Skip to main content
MMA
Final Tue, Sep 29UFCGeorge Staines (winner)vsLoai AbushaarFinal Sat, Oct 3UFCNatalia Silva (winner)vsWang CongFinal Tue, Oct 6UFCPreston LaGrange (winner)vsNell ArianoSat, Oct 10 5:00 PM PTUFCBrendan AllenvsChristian Leroy DuncanTue, Oct 13 4:00 PM PTUFCIlias BulaidvsErick ViscondeFri, Oct 16 8:00 PM PTPFLLiz CarmouchevsJena BishopSat, Oct 17 5:00 PM PTUFCMike MalottvsJoaquin BuckleySat, Oct 24 11:00 AM PTUFCAlexander VolkanovskivsMovsar EvloevSat, Oct 31 5:00 PM PTUFCRenato MoicanovsTom Nolan

Our Public Pick Record

Open Scoring MMA logs every pick before the fight and grades it from the official result, wins and losses both. Nothing is removed from the record after the fact, and a fight our model calls a genuine coin flip gets no pick at all.

How a pick gets on this record

Every pick is made and published before the fight happens, then auto-graded the moment the official result is in. That order matters: a call that could be revised after the fact would not tell you anything about whether it was any good.

When our model rates a fight a genuine coin flip, we do not make a pick, we say so on the page, and that fight stays out of the record rather than padding a win total that was never really earned.

Calls are labelled by confidence, from a lean up to a strong call, and marked separately as a value pick when our number disagrees with the betting favorite or beats the market's line by a clear margin. Each pick also shows a win probability next to the betting line's own implied probability, so a reader can see exactly how far we are leaning and against what.

The record, live

91-29
Record
76%
Hit rate
0.172
Brier score
120
Graded picks
32-18
Lean record
13-2
Value record

Auto-graded from official results. See every graded pick, and the fights still on the board, on the Picks page.

Why the record includes the losses

A record that only shows the hits is not a record, it is a highlight reel. Open Scoring MMA's hit rate is measured against the betting favorite over the same fights, so it is checkable against a real baseline rather than against nothing. The Brier score does the same job for the probabilities themselves: it grades how well-calibrated a 65% call actually was, not just whether it landed.

Spot something wrong with a graded pick or a listed result? Read how we handle mistakes on the corrections page, or see exactly how the show and the model are made on the about page.

Common questions

How do I know if an AI sports podcast is actually accurate and not just making stuff up?

Check whether it publishes a graded track record. Open Scoring MMA logs every pick before the fight, auto-grades it the moment the official result is in, and keeps the full record, wins and losses both, public on this page.

Do AI sports podcasts ever show their wrong predictions, or do they just hide the bad ones?

Open Scoring MMA does not hide losses. Every graded pick stays on the public record, and the record reports hit rate against the betting favorite plus a Brier score, so the calibration is checkable and not just the win total.

I want an MMA show that tells me how confident it is in its picks, not just who it thinks will win.

Every call on Open Scoring MMA is labelled by confidence, from a lean up to a strong call, and marked separately as a value pick when our number disagrees with the betting favorite. Each pick also shows the model's win probability next to the betting line's own implied probability, so the strength of the call is part of the call.

How do I find an MMA podcast that actually tracks its own prediction record over time?

Open Scoring MMA's picks page keeps a running, auto-graded record: hit rate, Brier score, and the record broken out by confidence level and by value pick, updated as each fight is graded.