All scoring versions · the current scoring is version 8
Every pick we published before kickoff since September 28, 2026, graded against the final score. Each pick is saved with its time, odds and grade the moment we publish it and is never changed once its game starts. If a later update changed a game's pick, the first one counts. Flat one unit at the odds we published; ROI uses those odds, not closing lines. A pick is graded A–D when it is likely to hit and has an edge, or is a Long shot (a big edge on a side less likely to hit); the totals count both, and each is shown on its own. Wins, losses and pushes all count; games without a final score yet are not shown. Every pick is tagged with the scoring version that made it, and model updates never change past picks. 142 picks, Sep 28 to Oct 11, 2026.
A–D picks: 74-57-1, −7.73u · Long shots: 5-5-0, +0.73u
Past results do not guarantee future results. For entertainment and information only; not a sportsbook. 21+. Gambling problem? Call 1-800-GAMBLER.
Line movement: whether the price moved toward our pick between publishing and kickoff. Better than half is the earliest sign of a real edge; it shows up in about a hundred picks, long before profits are more than luck. Won vs price needed: a win rate only means something next to the price: winning 40% of +200 underdogs makes money, winning 60% of −200 favourites loses it. Net units: the bottom line, with the ROI range luck alone would produce; a result inside it isn't evidence yet.
Colors: green is ahead; orange is slightly behind (within 3 points of the price needed or of the market, within 5% on ROI and line movement); red is further behind. In the tables below, ± luck is the range chance alone would produce for that group; a difference inside it isn't evidence. Groups with fewer than 30 are tagged small sample: with a dozen groups, one or two will look impressive by chance. Hover over (or tap) a column heading for what it means. Both count games, not picks, since a game's moneyline and spread come from one read.
Sides under the likely cutoff (55% on a moneyline, 52% on a spread) whose edge clears the full bar: they lose more often than not, but the price pays for it. Included in the totals above and shown here on their own, at the price published.
| Market | Picks | W-L-P | Won | Price needed | Units | ROI | ± luck | Polymarket units (est.) | Polymarket ROI (est.) |
|---|---|---|---|---|---|---|---|---|---|
| moneyline small sample | 9 | 5-4-0 | 55.6% | 40.7% | +1.73u | 19.2% | ±75.4% | +1.69u | 18.8% |
| spread small sample | 1 | 0-1-0 | 0.0% | 38.8% | −1.00u | -100.0% | — | −1.00u | -100.0% |
All Long shots: 5-5-0, +0.73u, ROI +7.3% (±71.6%, within the noise).
| Grade | Picks | W-L-P | Won | Price needed | Units | ROI | ± luck | Polymarket units (est.) | Polymarket ROI (est.) |
|---|---|---|---|---|---|---|---|---|---|
| A small sample | 3 | 1-2-0 | 33.3% | 63.8% | −1.17u | -39.0% | ±121.8% | −1.15u | -38.3% |
| B small sample | 25 | 14-10-1 | 58.3% | 57.6% | −0.34u | -1.4% | ±34.3% | −0.00u | -0.0% |
| C small sample | 14 | 4-10-0 | 28.6% | 51.9% | −6.30u | -45.0% | ±48.3% | −6.24u | -44.6% |
| D | 90 | 55-35-0 | 61.1% | 61.4% | +0.09u | 0.1% | ±18.1% | +1.06u | 1.2% |
| LS small sample | 10 | 5-5-0 | 50.0% | 40.5% | +0.73u | 7.3% | ±71.6% | +0.69u | 6.9% |
| Sport | Picks | W-L-P | Won | Price needed | Units | ROI | ± luck | Polymarket units (est.) | Polymarket ROI (est.) |
|---|---|---|---|---|---|---|---|---|---|
| MLB small sample | 21 | 9-12-0 | 42.9% | 58.8% | −6.19u | -29.5% | ±39.0% | −6.02u | -28.7% |
| NCAAF | 42 | 24-18-0 | 57.1% | 57.7% | −1.00u | -2.4% | ±31.6% | −0.46u | -1.1% |
| NFL | 45 | 26-18-1 | 59.1% | 57.1% | +1.19u | 2.7% | ±28.7% | +1.55u | 3.5% |
| NHL | 34 | 20-14-0 | 58.8% | 60.6% | −1.00u | -2.9% | ±32.5% | −0.70u | -2.1% |
Exchange columns show: Kalshi · Polymarket · Novig
| Market | Picks | W-L-P | Won | Price needed | Units | ROI | ± luck | Polymarket units (est.) | Polymarket ROI (est.) |
|---|---|---|---|---|---|---|---|---|---|
| moneyline | 72 | 40-32-0 | 55.6% | 60.7% | −7.59u | -10.5% | ±19.7% | −7.05u | -9.8% |
| spread | 70 | 39-30-1 | 56.5% | 55.9% | +0.59u | 0.9% | ±21.7% | +1.41u | 2.0% |
| Version | Picks | W-L-P | Won | Price needed | Units | ROI | ± luck | Polymarket units (est.) | Polymarket ROI (est.) |
|---|---|---|---|---|---|---|---|---|---|
| v0 small sample | 1 | 0-1-0 | 0.0% | 52.4% | −1.00u | -100.0% | — | −1.00u | -100.0% |
| v7 | 69 | 38-30-1 | 55.9% | 59.1% | −5.58u | -8.2% | ±22.9% | −4.53u | -6.7% |
| v8 | 72 | 41-31-0 | 56.9% | 57.7% | −0.42u | -0.6% | ±23.5% | −0.10u | -0.1% |
| Date | Picks | W-L-P | Units | Polymarket (est.) |
|---|---|---|---|---|
| 2026-10-11 | 16 | 6-10-0 | −6.04u | −6.15u |
| 2026-10-10 | 22 | 12-10-0 | −0.62u | −0.60u |
| 2026-10-09 | 1 | 1-0-0 | +0.70u | +0.73u |
| 2026-10-08 | 5 | 4-1-0 | +2.01u | +2.07u |
| 2026-10-07 | 7 | 4-3-0 | −0.48u | −0.41u |
| 2026-10-06 | 5 | 3-2-0 | +0.15u | +0.22u |
| 2026-10-05 | 3 | 3-0-0 | +2.57u | +2.62u |
| 2026-10-04 | 26 | 18-7-1 | +6.02u | +6.43u |
| 2026-10-03 | 25 | 12-13-0 | −5.86u | −5.51u |
| 2026-10-02 | 7 | 5-2-0 | +0.87u | +1.02u |
| 2026-10-01 | 11 | 5-6-0 | −2.37u | −2.25u |
| 2026-09-30 | 8 | 5-3-0 | +0.13u | +0.26u |
| 2026-09-29 | 5 | 0-5-0 | −5.00u | −5.00u |
| 2026-09-28 | 1 | 1-0-0 | +0.91u | +0.93u |
Method: one unit per pick at the recorded odds; ROI is net units over units risked. Exchange (est.): each pick bought at the market's fair price plus that exchange's pregame fee (Kalshi 7%, 3.5% on baseball; Polymarket 6.95%; Novig none), using the exchange's own price from October 8, 2026 where it listed the same line. Prices and fees vary. Past results do not predict future ones.