The record

3Wins

0Losses

100.0%

2–0 Favourites below even

1–0 Underdogs +100 to +400

0–0 Longshots over +400

+$367

Total profit through the site

$367 from 3 graded tier picks + $0 from 0 personal bets · $1,192 riding on 6 live tier picks

Too few graded bets to quote a return — 3 of the 30 needed. At this sample a percentage would be noise, not a rate, so we show the dollars and wait.

How often we pick the winner

every bout we called before the fight, graded — not just the bets

Full stackthe number we publish — the engine plus the analyst read

67% ± 16 pts · 6 of 9 fights

Engine alonethe simulator with the analyst layer switched off

67% ± 16 pts · 6 of 9 fights

One row per fight, the call frozen before it happened and never edited afterwards. The ± is one standard error: with this many fights the true rate could sit that far either side of the number. A fight where our number sat within 2 points of 50% is a coin flip — no call, so it is not scored (1 so far on the full stack, 1 on the engine alone).

By tier

Gap: our sim vs. de-vigged price

FAVOURITE

A favourite we rate higher still · 4+ pt gap · to win $150

2–0 Record

100.0%Win rate

+$199Profit

—Return

6 live · CLV +0.9 pts

UNDERDOG

A near-even underdog we rate higher · risk $150

1–0 Record

100.0%Win rate

+$168Profit

—Return

CLV −0.9 pts

LOW CONFIDENCE

Our read under 30% — tracked, never in the tier record

0–0 Record

—Win rate

—Profit

—Return

Kept out of the tier record on purpose.

ROI = profit over money risked · CLV = how far the no-vig market moved toward our side between the call and the close, in percentage points. Positive means the market came to us. Both ends are de-vigged consensus across every book in the capture, so the number is market movement and not the bookmaker's margin.

Live picks

6 frozen, waiting on a result

Pick	Gap	Price	Our %	Market %	Tier	Fight	Event
Karine Silva	13.9	-170 betonlineag	74.3%	60.4%	FAVOURITE	Karine Silva vs Gabriella Fernandes	UFC Fight Night 293 November 7, 2026
Elijah Smith	11.9	-172 betonlineag	77.9%	66.0%	FAVOURITE	Davey Grant vs Elijah Smith	UFC Fight Night 293 November 7, 2026
Tatiana Suarez	6.4	-160 betonlineag	69.1%	62.7%	FAVOURITE	Tatiana Suarez vs Virna Jandiroba	UFC Fight Night: Cortes-Acosta vs. Walker November 7, 2026
Waldo Cortes-Acosta	6.3	-185 betonlineag	69.2%	62.9%	FAVOURITE	Waldo Cortes-Acosta vs Johnny Walker	UFC Fight Night: Cortes-Acosta vs. Walker November 7, 2026
Chris Duncan	16.6	-155 betonlineag	74.9%	58.3%	FAVOURITE	Drew Dober vs Chris Duncan	UFC 334: Gane vs. Hokit November 14, 2026
Navajo Stirling	15.9	-140 betonlineag	69.7%	53.8%	FAVOURITE	Jiří Procházka vs Navajo Stirling	UFC Fight Night: Procházka vs. Stirling November 21, 2026

scroll for more →

Settled

3 graded · nothing is ever removed

Pick	Result	Gap	Price	Units	CLV	Tier	Winner	Event
Tom Nolan	WON	9.0	+112	+1.12	-0.9	UNDERDOG	Tom Nolan	UFC Fight Night: Moicano vs. Nolan October 31, 2026
Jacobe Smith	WON	7.8	-250	+0.40	+1.4	FAVOURITE	Jacobe Smith	UFC Fight Night: Moicano vs. Nolan October 31, 2026
Cody Durden	WON	5.6	-108	+0.93	+0.5	FAVOURITE	Cody Durden	UFC Fight Night: Moicano vs. Nolan October 31, 2026

scroll for more →

How to read this. Every call is graded like a real bet. A unit is $150, 1.5% of a $10,000 starting bankroll. A favourite is bet to win one unit — at -150 we risk $225 to win $150 — and an underdog risks one unit, so every call has about the same upside. No bet risks more than $450, and one card never carries more than $750 in total. How strongly we disagree with the bookmakers does not change the size: across hundreds of past fights, bigger disagreements did not win more often. Return is profit divided by everything we risked. A bout where our number sits within 2 points of 50% is a coin flip: no call, no stake, and it never reaches this ledger. The headline and the three price buckets count every graded bet, so favourites + underdogs + longshots always add up to the record above. The tier cards below use a stricter cut — they leave out the low-confidence board — because that is the number the engine is judged by.

Why the board starts at zero · engine backtest

No back-dated tier record — on purpose.

A tier needs the betting line it disagreed with. We started logging lines on 5 August 2026 and the odds feed doesn't sell the past, so the board starts at zero rather than inventing one.

What we can show: the simulator replayed over 8,312 real fights on information available before each one — 60.1% of winners overall, 64.0% from 2023 on. That's the engine, not a tier record.

Analysis and entertainment only. Nothing here is betting advice.