Polls & Odds

Methodology. The polling average is a weighted average of the most recent national polls of the 46th federal election: each poll's weight halves every 14 days and scales with sample size; no single pollster contributes more than 35% of the total weight, so weekly trackers cannot dominate the average. The seat projection applies each party's regional swing — our current regional standing minus its 2025 result — to all 343 ridings' official 45th general election results (Elections Canada), then runs 10,000 simulations with a ±3-point correlated regional error and ±4.5-point riding-level noise. It is deliberately simpler than dedicated seat models (no incumbency, candidate or riding effects): read the ranges, not the point estimates. The trend line applies the same average at weekly steps; each dot is one published poll. Poll figures are compiled into Open Data Canada's own poll registry directly from each polling firm's published releases — Léger, Abacus Data, Nanos Research, Liaison Strategies, EKOS, Mainstreet Research and others; where the release is public, the pollster's name in the table links to it. Provincial shading uses dedicated provincial polls where fresh ones exist; where they don't (Ontario has had no dedicated federal sample since May 2025), we pool the provincial crosstabs printed in at least two recent national releases — pooled subsamples carry larger error than full provincial polls and are labelled as such on the map. Prediction-market probabilities are live prices from Polymarket — the share price of a contract paying $1 if the outcome happens, read as the market's implied probability; markets under US$25,000 traded are flagged as thin, since a handful of bettors can move them. A poll measures support today; a market prices the final outcome — they answer different questions, which is exactly why comparing them is informative. This page is not affiliated with any pollster, party, or Polymarket.