · final 1
- WINDylan Beavers · BAL vs BOS · Hits Over 0.5 · final 1
- WINTaylor Ward · SEA vs ATH · Hits Under 0.5 · final 0
- WINCal Raleigh · SEA vs ATH · Hits Over 0.5 · final 3
- WINAlejandro Kirk · TOR vs KC · Hits Under 1.5 · final 1
- LOSSAlex Bregman · CHC vs MIA · Hits Over 0.5 · final 0
- LOSSFreddie Freeman · LAD vs WSH · Hits Over 0.5 · final 0
- LOSSTy France · SD vs NYY · Hits Over 0.5 · final 0
- WINLuke Keaschall · MIN vs CWS · Hits Over 0.5 · final 2
- WINJake Burger · TEX vs TB · Hits Over 0.5 · final 1
- WINWade Meckler · LAA vs PIT · Hits Over 0.5 · final 2
- LOSSDavid Fry · CLE vs DET · Hits Under 0.5 · final 1
- WINGleyber Torres · DET vs CLE · Hits Over 0.5 · final 2
- WINLawrence Butler · ATH vs SEA · Hits Over 0.5 · final 1
- WINLiam Hicks · TB vs TEX · Hits Over 0.5 · final 2
- WINBen Rice · NYY vs SD · Hits Over 0.5 · final 3
- LOSSNolan Gorman · STL vs COL · Hits Under 0.5 · final 1
- WINMichael Massey · KC vs TOR · Hits Over 0.5 · final 1
- WINBryce Harper · PHI vs ATL · Hits Over 0.5 · final 2
- LOSSOzzie Albies · ATL vs PHI · Hits Over 0.5 · final 0
- WINJunior Caminero · TB vs TEX · Hits Over 0.5 · final 2
- WINJavier Sanoja · MIA vs CHC · Hits Over 0.5 · final 1
- LOSSSamad Taylor · SD vs NYY · Hits Over 0.5 · final 0
- LOSSCarter Jensen · KC vs TOR · Hits Over 0.5 · final 0
- WINErnie Clement · TOR vs KC · Hits Under 1.5 · final 1
- WINBryan Reynolds · PIT vs LAA · Hits Over 0.5 · final 3
- WINRyan Kreidler · MIN vs CWS · Hits Under 0.5 · final 0
- VOIDWalker Jenkins · MIN vs CWS · Hits Under 0.5
Policy window: 7/3 confirming days — model demotions: market:batter_total_bases
2026-09-05
Missed slips · 27tap to expand
2026-09-04
Missed slips · 42tap to expand
2026-09-03
Missed slips · 52tap to expand
Every older graded day has its own full page — slip by slip, nothing dropped; each still counts in the lifetime and by-profile records above.
2026-09-022026-09-012026-08-312026-08-302026-08-292026-08-282026-08-272026-08-262026-08-252026-08-242026-08-232026-08-222026-08-212026-08-202026-08-192026-08-182026-08-172026-08-162026-08-152026-08-142026-08-132026-08-122026-08-112026-08-102026-08-092026-08-082026-08-072026-08-062026-08-052026-08-042026-08-032026-07-312026-07-302026-07-272026-07-262026-07-252026-07-242026-07-232026-07-222026-06-182026-06-172026-06-162026-06-152026-06-142026-06-132026-06-122026-06-112026-06-102026-06-092026-06-082026-06-072026-06-062026-06-052026-06-042026-06-032026-06-022026-06-012026-05-302026-05-292026-05-282026-05-27
By model profile
historical view · the framing we shipped earlierHit rate = wins ÷ (wins + losses). Pushes and pending slips are excluded from the denominator. "Pending" is shown separately so you always know how many slips are still in flight.
Learning signals· 1 confirmed · 2 in review · 13 tracking · 1 too smallclick to expand
- Conservative profileWithin the variance band; no action needed.Sample548 (n ≥ 60)Direction+15.7pp vs lifetimeTracking
- Balanced profileWithin the variance band; no action needed.Sample565 (n ≥ 60)Direction-4.6pp vs lifetimeTracking
- Aggressive profilePast the n + gap thresholds — qualifies for silent evaluation before any card-selection change.Sample574 (n ≥ 60)Direction-10.4pp vs lifetimeShadow-test candidate
- Star Power profileWithin the variance band; no action needed.Sample574 (n ≥ 60)Direction-1.5pp vs lifetimeTracking
- Low Risk sectionWithin the published floor; section cap stays unchanged.Sample350 (n ≥ 40)Direction39.1% hit · floor 35%Tracking
- Medium Risk sectionBelow the published floor and past the n=40 gate — flagged for shadow evaluation.Sample354 (n ≥ 40)Direction17.8% hit · floor 22%Shadow-test candidate
- High Risk sectionWithin the published floor; section cap stays unchanged.Sample354 (n ≥ 40)Direction12.7% hit · floor 12%Tracking
- Longshot sectionWithin the published floor; section cap stays unchanged.Sample357 (n ≥ 40)Direction6.7% hit · floor 6%Tracking
- NBA-only publicSample below n=40; needs 22 more decisive slips before this bucket can move audit-policy state.Sample18 (n < 40)Direction27.8% hit · 5-13Too small to act on
- MLB-only publicTracked at the audit-policy level — no behavior change yet.Sample1418 (n ≥ 40)Direction18.3% hit · 259-1159Tracking
- Mixed (NBA + MLB) publicTracked at the audit-policy level — no behavior change yet.Sample99 (n ≥ 40)Direction9.1% hit · 9-90Tracking
- Audit signal · mixedSportDownrankFired 0 of 3 required days; still under threshold.Sample0 (n < 3)Directionfires trackedTracking
- Audit signal · sameGameNbaCapFired 0 of 3 required days; still under threshold.Sample0 (n < 3)Directionfires trackedTracking
- Audit signal · market:batter_hits_runs_rbisFired 1 of 3 required days; still under threshold.Sample1 (n < 3)Directionfires trackedTracking
- Audit signal · market:batter_total_basesConfirming-days threshold cleared — pending operator approval before card selection uses it.Sample5 (n ≥ 3)Directionfires trackedConfirmed · operator review
- Audit signal · dnpGuardStrengthenFired 0 of 3 required days; still under threshold.Sample0 (n < 3)Directionfires trackedTracking
- Audit signal · longshotKeepCollapsedFired 0 of 1 required days; still under threshold.Sample0 (n < 1)Directionfires trackedTracking
Every signal here is tracked first, reviewed internally second, and only then considered for card selection. We never say the model "learned" from a single slate, and we do not consume an audit signal until the confirming-days threshold in the policy clears.
How the model improves over time.
Three things happen after every slate. The first two are automatic. The third is a slow loop — by design.
- Settlement
Once games finish, every saved slip is graded against final stats. Wins, losses, pushes, and pending are recorded by the nightly settler. We never edit outcomes by hand and we never use a slate's final results to alter that slate's pre-game suggestions.
- Daily projection audit
Per-prop accuracy is compared to the model's pre-game projection on every settled lean. Per-sport audits live at /results/nba and /results/mlb. The raw audit JSON is published alongside this site, so every number here can be checked end-to-end.
- Signal review
Underperforming markets, tiers, or matchup shapes get flagged via the audit policy you can see in the banner above. A signal has to repeat across a configured window of confirming days before it can affect card selection. One slate never changes the model on its own.
- We found a grading blind spot. Hits + Runs + RBIs props used to come back unresolved because the MLB grader didn't know how to read them. They now grade off the box score the same way Hits and Total Bases do.
- We track public risk sections separately. Low / Medium / High / Longshot get their own results row. The hit rate you see on Results matches what you saw on the Parlay Center the day before, not a re-bucketing.
- We still need a sample before any model behavior moves. Every Learning Signals row says exactly how many more decisive slips it needs before its gate can fire. Confirmed signals stay operator-gated and aren't consumed automatically.
What happens next: audit signals are tracked first, reviewed internally second, and only then considered for the card selection. Advanced ML ideas are on the roadmap, but not active until they beat the current rules in out-of-time tests. Full sequence — MODEL_LEARNING_ROADMAP_2026-05-28.md.
- We do not call this a learning model. Until an audit signal is explicitly consumed by card selection, every flag in the banner above is informational only.
- Custom parlays from Build Your Own are not included in the public hit rate. They're exploratory.
- Bankroll Plan is an educational allocation planner — not financial advice.
- Single-game NBA cards in the Parlay Center are labeled higher variance because every leg shares one matchup. That label is the framing, not a hedge.
