What Cannon Fight Lab is, what it measures, and what it refuses to claim.
Cannon Fight Lab is a UFC research and market-intelligence tool. It exists to let you work a fight card faster and more honestly than you could with a dozen browser tabs open — and to say plainly which of the things people believe about fights are actually supported.
Every other corner of MMA betting is dominated by hot takes, picks pages, and Discord touts charging $50/month to tell you what to bet. The actual data — strike accuracy by division, age advantage, reach gap win rates, post-loss form, finish patterns — is either locked behind premium subscriptions, scattered across a dozen forums, or never surfaced at all.
This site consolidates that data layer into one fast, searchable place, and adds two things on top. The first is the market: sportsbook prices captured on a schedule and written to an append-only record, so we can show you what a line was, what it is, and what moved in between — with the book count, the source and the timestamp attached to every number. The second is derived analysis: matchup comparisons, cardio and style ratings, and public testing of which factors survive once you control for what the betting market already knew.
Fight statistics are updated nightly from UFCStats; sportsbook prices come from The Odds API on a capture schedule that tightens as a card approaches. Both are inputs to CFL's research, not the product — we are not in the business of charging anyone for access to somebody else's feed. The platform doesn't sell picks. It doesn't run a Discord. It doesn't tell you who to bet on.
Card Lab is the fight-week screen: every bout on the next card with the vig-free sportsbook consensus, the best price and which book is posting it, and how far the line has travelled since we started watching. Market Lab is the same card broken out book by book, with movement over the last 24 hours and since the first time we saw the fight priced by three or more books, and where the sportsbooks currently disagree. Fight Lab is one matchup in full: the market, the per-book table, and the measurable differences between the two corners. The Factor Lab is the research floor — it tests the stats everyone cites twice, once raw and once on only the fights the market priced as a coin flip, and reports which ones survive. Most do not, and it says so.
Until September 2026 the front of this site was a forecast: a calibrated win probability for every fight, a confidence tier, and a flag when our number disagreed with the market price by enough. We tested whether that disagreement was worth acting on. We could not show that it was — our own benchmark has the engine landing within a few points of the closing line without ever seeing one, which is a respectable model and is not an advantage anyone can bet.
So the forecast came off the public product. Nothing was deleted and nothing was quietly re-labelled: the engine still runs, still writes an immutable record of what it would have said before every bell, and is still graded in the open on the Proof Center with losses at the same size as wins. If prospective evidence ever shows it beats the price, the evidence goes up before the number comes back. We changed the product rather than change the test, and we would rather be known for that than for the version of this page that quoted a win rate.
Cannon Fight Lab is operated by Cannon Fight Lab LLC, a Tennessee limited liability company. It's an independent project — no investors, no parent company, no editorial agenda beyond "show the data clearly and let people decide." We take no sportsbook money, and no book pays to sit higher in a price comparison here; "best price" means the best price in the data, always.
The site is in active development. Some features are still rolling out. Some data points (per-fight striking accuracy, control time, head-strike ratios, round-by-round cardio drop-off) are coming as additional scraping completes. If you find a bug, a fighter that's missing, or a stat that looks wrong, reach out.
This is not a sportsbook. We don't take bets. We don't issue picks, and we no longer publish a forecast of who wins. The statistics and projections shown here describe historical patterns — they don't predict any individual fight. A pattern that holds 56% of the time fails 44% of the time. Prices on this site are the last ones we captured, not live quotes; check the sportsbook before you act on any of them.
Always gamble responsibly. If you or someone you know has a gambling problem, call 1-800-GAMBLER or visit 1800gambler.net.