Trifecta doesn't try to predict every matchup — it plays only its biggest edges. That selectivity is why each week's five most confident picks have covered at a remarkable rate, graded in public every week.
The model never sees a spread, a moneyline, or a bet split while forming its view of a game. It's built entirely from team and player performance data — the closing line is only used afterward, to see whether the model actually found something the market missed.
Trifecta is a purpose-built model — not a public formula, not a rented algorithm. The exact mechanics are ours; the results are public.
Every input is what was actually knowable before kickoff — no accidental use of information from later in the season. Checked automatically on every model run, not just assumed.
Most games, the model has no strong opinion — and says nothing. It only surfaces a pick when its view diverges sharply from the market. That's the entire point of the Top 5 and Most Confident tiers.
No hiding behind a filter that changes after the fact. Every Top 5 pick — including each week's single most confident call — is on the Track Record page, right or wrong.