NCAA Basketball Indicator Framework
A basketball-specific research architecture for finding causal mechanisms, choosing the market that most directly expresses them, respecting exact price, and learning prospectively without pretending an early framework is already a finished model.
Sophisticated enough to investigate. Small enough to operate.
This framework is not a betting model, probability engine, rigid checklist or historical backtest dressed up as certainty. BrownBagBets is beginning with a small set of basketball mechanisms, preserving decisions prospectively and allowing repeated evidence to determine what deserves promotion, revision, merger or retirement.
Scan broadly for games, conditions or prices worth understanding.
Research only the basketball mechanisms actually relevant to the matchup.
Require causal, supported, current and expressible evidence. Challenge only when something material needs resolving.
State what should happen on the court before choosing a market.
Exact market, line and price determine whether the idea still has permission.
BET, WAIT or PASS. TRACK never means BET.
Freeze the pregame reasoning before the outcome can rewrite it.
Grade mechanism, expression, price, allocation and outcome separately.
BrownBagBets has prior NCAA Basketball portfolio history, but Phase 1 begins prospectively. Older results are not re-tagged after the fact to manufacture validation for a system that did not yet exist.
Count causal paths, not statistics.
A parent indicator exists because it represents a distinct way a basketball game can be separated, accelerated, suppressed or made volatile. Multiple measurements of the same reason remain one mechanism.
Functional Team Quality
The baseline layer. Opponent-adjusted offense and defense, quality of opposition, possession-level performance and schedule distortion establish how good the team appears to be before matchup-specific mechanisms are applied.
Possession Economy
Ball security and pressure sit beside rebounding and possession recovery. Turnovers and rebounds can remain separate causal paths inside the parent when their basketball effects are genuinely independent.
Shot Profile Creation & Prevention
Research emphasizes shot creation and denial rather than recent shooting results. Rim access, three-point quality, location profile and defensive interaction matter more than a short hot or cold percentage streak.
Interior Pressure & Foul Economy
Size matters only when it changes basketball function. Rim access, rim protection, foul drawing, defensive foul behavior and frontcourt depth can create points, free throws, lineup stress or margin.
Tempo & Expected Game Shape
Tempo is not simply season-average possessions. Transition, half-court dependence, pressure, offensive rebounding, late-game foul extension and expected competitiveness shape the possession environment.
Roster, Role & Rotation Integrity
Injuries, transfers, freshmen, creator availability, lineup changes, minutes concentration and evolving roles can invalidate otherwise useful statistics. Experience and continuity receive no automatic positive weight.
Scheme & Tactical Interaction
Zone, switching, drop coverage, ball-screen defense, press, transition emphasis, post doubles and offensive-rebounding philosophy matter only when the tactic is actually used, can be executed and has a plausible matchup consequence.
Context can matter without receiving another vote.
Phase 1 deliberately keeps familiar betting narratives outside the parent count unless they change an identifiable basketball function.
Home, travel, rest, altitude
Relevant when they alter shooting environment, conditioning, rotation, pace or another basketball mechanism. Not standalone reasons by default.
Rivalry, revenge, must-win, Senior Night
These can trigger investigation. They do not independently qualify a wager unless they create observable changes in role, pace, pressure, rotation or game behavior.
Strength of schedule
Used to interpret efficiency and raw performance rather than treated as a separate causal mechanism.
Market movement & price
Market information can change permission and research priority. It cannot invent the underlying basketball case.
A good basketball thesis does not automatically identify the right wager.
BrownBagBets states the court-level prediction first, then chooses the available market that captures it while requiring the fewest additional assumptions.
ML and spread are different claims.
A team can be more likely to win without having enough possession, shot-quality or closing advantage to justify laying a large spread.
More possessions do not automatically mean Over.
Totals require a view of both expected possession count and expected scoring efficiency.
Not simply half of the full-game handicap.
Starting-lineup edges, early rotation strength, late foul variance, depth and halftime adjustment risk can make the first half the cleaner expression.
Target that offense specifically.
A defensive advantage is not enough. Expected possessions, efficiency, turnovers, rebounding, transition, rim access, free throws and game state must explain why that team's scoring output should move.
Ask: What does this mechanism predict? Then ask: Which available market expresses that prediction while requiring the fewest additional assumptions?
Challenge when something meaningful needs resolving. Do not manufacture a countercase.
Phase 1 does not require a ritual opposing argument for every game. A dedicated challenge becomes necessary when material contradiction, uncertainty, price movement or expression conflict could change the decision.
| Trigger | Question | Possible consequence |
|---|---|---|
| Strong contradictory basketball evidence | Does it materially weaken or invalidate the primary mechanism? | Reduce confidence, WAIT or PASS |
| Expression is unclear | Does the evidence support winner, margin, scoring or a narrower first-half/team-total claim? | Change market |
| Roster uncertainty | Does the thesis survive the unknown role or availability? | WAIT, reduce allocation or PASS |
| Material line or price move | Is the same decision still justified at the new number? | Change expression or remove permission |
| Mechanisms point opposite directions | Which pathway is more likely to dominate this matchup? | Resolve, downgrade or PASS |
November data does not get February authority.
College basketball changes quickly. Phase 1 explicitly changes the evidence standard as current-season identity becomes more trustworthy.
Observe function before trusting rankings.
Heavily discount raw shooting percentages, tiny-sample turnover rates, early opponent-adjusted rankings and blowout-inflated scoring. Prior-year information remains useful only where players, roles or systems actually persist.
Transition from priors to current identity.
Increase current-season authority gradually. Audit opponent quality, non-D1 distortion, role stabilization, minutes concentration and whether shot and possession profiles are repeating.
Current-season evidence becomes primary.
Comparable opponents, repeated styles and stable rotations make matchup interaction and opponent-adjusted current evidence more useful.
Current team identity dominates.
Previous-season data becomes background unless the same basketball function is still directly relevant. Narrative still does not earn standalone weight.
The research card should be compact enough to use every time.
The working record preserves what mattered before tipoff without forcing every game through every parent or inventing fields that do not change the decision.
The framework earns permanence through prospective evidence.
The first objective is not to produce the prettiest record. It is to learn whether the seven parents identify useful mechanisms, whether the right games are rejected, and whether market expression and price discipline improve.
Did the expected basketball actually appear?
Outcome cannot answer this by itself. Review the causal pathway independently.
Did the selected market capture the thesis?
A good mechanism can be paired with the wrong spread, total, team total or game segment.
Was participation still justified at execution?
The sporting thesis can remain correct after the wager loses permission.
Merge, split, demote or retire only when evidence earns it.
No arbitrary sample size makes a parent permanent. Repeated distinct causal behavior does.
A win can contain a weak decision. A loss can contain a sound one. BrownBagBets grades mechanism, expression, price, allocation and outcome separately so the result cannot quietly rewrite the process.