The Operating Standard
Principles establish responsibility. Standards make it observable.
The Research Constitution defines what BrownBagBets agrees to honor. These Standards define what qualifying research must document, preserve, disclose, review, and make visible.
A principle cannot govern work that leaves no evidence of its application.
It is not enough to claim that research considered uncertainty, respected price, preserved contradictory evidence, or reviewed the outcome honestly. A credible public system must create a record showing where those responsibilities were carried out.
These Standards convert the BrownBagBets Research Constitution into required publication controls. They establish the minimum information needed to evaluate a decision before the event, examine it after settlement, and understand how the institution changed its knowledge over time.
The purpose is not to make every publication longer. It is to ensure that brevity never becomes permission to omit the facts necessary for accountability.
Make decisions traceable.
Preserve the evidence, price, timing, allocation, uncertainty, and judgment attached to the original decision.
Make reviews comparable.
Apply consistent fields and review standards across winning, losing, revised, and unresolved research.
Make learning cumulative.
Convert isolated outcomes into evidence that can strengthen, narrow, challenge, or retire prior institutional beliefs.
The Constitution defines the law. The Standards define the required record.
The Constitution establishes durable principles. These Standards specify the operational fields, disclosures, checkpoints, classifications, and publication gates used to demonstrate that those principles were applied.
A publication may be accurate and still fail the Standards if it does not preserve enough evidence to be independently reviewed. Likewise, compliance with the Standards does not guarantee that the final conclusion will be correct.
Publication Classes
Different publications require different depths of evidence.
The same constitutional principles govern every BrownBagBets research asset, but the required record must be proportionate to the claim being made. A short market note should not require a full dossier. A governing pattern should never rely on a short market note.
Decision Record
A documented market position, price, allocation, timing, and settlement record.
- Selection and market
- Published price and allocation
- Decision timestamp or publication window
- Settlement and outcome
- Material correction or void status
Research Observation
A preserved fact, change, market response, or contextual event that may contribute to later research.
- Raw fact or observed change
- Source and observation time
- Initial interpretation
- Contradictory or missing evidence
- Current qualification state
Case Study
A structured examination of a specific event before and after its outcome became known.
- Pregame evidence sequence
- Decision and market context
- Outcome and settlement
- Post-event review
- Mechanism judgment and limitations
Pattern Entry
A governing public research asset describing a recurring mechanism, its required conditions, limitations, and current evidence state.
- Pattern definition and scope
- Mechanism and qualifying conditions
- Supporting and contradictory cases
- Status, confidence, and revision history
- Public-versus-protected boundary
Research Letter
Public synthesis that translates current research into durable learning without overstating evidence.
- Primary research question
- Supporting evidence classes
- Limitations and unresolved questions
- Links to underlying public assets
- Publication date and edition number
Research Revision
A visible institutional update explaining how and why prior knowledge changed.
- Prior belief or classification
- New evidence requiring change
- Updated conclusion
- Effective date and revision number
- Impact on future research or decisions
Required evidence scales with the authority of the claim.
BrownBagBets applies three publication tiers. A higher tier may incorporate assets from a lower tier, but lower-tier evidence cannot independently support a higher-tier institutional claim.
Documents what occurred, what was decided, and when it was observed or published.
Explains what the evidence may mean while preserving uncertainty and alternatives.
Establishes a durable pattern, rule, or revision supported by multiple evidence objects and formal review.
The authority of the conclusion may never exceed the quality of the evidence preserved beneath it.
Publication Classes
Different publications require different depths of evidence.
The same constitutional principles govern every BrownBagBets research asset, but the required record must be proportionate to the claim being made. A short market note should not require a full dossier. A governing pattern should never rely on a short market note.
Decision Record
A documented market position, price, allocation, timing, and settlement record.
- Selection and market
- Published price and allocation
- Decision timestamp or publication window
- Settlement and outcome
- Material correction or void status
Research Observation
A preserved fact, change, market response, or contextual event that may contribute to later research.
- Raw fact or observed change
- Source and observation time
- Initial interpretation
- Contradictory or missing evidence
- Current qualification state
Case Study
A structured examination of a specific event before and after its outcome became known.
- Pregame evidence sequence
- Decision and market context
- Outcome and settlement
- Post-event review
- Mechanism judgment and limitations
Pattern Entry
A governing public research asset describing a recurring mechanism, its required conditions, limitations, and current evidence state.
- Pattern definition and scope
- Mechanism and qualifying conditions
- Supporting and contradictory cases
- Status, confidence, and revision history
- Public-versus-protected boundary
Research Letter
Public synthesis that translates current research into durable learning without overstating evidence.
- Primary research question
- Supporting evidence classes
- Limitations and unresolved questions
- Links to underlying public assets
- Publication date and edition number
Research Revision
A visible institutional update explaining how and why prior knowledge changed.
- Prior belief or classification
- New evidence requiring change
- Updated conclusion
- Effective date and revision number
- Impact on future research or decisions
Required evidence scales with the authority of the claim.
BrownBagBets applies three publication tiers. A higher tier may incorporate assets from a lower tier, but lower-tier evidence cannot independently support a higher-tier institutional claim.
Documents what occurred, what was decided, and when it was observed or published.
Explains what the evidence may mean while preserving uncertainty and alternatives.
Establishes a durable pattern, rule, or revision supported by multiple evidence objects and formal review.
The authority of the conclusion may never exceed the quality of the evidence preserved beneath it.
Required Evidence Fields
Every conclusion must preserve the evidence beneath it.
Qualifying research must retain enough information to show what was observed, where it came from, when it was known, what remained uncertain, and how the evidence became a judgment.
Evidence Standard
The record should make it possible to reconstruct the decision without reconstructing the story.
The purpose of evidence fields is not to create administrative volume. It is to prevent later memory, outcome knowledge, or promotional framing from replacing the information available when the original conclusion was formed.
Observation or Claim
The underlying fact, market change, role change, statistical condition, or research claim being preserved.
- State the observation without promotional language.
- Separate the raw fact from what it may imply.
- Use precise teams, markets, dates, and participants.
- Preserve the original wording when materially relevant.
Source and Provenance
The origin of the information and the chain through which it entered the research system.
- Identify the primary or secondary source.
- Preserve links, screenshots, files, or source references.
- Distinguish direct evidence from aggregated reporting.
- Note when multiple outlets repeat the same original source.
Time Known
The date, time, or reasonably bounded window in which the information became available to BrownBagBets.
- Preserve exact timestamps where available.
- Use a bounded publication window when exact time is unknown.
- Separate event time from observation time.
- Do not present post-event knowledge as pre-event evidence.
Market State
The price, direction, range, consensus, or liquidity context relevant to the research decision.
- Record the BrownBagBets entry price separately.
- Distinguish consensus from best available price.
- Preserve meaningful movement and market direction.
- Note whether the price was stale, limited, or widely available.
Interpretation
The current explanation of what the evidence may mean and why it matters to the research question.
- State the proposed relationship or mechanism.
- Avoid claiming causation from sequence alone.
- Identify reasonable alternative explanations.
- Keep interpretation separate from the final decision.
Uncertainty and Missing Evidence
Information that remains unavailable, unconfirmed, incomplete, or too weak to support a stronger conclusion.
- Identify missing confirmations or unresolved variables.
- State when sample size is too small.
- Note when data quality or timing limits confidence.
- Preserve unresolved questions for later review.
Contradictory Evidence
Material information that weakens, challenges, narrows, or potentially invalidates the primary interpretation.
- Record evidence supporting the opposing view.
- Preserve model disagreement and market disagreement.
- Do not remove contrary evidence after a favorable result.
- Explain whether it changes confidence, price, or classification.
Qualification State
The current status of the evidence object inside the research system.
- Unverified, observed, supported, challenged, or retired.
- Identify whether the evidence is pregame or postgame.
- State whether the case remains open or complete.
- Connect the evidence to any pattern, case, or revision.
Every statement must be classified.
Classification prevents a plausible interpretation from quietly becoming a verified fact as it moves through the research system.
Confirmed by a reliable source, official record, preserved market screen, or observable event.
A conclusion drawn from evidence but not independently verified as fact.
The institutional assessment reached after weighing evidence, alternatives, price, and uncertainty.
Information that is missing, unavailable, contradictory, or too weak to classify more strongly.
No qualifying item may be published without a reconstructable core record.
- What was observed or decided
- When it was known or published
- Where the information came from
- What interpretation was applied
- What uncertainty remained
- What evidence challenged the conclusion
A missing field must be disclosed, not silently invented.
Research may proceed with incomplete information when the limitation is visible and the claim remains proportionate. It may not proceed by treating unavailable evidence as though it were favorable evidence.
- Do not infer confirmed lineups from projections.
- Do not label public money as sharp without support.
- Do not claim causation from line movement alone.
- Do not describe consensus as the best available market.
A conclusion may be concise. The evidence record may not be invisible.
Settlement and Review Standards
The outcome closes the wager. It does not complete the research.
Every qualifying decision must be settled accurately and then reviewed against the conditions that existed when it was made. Outcome, closing price, decision quality, and research learning must remain separate fields.
Review Standard
The result tells us what happened. The review asks what the decision deserved.
A winning outcome may conceal a weak decision. A losing outcome may emerge from a disciplined process. Settlement records the event. Review evaluates the evidence, price, allocation, and mechanism without allowing the final score to rewrite the original record.
Record the Official Settlement
The final result must be entered according to the governing market rules rather than general impressions of how the event ended.
- Win, loss, push, void, partial settlement, or postponed
- Official score and relevant market result
- Settlement date and governing sportsbook rule
- Any disputed or corrected settlement
Preserve the Closing Market
Closing context should be recorded whenever reasonably available so that price quality can be evaluated independently of outcome.
- Closing consensus price
- Best widely available closing price
- Direction and magnitude of meaningful movement
- Availability or liquidity limitations
Separate CLV From Outcome
Closing-line value is a pricing observation, not proof that the underlying prediction was correct or that the research mechanism was valid.
- Record whether the entry beat consensus close.
- Record whether it beat the best available close.
- Do not describe all favorable movement as validation.
- Do not treat unfavorable movement as automatic disproof.
Identify the Failure Mode
When the decision loses or the thesis weakens, the review should identify where the process may have failed.
- Evidence failure
- Interpretation failure
- Price or valuation failure
- Allocation or portfolio failure
- Unavoidable outcome variance
Reassess the Mechanism
The post-event record should examine whether the proposed causal explanation was supported, weakened, or left unresolved.
- What occurred as expected
- What occurred for a different reason
- What evidence remains ambiguous
- What should be tested in future cases
Assign a Process Judgment
Each completed case should receive a process classification that is distinct from its financial result.
- Sound decision, favorable outcome
- Sound decision, unfavorable outcome
- Weak decision, favorable outcome
- Weak decision, unfavorable outcome
- Insufficient evidence to classify
The review must follow a fixed sequence.
A stable sequence reduces hindsight bias by forcing the institution to examine the original record before explaining the outcome.
Preserve the published thesis, price, allocation, and evidence before interpretation begins.
Enter the official result and any governing settlement notes.
Add consensus, best price, movement, and availability where possible.
Review evidence, valuation, allocation, and known uncertainty.
Support, challenge, narrow, or leave the relevant pattern unresolved.
Would the decision still be defensible without knowing the result?
The answer should be based on the original evidence, available price, allocation, and uncertainty—not on whether the event ended favorably.
- Was the evidence available at publication?
- Was the interpretation proportionate?
- Was the price inside the approved range?
- Was the allocation consistent with risk?
The outcome may not be used to invent confidence that did not exist.
Post-event explanation must not convert incomplete evidence, unresolved assumptions, or uncertain mechanisms into claims of inevitability.
- No “we knew it all along” reconstruction.
- No removal of contradictory evidence after a win.
- No dismissal of sound process after a loss.
- No selective preservation of favorable cases.
Winning and losing are settlement states. Learning requires a separate judgment.
Correction, Revision, and Archival Standards
Research must be allowed to change without allowing the record to disappear.
BrownBagBets distinguishes between correcting an error, revising a conclusion, and archiving a superseded asset. Each action serves a different purpose and requires its own visible record.
Governing Revision Principle
Knowledge may change. History may not be quietly rewritten.
A credible research system must improve when stronger evidence arrives. It must also preserve enough of the prior record for readers to understand what changed, when it changed, and why the institution reached a different conclusion.
Correction
A correction addresses a factual, clerical, pricing, attribution, timestamp, or settlement error in the published record.
- Identify the incorrect information.
- State the corrected information.
- Display the correction date.
- Explain whether the conclusion was affected.
- Preserve material prior language when practical.
Research Revision
A revision changes the institutional interpretation, confidence, status, scope, or mechanism because new evidence altered the research judgment.
- State the prior belief or classification.
- Identify the new evidence.
- Explain why the prior view changed.
- Publish the updated conclusion.
- Assign an effective date and revision number.
Archive
Archival preserves a prior asset that is no longer current, active, or governing while keeping it available for historical review.
- Mark the asset as archived or superseded.
- State the reason for archival.
- Link to the current replacement when one exists.
- Preserve original publication and revision dates.
- Prevent archived material from appearing current.
Every material change should leave an audit trail.
The change record should be clear enough for a reader to understand the prior state, the new state, and the evidence responsible for the transition.
Determine whether the issue requires correction, revision, or archival.
Preserve the material state of the asset before updating it.
Record the error, new evidence, or governance reason requiring change.
Make the corrected or revised information visible and current.
Link the current asset to the prior version or revision record.
Material changes must remain discoverable.
Corrections and revisions should appear on the affected asset or within a clearly linked change record.
- Version or revision number
- Effective date
- Summary of what changed
- Reason for the change
- Impact on the prior conclusion
Silence may not be used as a substitute for correction.
BrownBagBets may not erase a material error, revise a conclusion, or remove an unfavorable record without leaving evidence that the prior state existed.
- No silent changes to published prices or allocations.
- No deletion of losing decisions to improve appearance.
- No replacement of original language without notice.
- No backdating of revisions.
- No presentation of archived research as current.
Revision is evidence of learning. Invisible revision is evidence of evasion.
Compliance and Publication Gate
Research becomes public only after the record is ready.
The publication gate is the final operational control. It does not determine whether a conclusion will win. It determines whether the conclusion has been documented, bounded, priced, allocated, and communicated in a form the institution is prepared to defend.
Final Publication Checklist
A qualifying research asset must pass every applicable standard before release.
A missing requirement may result in a hold, a reduced claim, a lower publication class, or a decision not to publish.
The asset has been classified as a decision record, observation, case study, pattern entry, research letter, or revision.
The authority of the language does not exceed the depth, independence, or quality of the preserved evidence.
Material sources, timestamps, market records, files, and provenance are documented or visibly marked unavailable.
Verified facts, reasonable inferences, research judgments, and unknowns remain distinguishable.
Material evidence supporting an alternative conclusion has not been omitted or minimized.
Missing confirmations, weak samples, model limitations, and unresolved variables are stated where relevant.
The published price remains available or the acceptable range and limitation are stated clearly.
Exposure is inside the bankroll framework and accounts for portfolio concentration and information risk.
The publication avoids certainty claims, reckless promotion, and language that presents probability as inevitability.
The asset has a settlement, revision, archival, or future review pathway appropriate to its publication class.
Publish
The applicable fields are complete, the claim is proportionate, the price and allocation are approved, and the language meets the governing standards.
Hold or Reduce
The asset may require additional evidence, a narrower claim, a lower publication class, a revised price range, or reduced allocation.
Do Not Publish
The evidence is insufficient, the claim is overstated, the market is no longer actionable, or publication would violate the Constitution or these Standards.
The final standard is not confidence. It is whether the work remains defensible after certainty disappears.
Evidence before publication · accountability after release