Reading the arbitrage board
Which pairs of books currently disagree enough that covering both sides returns more than it costs, and exactly what to stake on each.
The screen, annotated
Every column, and what its number actually means.
The green badge at the top-left of the card: your locked return on the whole position, not on either leg and not on your own money after the hedge. A $1,000 two-leg position at 2.39% comes back as $1,023.86 whichever side lands.
Each leg's price at its own book, in the odds format you chose in Settings. These two imply 47.2% and 50.5% — 97.7% together. The 2.3% they fall short of 100% is the arb, and it exists because the two books are pricing the same market against different books.
What to place at that book. The boxes are linked: type into either one and the other resizes so both outcomes pay the same. The box you typed in is bright, the computed one is dimmed, and typing in the dimmed one moves the anchor to it.
The outcome line under the two legs, recomputed as you type. It is the worst of the two returns, so it is what you actually collect. If it turns red you no longer have an arbitrage on those stakes.
Opens that book at this selection and copies this leg's exact stake to the clipboard. Exact, not rounded — rounding one leg breaks the equal-payout sizing that makes the return locked.
The engine's judgement that one book's number is an error rather than a disagreement, pinned to the leg it blames. Tap it for the reasoning. These are the fastest way to get an account limited, and they are also the prices most likely to be voided after you place.
When this pair was first seen at these prices. Arbitrage lives minutes and sometimes seconds, so a timestamp several minutes old means re-check both prices at the books before you commit.
Shown when a leg sits on an exchange. The headline percentage is already net of the venue's per-contract fee; this line says how much was taken out and what the figure was before it.
Every book's price on this same line, in place. Worth opening BEFORE you place anything: it tells you what your second-best cover is, which is the number you will need if the planned one moves.
The three controls that matter
Ignore everything else on a first read.
Sportsbooks (the chip list at the bottom of Filters)
An arbitrage between two books you cannot place at is not an arbitrage. Tick only the accounts you hold and can fund today. This removes more rows than every other filter combined, and it removes the ones that were never available to you.
Max quote age (s)
Set it to 60-120. The percentage on a card is true of the prices when they were last confirmed, not of the prices now, and a 2.4% arb quoted ten minutes ago has usually been repriced away. A tighter board of live rows beats a longer board of dead ones.
Min edge %
Your floor. Below about 1% the locked profit is smaller than what one mistimed leg costs you, so the row is not worth the execution risk. Everything else on the filter bar — segment, dates, event picker, the hedge odds bounds — can be left alone on a first read.
A worked example
From a real row to a placed bet, including the stake.
Two things about the $23.86. It is 2.39% of the money at risk, not of your bankroll, so the strategy scales with stake rather than with skill. And it is the entire return from an account whose working life, used only this way, is measured in weeks. A book that limits you to a $20 maximum has taken every future $23.86 — plus every promotion and every +EV bet you would have placed there. Price the account, not the row.
When it goes wrong
The price moved, a leg was voided, the row vanished. What to actually do.
The first leg is on and the second price has moved or vanished. You hold $483 on Over 45.5 at +112 and the -102 Under is gone.
You now have a straight bet, and the only question is what it costs to close it. Do not refresh the board hoping the old price comes back; it is being taken by other people and it is not coming back. Open All books on that row and read the best remaining price on the side you need. Say the best Under is now -118: covering needs $1,023.96 / 1.847 = $554.25, making the position $1,037.25 staked for $1,023.96 back — a locked loss of $13.29. Take it. Paying $13 to close is the correct price for removing a $483 swing you never intended to have, and a reader who refuses on principle is now a punter with a view on a total. If no book covers the side at all, you have two honest choices. Hold it as the straight bet it is, or cover at the nearest different line, which turns the position into a middle. Decide deliberately rather than leaving it open by not choosing. Next time, place the leg at the softer book first: it is the one most likely to move, and if it moves before you have committed anything you have lost nothing.
The book accepted the bet at a worse price than the card showed, or accepted only part of the stake. The slip confirms $483 at +106, not +112.
Re-anchor before placing the second leg. Type the stake you actually got on into its box and read the outcome line again. The other leg resizes itself, and the line tells you whether the position is still positive at the real price. At +106 against -102 the pair implies 48.5% + 50.5% = 99.0% — thinner, but still locked. At +98 it would be a loss, and the right move is to cover for the smallest loss available rather than finish a plan that no longer exists. If only part of the stake was accepted, that is a book telling you it has cut your maximum — treat it as the first sign of a limit, not as a technical hiccup.
One leg is voided after the event. A player was scratched and one book voided the prop while the other graded it, or the price carried the mispricing badge and was voided under the book's palpable-error term.
You are holding a single live bet that you sized as half of a hedge — on the $1,000 plan that is up to $517 exposed. Nothing can be done after settlement, so the work is before it. Keep arbitrage on game lines, where void rules mostly agree across books. Treat player props as a place where two books can disagree about whether the bet happened at all. If the card carried a mispricing badge, that badge was the warning. The price is an error, errors get corrected, and the correction is often a void — on the leg you wanted, not the leg you hedged with.
Common mistakes
The specific ways people lose money with this tool.
Placing only the outlier price at one soft book, every time, and nothing else there.
The account is limited — the maximum stake drops to around $20 and usually never comes back. An account taking $500 a side three times a week at 2.4% produced about $36 a week in locked profit. It also held every signup and reload promotion you had not used, and every +EV price you would have taken there. The limit takes all of it at once, and there is no appeal.
Spread the volume across ten books rather than hammering two. Take the promotions at each — a promo-heavy account is one of the healthiest profiles a book sees. Keep ordinary mainline bets in the mix, and work the Live board if you have it. An in-play bet sits in a market repricing every few seconds — far harder to pattern-match than a pre-match price lifted the instant it moved.
Taking the rows badged "⚠ likely mispricing" because they carry the biggest percentages on the board.
Two ways. The book voids the leg under its error terms and you keep the hedge alone, which on a $1,000 plan is $517 riding on one side of a game you had no view about. And hitting book errors is the behaviour that gets an account limited fastest of all — faster than ordinary arbitrage, because the pattern is unmistakable.
Read the badge as a filter rather than a feature. Skip those rows. If you want one anyway, size it as a bet you can afford to have half of, not as a locked position, and expect the leg to be voided.
Rounding $483 and $517.02 to $500 and $500 because the plan's numbers look odd.
You no longer hold an arbitrage. $500 at +112 returns $1,060; $500 at -102 returns $990.20. The position wins $60 if the Over lands and loses $9.80 if it does not — a real bet on a total, taken by accident. Matched round pairs at two books are also among the clearest patterns a trading desk can search for.
Use the exact figures the stake boxes compute, and let the Bet at link copy them for you. Round stakes are good account hygiene on single bets, where the payout does not have to match anything; on a hedged pair the sizing is the product.