Evidence Quality
Sports betting needs a standard for information—not merely more information.
Weak evidence can still look sophisticated.
A bettor can cite several statistics, repeat a respected analyst, reference a recent trend, and still build a poor decision case.
Every later stage of the BrownBagBets system depends on the quality of what enters it. Weak inputs can create a convincing but distorted indicator stack. That distortion can then produce inflated confidence, excessive price tolerance, and a cash allocation that appears disciplined while resting on a faulty foundation.
Evidence Quality does not eliminate uncertainty. It prevents unqualified information from pretending to reduce it.
Six dimensions determine how much influence information deserves.
Strong evidence does not need to be perfect. It needs to survive inspection.
Does it change the actual decision?
Identify what should change in the probability, price, confidence, uncertainty, or cash allocation because the information is true.
Can the claim be inspected?
Evaluate how the information was produced, whether the source is appropriate, and whether the underlying claim can be checked.
Does it describe the current environment?
Roles, injuries, tactics, lineups, and market conditions change. Evidence must still fit the present decision.
Does it add new information?
Multiple statistics may repeat the same underlying factor. Repetition should not be mistaken for confirmation.
Can the sample support the claim?
The sample should be representative and stable enough for the strength of the conclusion being drawn.
Is there a plausible mechanism?
Patterns are stronger when a credible mechanism connects the evidence to the expected outcome or market price.
What qualifies—and what does not.
Information can be accepted, downgraded, or rejected depending on how well it survives inspection.
Decision-connected evidence
Information directly tied to the market under review, supported by a credible source, current conditions, an appropriate sample, and a plausible mechanism.
Independent contribution
An input that adds meaningfully distinct information instead of repackaging the same underlying factor.
Interesting but irrelevant facts
Memorable history, rivalry records, broad trends, and television narratives that do not materially change the current decision.
Precision without mechanism
A model output, threshold, or probability estimate that appears authoritative while hiding weak assumptions or poor inputs.
Screen the information before building the case.
This is a hypothetical teaching example. It is not a historical BrownBagBets wager and should not be interpreted as a current recommendation.
Hypothetical pitcher-outs market
Qualified evidence
- A verified team report says the pitcher is returning from a minor arm issue under a monitored workload.
- His three starts since returning ended at 65, 72, and 78 pitches, consistent with the current usage report.
- The opposing lineup is patient, creating a plausible mechanism for a higher pitch count before 18 outs are recorded.
- The available market is 17.5 while comparable books are beginning to show 16.5, making the current number materially relevant.
Limited evidence
- The pitcher stayed under 17.5 in four of his last five starts. Relevant, but reduced because the sample is small and conditions changed.
- Projected left-handed hitters have performed well against his primary pitch type. Potentially useful, but conditional on the confirmed lineup and sample stability.
Rejected information
- The opponent won the previous night.
- The pitcher has a poor career record in day games.
- A social account calls it the strongest play of the slate.
Decision effect
The qualified evidence supports continued consideration at 17.5 outs. It does not justify participation at every number. A move to 16.5 or a materially worse price requires a new valuation, and any cash allocation remains constrained by workload and lineup uncertainty.
Weak evidence often enters through familiar shortcuts.
Mistaking detail for relevance
Ask exactly what part of the decision changes because the fact is true.
Treating correlation as independence
Identify the mechanism behind each indicator and reduce the weight of overlapping inputs.
Trusting freshness without reliability
A recent claim still needs a credible source and an inspectable basis.
Trusting sample size without mechanism
A large dataset can describe what happened without explaining why it should continue now.
Accepting consensus as proof
Agreement may reflect shared information—or merely shared assumptions.
Confusing outcome with evidence quality
A winning wager can contain poor evidence. A losing wager can contain strong evidence.
Information does not become evidence merely because it is available.
BrownBagBets records why an input is relevant, reliable, current, and sufficiently independent before allowing it to strengthen a position. Sample limitations and causal uncertainty must remain visible. Inputs that fail those standards are rejected, downgraded, or carried forward explicitly as uncertainty.
Evidence Quality is the first gate—not the complete decision.
Qualified evidence must next be matched to a recognized sport-specific indicator. Only applicable registered indicators may proceed into Indicator Stacking.
The public standard is reinforced by the Research Constitution and Research Standards. Return to the Pattern Literacy Library for the broader market-reading framework.
The outcome is evidence. It is not the only verdict.
Evidence Quality should be reviewed after the event without judging the original process solely by whether the wager won.
- Original source and availability time
- The claim being made
- Why it was considered relevant
- Whether it was independent
- The mechanism connecting it to the decision
- Sample or source limitations
- The influence it received
- What later evidence confirmed or weakened it