We pulled these three signals from the public /stats dataset of 607 entries. Cherry-picking is intentional only in showing variety of archetypes — wide gap, late-inning lockdown, tighter call — not in selecting for outcomes. Full dataset available on request.
Case 1 — Wide Gap Archetype
SEA @ KC
Game state at signal
Top 7th · KC leads 5–2 · run diff +3
Model output
P(KC wins) = 95.4%
(empirical run-diff table)
(empirical run-diff table)
Kalshi market at signal
KC contract $0.77 · implied 77.0%
Gap
+18.4 percentage points
Filters met
gap ≥10pp · model ≥0.85 · not tied · KC ≤$0.90 · inning ≥3 · gap ≤25pp
Subscriber action (60-second window)
Buy KC YES @ $0.77 → $10 = 12.99 contracts
Game result
SEA 6, KC 8 — KC won
Settlement
$10 in → $12.99 out on resolution
+29.9%gross return
Case 2 — Late Lockdown Archetype
WSH @ ATL
Game state at signal
Bottom 9th · WSH leads 2–1 · run diff +1
Model output
P(WSH wins) = 94.0%
(empirical run-diff table)
(empirical run-diff table)
Kalshi market at signal
WSH contract $0.73 · implied 73.0%
Gap
+21.0 percentage points (largest of the three)
Subscriber action (60-second window)
Buy WSH YES @ $0.73 → $10 = 13.70 contracts
Game result
WSH 2, ATL 1 — WSH won
Settlement
$10 in → $13.70 out on resolution
+37.0%gross return
Case 3 — Tighter Call Archetype
CLE @ PHI
Game state at signal
Bottom 7th · CLE leads 2–0 · run diff +2
Model output
P(CLE wins) = 86.3%
(lower confidence vs. cases 1–2)
(lower confidence vs. cases 1–2)
Kalshi market at signal
CLE contract $0.72 · implied 72.0%
Gap
+14.3 percentage points
Subscriber action (60-second window)
Buy CLE YES @ $0.72 → $10 = 13.89 contracts
Game result
CLE 3, PHI 1 — CLE won
Settlement
$10 in → $13.89 out on resolution
+38.9%gross return
What these three show
- Archetype variety, not outcome variety. Wide gap (Case 1, +18.4pp mid-game), late-inning lockdown (Case 2, +21.0pp in the bottom 9th), and a tighter floor case (Case 3, +14.3pp at 86.3% model confidence) — three distinct game-state shapes the detector fires on.
- Edge comes from the gap, not just the model probability. Case 3 has the lowest model probability (86.3%) but the highest gross return (+38.9%), because the Kalshi price was cheapest relative to the model. The signal targets mispricing, not certainty.
- Outcomes confirm the archetypes the filters were designed to catch. All three cleared the same filter stack (gap ≥10pp, model ≥0.85, inning ≥3, gap ≤25pp) and all three resolved on the YES side. This is what the filter is for; the underlying win-rate over the full 607-entry dataset is on /stats.
Honest caveat about the data above
Our current logging captures signal_at, model_p, market_p, gap, and outcome. It does not yet capture Kalshi price-over-time after the signal fires — which would let us document the exact exit window and the reprice arc as the lag closes. We're adding that to the next bot revision. The numbers above are real, but they describe the entry and the resolution, not the seconds in between.
Risk reality. Three signals do not constitute a win rate. About one in five MLB signals loses 100%, sometimes in clusters. A high win rate describes the past — it does not by itself make anyone money: your entry price, bet sizing, and discipline through losing streaks decide your actual outcome. See the full risk disclosure on /method/. This is not investment advice. Sports betting and prediction-market trading involve substantial risk — bet only with capital you can afford to lose.
Want the live alerts in your pocket?
The full methodology is on /method/. Pricing and plans below.