What is visible from outside
A reader can observe the operator, event, market, selection, decimal price, timestamp, overround, availability and later changes. A user may also observe whether a selected price changed before acceptance. Pinnacle's current bet-placement documentation is one first-party example of a product allowing price-change preferences; it does not reveal the cause of any particular move.
What usually remains private
| Question | Publicly verifiable by default? | Evidence needed |
|---|---|---|
| Which prices were displayed? | Yes, with timestamped capture | Feed or screenshot plus IDs |
| What was the market overround? | Yes, with a complete snapshot | All mutually exclusive prices |
| Which statistical model set the opener? | No | Operator methodology or reproducible system record |
| Did a named customer cause a move? | No | Direct first-party event evidence |
| Was a move driven by liability or new team news? | No | Documented causal evidence |
| Which stake could be accepted? | Only at execution | Accepted-bet record |
Do not turn an observed move into a story about “sharp money” or a trader's opinion. Multiple mechanisms can produce the same price path.
A defensible generic workflow
Without attributing it to every operator, the following is a local audit abstraction rather than a description of a named firm's current system. The need to preserve source differences is supported by bookmaker forecast research:
- Define the event and settlement rules.
- Estimate outcome probabilities using a model, market references, judgement or a combination.
- Transform those estimates into quoted prices and a commercial margin.
- Apply availability, limit and risk controls.
- Update quotes as information, reference prices or liabilities change.
- Suspend and settle under current rules.
The implementation and weighting at each stage are operator-specific and often unknown.
Why bookmakers can disagree
Differences can reflect data, models, margin allocation, timing, limits, product rules or commercial decisions. A historical study of 10,699 football matches found that odds-based forecast quality differed among bookmakers and leagues (Štrumbelj and Robnik-Šikonja, 2010). A broader 37-competition study also found source and market-size effects (Štrumbelj, 2014). Those results justify source-specific evaluation, not a permanent current ranking.
Worked observable audit
At 10:00, a 1X2 market shows 2.20, 3.40 and 3.30.
- Booksum = (1 / 2.20) + (1 / 3.40) + (1 / 3.30) = 1.05169.
- Displayed overround = 1.05169 - 1 = 0.05169, or 5.169%.
At 15:00 the home price is 2.05. Record the new complete market and executable depth. The observation supports “the home price shortened from 2.20 to 2.05 between captures.” It does not support “professionals backed the home team” without separate evidence.
Reader checklist
Apply these as SportSignals editorial controls; the source-comparison requirement follows the documented bookmaker differences in Štrumbelj and Robnik-Šikonja:
- Attribute product behaviour only to current first-party documentation.
- Store complete simultaneous markets when discussing margin.
- Use accepted prices for return calculations.
- Compare sources over a declared period with proper probability scores.
- Label causes unknown unless direct evidence exists.
Continue learning
- Next guide: Kelly Criterion for Value Bettors
- Related guide: Luck vs Edge in Betting
Assumptions and limitations
This page describes an audit framework, not the confidential process of any named operator. Models, limits and risk policies can change. For a clean observation protocol, use odds movement explained; for margin arithmetic, use overround and value.
GambleAware
GamCare
Gambling Therapy
GAMSTOP