A cognitive bias is a systematic pattern in judgement observed under defined conditions. Foundational heuristics-and-biases research supports narrow tests of judgement, not a claim that every person displays every bias. In betting, use the labels to audit a recorded decision rather than diagnose a person or claim an advantage.
Bias map
| Bias | Narrow meaning | Betting audit question | Different from |
|---|---|---|---|
| Anchoring | An initial value influences a later estimate or adjustment | Was the forecast made before seeing a price or pundit estimate? | Simply comparing prices |
| Confirmation bias | Evidence search or testing favours a current hypothesis | What evidence would have counted against the selection? | Having a prior view |
| Gambler's fallacy | A recent independent sequence is treated as changing the next fixed chance | Did the process actually change, or only the sequence? | Regression to the mean |
| Illusion of control | Skill or action is treated as controlling a chance outcome beyond its real influence | Which part of the outcome can this action causally change? | Doing valid research |
| Loss aversion | Choices respond differently to gains and losses relative to a reference point | Would the decision change under an equivalent neutral frame? | Any dislike of losing |
| Outcome bias | A decision is judged differently because its outcome is known | Can the process be reviewed before revealing the result? | Updating from new pre-event evidence |
| Sunk-cost effect | Irrecoverable past resources influence a new decision | Would the same next action be chosen if the past cost were zero? | Considering future switching costs |
These concepts come from different studies and should not be collapsed into one explanation. The original research includes confirmation testing, illusion of control, prospect theory, outcome bias, and the sunk-cost effect. Each result is tied to its design, sample, and task.
One scenario, several possible errors
An illustrative bettor first sees decimal odds of 2.50, searches for team news supporting the selection, increases the stake after a previous loss, and later calls the decision correct because it won.
Several labels might be relevant, but the evidence is not interchangeable:
- Anchoring requires evidence that the first price influenced a later estimate.
- Confirmation bias requires a one-sided evidence-search or test.
- Sunk cost concerns the influence of an irrecoverable past cost on a new choice.
- Outcome bias concerns evaluating the prior decision after learning the win.
The scenario does not prove any internal cause. A better audit preserves timestamps, the forecast before the first price, search terms, disconfirming evidence considered, stake rule, and a process rating made before the result is revealed.
Four-column bias audit
Use one row per material decision:
| Before the decision | Possible distortion | Counter-check | Recorded result |
|---|---|---|---|
| Forecast made before price | First displayed price | Recalculate from frozen inputs | Was the estimate changed? |
| Current hypothesis | One-sided evidence search | Write a falsifier and search for it | Was adverse evidence recorded? |
| Current stake rule | Previous win or loss | Restore preset maximum loss | Did exposure change? |
| Process score hidden from result | Hindsight | Score process first | Did the rating change after reveal? |
The final column should record what happened, not assign a personality trait. Across many consistently recorded decisions, repeated failures can guide a simpler rule or a stop condition. They still do not establish a medical diagnosis or an economic edge.
Retrieval guide
Use the dedicated anchoring guide when the first number is central, the outcome-bias guide when a result has already been revealed, and the betting journal when the main problem is reconstructing what was known at the time.
Continue learning
- Next guide: Gambling Deposit Limits
- Related guide: GAMSTOP Self-Exclusion
Assumptions and limitations
Bias research describes patterns under specified conditions and does not imply universality. Similar observable behaviour can have different causes, and self-report can be incomplete. Bias awareness is not treatment, does not make gambling safe, and does not demonstrate that a forecast or price is favourable.
GambleAware
GamCare
Gambling Therapy
GAMSTOP