FAQ

VigBreak detects hundreds of signals, but the paper-trader applies filters before logging a bet. EV bets with decimal odds above 5.0 are excluded to reduce variance - they are real edges, just not paper-traded. Kelly staking also means very small edges receive near-zero stakes and may be skipped entirely.

The paper-trader is an automated bot that simulates placing a bet at the moment each qualifying signal fires, always before kickoff. It uses half-Kelly sizing on a 100-unit bankroll. No bets are backdated or edited.

A signal is an opportunity detected by our models: either a positive expected-value (EV) bet at a soft bookmaker, or an arbitrage opportunity across two or more bookmakers. Every signal is timestamped before the event starts.

Available: odds are still live and the event has not started. Odds Gone: the bookmaker moved or removed the price. Event Started: the event has kicked off. Completed: the event finished.

An EV bet is one where our model believes the true probability of an outcome is higher than the bookmaker's implied probability. Over a large sample, positive EV bets should return profit even if individual bets lose.

An arb covers all outcomes of an event across multiple bookmakers at odds that guarantee a profit regardless of the result. The profit % is locked in at the time of the signal.

All results are shown in units (u) rather than currency. One unit equals 1% of a 100-unit bankroll. This makes performance comparable regardless of individual stake size. For example, a +2.3u result means the same whether you bet £10 or £100 per unit.

The stats page only includes signals that were paper-traded. Signals filtered out (high odds, very low Kelly stake) are shown on the Signals page but not in the performance numbers. We don't place paper bets on high odds signals to avoid skewing the stats with longshot wins/losses that aren't representative of the bot's core edge. These are winning bets, but the variance is extreme and they are not the focus of our strategy, which is to find consistent edges at reasonable odds.

CLV measures whether the odds we bet at were better or worse than the final odds before kickoff. If we bet at 2.10 and the closing price was 1.95, our CLV is positive. Consistently beating the closing line is the strongest evidence of a genuine edge.

CLV profit = sum of (stake × CLV %), with CLV % being how much better the placed odds were than the closing odds, across all settled EV bets. It represents the theoretical profit implied by how well we beat the market, independent of win/loss variance. If CLV profit is much higher than actual P&L, we are running below expectation; if much lower, we are running above. This is the bot's EV.

The average return of each individual bet. Since our bets use kelly staking, we expect this number to be below our overall ROI. Higher confidence bets get larger stakes.

Arb staking is fundamentally different - you commit capital to all legs simultaneously to guarantee profit. Combining EV and ARB stakes would misrepresent how the bankroll is being used. EV stake gives a clean picture of Kelly-sized directional bets. We don't have to prove that our arbs work, the maths does that for us.

Every paper bet placed by the bot, with the odds at the time of placement, the result, and the P&L in units. ARB bets show the combined result for all legs of that arb. The bot will only place one bet per event, even if there are signals across multiple bookmakers.

For each settled EV bet, CLV shows how our placed odds compared to the closing odds - positive means we got better than the closing price, and placed a good bet.

Completed shows all settled bets (won, lost, voided) - everything that is no longer pending. This lets you look into specific bets and see the details of how they performed, including CLV.

The Event date filter narrows bets by when the event took place, not when the bet was logged. Useful for looking at a specific match day or week.

Player props (e.g. Over 22.5 points, Over 4.5 assists) have no reliable result feed we can verify automatically. Rather than skip the bet, we paper-trade it and record CLV - how much better our odds were than the closing price. CLV is the real performance signal for props: consistently beating the closing line means we have an edge, regardless of the individual outcome. The push result means the bet has no P&L impact on the portfolio.

For ARBs on sports without a score source (boxing, tennis, MMA, rugby) and for player prop ARBs, we cannot determine which leg actually won. We record Leg 1 as Won and Leg 2 as Lost by convention - the assignment is arbitrary. The total P&L is always correct and always equals the guaranteed profit locked in at signal time. A note is shown on any bet where this applies.

VigBreak is free to use during the beta period. A subscription will be introduced soon - we will give plenty of notice before any charges begin, so you will have time to decide whether you want to continue.

Our focus right now is on making sure the product works well before we start charging. Early users will be the first to know when pricing goes live.

Visit our contact page and we'll get back to you as soon as possible.