◎ Analyze any market — Got a market that isn't in the list? Paste its link or ID and hit Get the odds. We price it on the spot and show the same card: each side's price, the model's fair value, and the contract payoff.
Model consensus — A live picture of our model ensemble voting on the featured market. Green dots = models that say “Yes”, red = “No”. A tight one-colour cluster means strong agreement; a mixed blob means it's a coin-flip and we stay out.
Gap map — The whole board at a glance: every coin (rows) across time-to-resolution (columns). A green cell = the model sits above the price, red = below, the number is the gap in points, ● = wide, ○ = worth watching, blank = no market.
Today's board — Your main screen. Each card: each side's price, the market price vs the model's fair value, the divergence with a 1–5 dot strength meter, and the contract payoff (win/lose per $1). Open “What would have to be true” for the reasoning, “Quant detail” for the model, or “Sentiment Scan” for a live news-and-social read.
TABLE vs CARDS — TABLE is the scanning view: one row per market across every category, and every column sorts (click a header — it opens on the biggest divergence first). Click any row to open that market's full card underneath it. CARDS is the reading view. Your choice is remembered. ANY HORIZON / ≤7D / ≤30D filters by how far away a market is — for a game, the day it is played; for everything else, the day it resolves. The board opens on ≤7D, and ANY HORIZON shows everything we price. PRICE GAPS / ALL MARKETS filters by whether there is a gap at all.
◎ Bookmakers say… (sports) — Every match card shows the same game priced by the world's leading sports bookmaker, with its hidden profit margin removed. One glance tells you whether this market's price is fair: if the two numbers agree, it is; if they're far apart, you're looking at a disagreement worth knowing about before you act. Information only — it never changes our signal.
Event odds — Multi-runner events (elections, leagues, awards) presented as a fair-value board. Each runner's probability is recovered with Shin's method, which removes the book's overround (the operator margin) while correcting the favourite-longshot bias — more accurate than proportional normalisation. Prices summing over 100% is the normal margin, not a gap; for an independent estimate use AI true-odds.
Price-gap spread — A histogram of how big the gaps are across everything we priced. Bars to the right = the model sits above the price on “Yes”; to the left = below.
Other markets — Markets we don't have a model for yet, shown at their live price for reference only — no model estimate yet.
Controls — AUTO refreshes every 60s · REFRESH recomputes now · LIGHT/DARK switches theme · GUIDE jumps here.
A probability is not a prediction: a 70% call is meant to be wrong three times in ten, and a run of those is ordinary rather than a sign that something broke. Our own record is published in full at /accuracy, including where it does not flatter us — read it before you rely on any of this. Never risk more than you can afford to lose. Kanon does not tell you what to do with any market, or how much to commit to it; that is entirely your decision. It is an information and analytics tool: it does not place trades, hold funds, or route you to any venue, and nothing here is financial or betting advice.