CFL Predictions & Picks
Model predictions for every game • Picks •2026 Season
Error loading data
> Our CFL prediction model outputs moneyline win probability, spread cover lean, and over/under direction for every game of the Canadian Football League season. The model has been running since 2022 and is still building its baseline data set -- V1 ML results by season: 82.1% (2022 -- partial season, small sample), 66.7% (2023), 60.7% (2024), 58.7% (2025). We're transparent about where the dataset is small and the signal is still developing. Full records are on the historical records page. Beta model tiers are planned for the upcoming season.
NFL 2025-26 Season Complete
47-38 | +100.5u
vs last season: +55.3u improvement ↑
Season begins October 2026
CFL 2025-26 Season Complete
47-38 | +100.5u
vs last season: +55.3u improvement ↑
Season begins October 2026
Today's Games • Apr 18
No games schedule today
MODEL PERFORMANCE
TEAM ACCURACY
Season: 12-8 | +4.50u
Today's Picks • Apr 17
No picks posted yet for today
Last night: 1-0 | +0.91u
No results yet
LAST NIGHT: 1-3 • -2.17u
BETTING RECORD
1-3 | -2.17u
TEAM PICKS RECORDS
CFL Market Dynamics
The CFL betting market is smaller and less liquid than NFL, which creates opportunities. Books set CFL lines with less sharp money input than a comparable NFL week -- mispriced lines tend to stay in place longer and move more predictably once the public catches on. The model targets the ML, spread and totals markets in particular, where the smaller market tends to lag on adjustments.
CFL also kicks off in June in overlaps with MLB season, making it one of the most active summer betting periods for multi-sport bettors.
Beta Models Coming For CFL
The same multi-tier model architecture launched for WNBA in 2026 is planned for CFL in the upcoming season -- V1 plus two progressively tighter confidence models. The process is consistent across every sports: build the baseline with V1, publish all the results, then add tighter buckets once there's enough history to calibrate them. The CFL dataset is still young, which means model improvement here has the most room to run.