An event contract and a sportsbook moneyline can describe the same question in different dialects. One quotes a share price. The other quotes odds. Translation is easy. Proving the two tickets settle the same event is the hard part.
Start with the rulebook, not the calculator. If timing, cancellation, overtime, participant eligibility, or settlement source differs, the prices are not interchangeable no matter how clean the math looks.
A contract price is only a probability under declared terms
For a binary contract that pays one full unit on YES and nothing on NO, the executable YES price can be read as an implied probability before fees and market frictions. The word executable matters. A last trade, midpoint, bid, and ask can all show different values.
Use the side you could actually trade. A buyer compares with the ask. A seller compares with the bid. Then account for the venue’s current fee schedule and any settlement friction described in the rules.
Do not assume every displayed market follows the same payout convention. Confirm the contract value, collateral treatment, and resolution terms first.
Convert probability to odds without changing the event
Let p be the implied probability of YES and q be the remaining probability. Let base be the standard American-odds base.
favorite odds = -(base × p / q)
underdog odds = base × q / pThe formula changes the display, not the belief. Decimal odds are the reciprocal of the implied probability under the same simplified assumption. Fees and bid-ask spread still sit outside that display conversion.
Keep enough precision during the calculation and round only for presentation. Early rounding can turn a small apparent difference into a fake edge.
Devig the sportsbook before comparing
A two-sided sportsbook market usually embeds hold. Convert both quoted sides to implied probabilities, then apply one declared vig-removal method. Normalizing the sides to a complete probability mass is a common method; other methods may behave differently when the prices are asymmetric.
The important rule is consistency. Do not use one devig method when it favors the contract and another when it favors the sportsbook. Save the raw prices and the method beside the normalized result.
The vig and hold guide covers this step in more detail. Without it, a direct comparison can mistake sportsbook margin for disagreement about the event.
Match the settlement language line by line
Two markets can look identical while resolving differently. Check the named event, participant, start and end condition, overtime treatment, postponement rule, cancellation rule, official data source, and dispute process.
A sportsbook moneyline may include overtime while a contract uses regulation only. A player market may void on non-participation while another venue settles under a different appearance rule. A season contract may define qualification or vacancy in a way a casual label does not reveal.
When the terms differ, stop calling the price gap arbitrage. You are comparing related risks.
Probability does not map to a point spread by algebra
A moneyline probability can be converted to odds directly. A point spread needs a model of the scoring-margin distribution. That model is sport-specific and can vary by matchup and market regime.
Key numbers matter in low-scoring sports. Possession volume matters in higher-scoring sports. Draws matter in three-way markets. Run lines and puck lines need their own margin distributions. A universal lookup table is attractive because it is simple. It is also a good way to print unsupported precision.
Use a sourced, point-in-time calibration curve when one exists. Name the historical window, sample, market source, and validation method. Without that evidence, keep the comparison in implied-probability space.
Threshold contracts need exact boundary matching
A total posted at a half-step and a contract resolving at an integer threshold may differ on one exact score. That boundary can carry meaningful probability. The same issue appears in player milestones and election margins.
Write the event sets explicitly. Which outcomes win the sportsbook ticket? Which outcomes settle YES? The difference between those sets is the adjustment you must model. Do not patch it with a remembered “key-number percentage.”
The bid-ask spread can eat the headline gap
A contract screen may show a midpoint that nobody can trade. A sportsbook screen may reprice between selection and acceptance. Compare executable terms at the same timestamp, then include fees and size limits.
Liquidity matters because the best visible price may cover only a small order. A larger position can walk the book and change its own average price. Record the available size beside the quote when position size matters to the decision.
Route only after both prices share one probability frame
- Define the event. Confirm the tickets resolve the same outcomes.
- Capture executable quotes. Save bid, ask, sportsbook sides, timestamps, and available size.
- Normalize the sportsbook. Convert both sides and remove hold under one method.
- Apply contract costs. Include the current fee and spread under the intended action.
- Compare with the model. Use one calibrated probability estimate for both venues.
- Return no trade when incomplete. Missing rules or stale quotes are not invitations to interpolate.
A workflow in the bettor desk can keep the raw and normalized quotes together. It should expose the inputs and cure string when a comparison fails, not emit a venue recommendation from missing data.
Save both quotes for later review
After settlement, preserve the original contract and sportsbook snapshots. Track whether the executable market moved, whether the model was calibrated, and whether the rule match held. Do not grade the translation only by which ticket won.
The CLV guide supplies the discipline: entry first, reference later, exact market match throughout. A cross-venue log adds another column for rule compatibility.
Bottom line
Converting a binary price to odds is arithmetic. Comparing venues is market work. Match the event, use executable quotes, remove sportsbook hold consistently, include contract fees and spread, and refuse unsupported point-spread conversions.
The calculator should be the easy part. If it is doing more work than the rule comparison, you are probably translating the wrong thing.
Breakeven win rate at recorded American prices
Breakeven probability is calculated only from American prices that were actually captured in the odds-history table.
Expected value from graded outcomes
Expected-value cells render only when a verified source binds observed win outcomes to the price paid for the same bets.
