Where the edges are, and how to stake them

Where the edge is

Moneylines on a main event are the most efficient market on any card. They are the line the book spends the most time on, the one that attracts the most money, and the one where a public opinion is already priced in. If your only view is who wins, you are competing with the book at its strongest point.

The edge in bare-knuckle sits in the derivative markets: method of victory, inside the distance, round groups and knockdown totals. These are priced by algorithm off gloved-boxing assumptions, and bare-knuckle does not behave like gloved boxing.

The pacing mismatch

Bare-knuckle fights end early and they end often. On a recent BKB card, ten of eleven bouts ended by KO or TKO and seven did not survive the opening round. A book pricing a total at two and a half rounds is using a distribution built for a sport where fighters wear fourteen ounces of padding on each hand. Exposed knuckles change both how much damage lands and how long a fighter can keep throwing.

The complication is that round length is no longer consistent across the sport. BKFC uses two-minute rounds. BKB moved to three-minute rounds after merging with BYB Extreme. A total of two and a half rounds means something materially different on each card, and anyone carrying one promotion's instincts to the other will be wrong in a predictable direction.

The cut variable

A ringside physician can stop a bout on a laceration at any point, which means a fighter who is comfortably ahead can lose to a wound opened by an accidental head clash. No other combat sport has a loss condition this arbitrary and this frequent.

Books price this crudely, if at all, usually by folding doctor stoppages into a general KO/TKO bucket. That is a reasonable simplification for them and an opportunity for anyone tracking cut history at the level of individual causes. A fighter who cuts from clean punching has a durability problem. A fighter who cuts from head collisions has a technique problem or a bad-luck profile. Those are different risks and they should not be priced identically.

Hand integrity

The most under-reported variable in the sport. Wrapping stops at the knuckle, so a fighter's offence can degrade mid-fight from damage to his own hands. Watch enough tape and you can see the moment a man stops throwing his right, or switches to hooks only, or starts leading with the shoulder. No public statistic tracks it.

This matters because it predicts a fighter failing rather than his opponent winning, and those are different bets. On a moneyline you cannot express "I think this man's hands go." On an exchange, where laying a fighter is possible, you can.

Staking

Bare-knuckle carries more variance than almost any market a recreational bettor will touch. Small samples, thin liquidity, and a stoppage condition that can end a fight at random. Stake sizing should reflect that, not the strength of a feeling.

• Flat stakes. One unit is the standard position. Two units is the ceiling, reserved for the rare case where the model and the tape agree strongly and the price is clearly wrong.
• A unit is one per cent of the bankroll or less. At that size a ten-bet losing run is survivable, and in this sport a ten-bet losing run is not unusual.
• No chasing and no live-betting a losing position. The momentum swings that make bare-knuckle compelling to watch make it hostile to in-play recovery.
• Log the price you took and the closing price. Every time, including losers.

Closing line value

Whether a bet won is mostly noise over a small sample. Whether the price shortened after you took it is signal. If you consistently take a fighter at +150 and he closes at +120, the market moved toward your opinion, and over a long enough run that is what profit looks like before variance is accounted for.

Offshore books post bare-knuckle lines several days before regulated books, which is where most closing-line value comes from. Taking a number early and watching it shorten is the whole exercise. BKBX logs the taken price and the closing price on every position for exactly this reason, and publishes the average.

A pre-fight checklist

1. Do both fighters have three or more charted fights? If not, the read is a prior, not a prediction.
2. What is the round length on this card, and does the total reflect it?
3. Has either fighter been stopped on a cut before, and from what cause?
4. Has either fighter shown hand failure in a previous fight, and in which round?
5. How long has each man been out? Layoff is the least priced variable in the sport.
6. Is there a common opponent, and what happened?
7. Is the market price outside the model's fair price by enough to cover the vig?
8. If the answer to any of the above is unknown, does the stake reflect that?

A position that fails this checklist is a preference, not a bet. Log it as such or do not take it.

A note on the obvious

Bare-knuckle is a thin market with small samples, and no model built on it should be mistaken for a prediction. Everything on this site is published as a prior to be argued with. Bet only what you can afford to lose, and if it stops being enjoyable, stop.