Goalies
No official pick is taken until both goalies are confirmed. A model signal before that confirmation stays a trend.
AI Trends Journal — NHL
An AI trend is a model signal that does not meet every requirement of an official pick: goalies not yet confirmed, a read not validated by our analyst, a probability outside the zone where our model is reliable, or a market under observation (the 1st period). We publish them all, recorded before the game, and settle them against the score — won or lost. This is our public proving ground: it lets anyone check, in real conditions, that our official thresholds sit in the right place.
Why isn't a trend an official pick?
Moneyline
Official only when our analyst validates the model's read and the stated probability stays below 70%: above that, our tracking shows the model overrates its chances. Other signals stay trends.
Total goals
Official after our analyst's validation, with both goalies confirmed. Weak or unvalidated signals stay trends.
1st period
Market under observation since October 2026: no official pick, every read is published here as a trend.
Per-market summary — settled trends
How to read this table: it only shows trends (never our official picks, tracked on the track record). A winning trend does not mean it should have been played: it had not cleared our thresholds. This journal exists to verify, over time, that our thresholds sit in the right place — in both directions.
| Market | Settled trends | Record | Hit rate | Units |
|---|---|---|---|---|
| Moneyline | 664 | 321W-343L | 48.3% | -28.5 u |
| Total goals | 701 | 321W-337L | 48.8% | -39.0 u |
| 1st-period total | 24 | 17W-7L | 70.8% | +9.6 u |
| Date | Game | Market | Trend | Model gap | Odds | Result |
|---|---|---|---|---|---|---|
| 15/10 | NYI @ OTT1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 1.93 | Won |
| 15/10 | BUF @ TOR1st per. 3 goals | 1st-period total | Over 1.5 | +0.1 goals | @ 1.68 | Won |
| 15/10 | PHI @ DET1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 1.97 | Won |
| 14/10 | MTL @ WSH1st per. 4 goals | 1st-period total | Under 1.5 | +0.3 goals | @ 2.21 | Lost |
| 14/10 | DAL @ COL1st per. 0 goals | 1st-period total | Under 1.5 | +0.3 goals | @ 2.02 | Won |
| 13/10 | FLA @ CBJ1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 2.02 | Won |
| 13/10 | STL @ OTT1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 1.97 | Won |
| 11/10 | VAN @ NYR1st per. 2 goals | 1st-period total | Under 1.5 | +0.2 goals | @ 2.08 | Lost |
| 11/10 | SEA @ WSH1st per. 4 goals | 1st-period total | Under 1.5 | +0.2 goals | @ 2.13 | Lost |
| 10/10 | EDM @ SJS1st per. 1 goal | 1st-period total | Over 1.5 | +0.1 goals | @ 1.69 | Lost |
| 10/10 | TBL @ NYI1st per. 2 goals | 1st-period total | Under 1.5 | +0.1 goals | @ 1.96 | Lost |
| 10/10 | DET @ MTL1st per. 0 goals | 1st-period total | Under 1.5 | +0.1 goals | @ 2.12 | Won |
| 10/10 | PHI @ BOS1st per. 1 goal | 1st-period total | Under 1.5 | +0.2 goals | @ 1.97 | Won |
| 10/10 | CBJ @ STL1st per. 2 goals | 1st-period total | Under 1.5 | +0.3 goals | @ 2.02 | Lost |
| 10/10 | MIN @ FLA1st per. 2 goals | 1st-period total | Over 1.5 | +0.2 goals | @ 1.83 | Won |
| 10/10 | DAL @ PIT1st per. 0 goals | 1st-period total | Under 1.5 | +0.1 goals | @ 2.09 | Won |
| 08/10 | SJS @ STL1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 1.99 | Won |
| 08/10 | PHI @ OTT1st per. 1 goal | 1st-period total | Under 1.5 | +0.2 goals | @ 1.97 | Won |
| 08/10 | CHI @ NYI1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 2.01 | Won |
| 08/10 | DAL @ BUF1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 2.04 | Won |
| 08/10 | COL @ CGY1st per. 2 goals | 1st-period total | Over 1.5 | +0.3 goals | @ 1.76 | Won |
| 08/10 | VAN @ CAR1st per. 3 goals | 1st-period total | Under 1.5 | +0.2 goals | @ 2.17 | Lost |
| 05/10 | WPG @ PIT1st per. 0 goals | 1st-period total | Under 1.5 | +0.2 goals | @ 2.15 | Won |
| 12/01 | NJD @ MIN1st per. 1 goal | 1st-period total | Under 1.5 | +0.1 goals | @ 2.04 | Won |
Frequently asked questions
What is the "model gap"?
For the moneyline, it is the gap in probability points between our model and the probability implied by the odds: +5 pts means the model gave the team 5 more points of chance than the market. For totals, it is the gap in goals between the total projected by our model and the market line.
Why publish losing trends?
Because transparency is our product: the journal publishes every signal's outcome, won or lost. That is what proves our official picks come from genuinely validated zones, not hindsight selection.
Should a winning trend have been an official pick?
Not necessarily: at single-game scale, variance dominates. A zone only turns official if its hit rate holds on sufficient volume. This journal exists precisely to document those zones in real conditions.
Where can I see official picks and their results?
On the Track Record page — the raw, timestamped journal of every settled official pick — and on the NHL page for today's picks.
