OddsIQ
Paper #11PlannedNBA

Component Agreement and ROI: When NBA Model Sub-Signals Stack, How Much Edge Compounds?

Across 3,708 graded picks, we score every bet by how many independent model components (efficiency, schedule, context) agreed on a side — and report ROI by stack depth.

Read it
$0

Free executive summary, key chart, key takeaway. Delivered to your inbox the day the paper drops.

Free forever. Weekly digest. Unsubscribe anytime.
Buy it
$7/paper

Full 15-20 page PDF, print-resolution charts, reproduction appendix. No subscription, one-time payment.

Available when published.

License it
$1K–3K/paper

For sportsbooks, hedge funds, media. Above plus citation rights, internal redistribution, and a custom analysis on your data.

Email research@oddsiq.co
Buyer
NBA bettors, model builders, sportsbook traders
Effort
2 weeks
Enterprise license
Books, model-building hedge funds — $1-3K

What's in the paper

  • ~15-20 page PDF, citation-ready
  • Full methodology, reproducible from any pick history with the same fields
  • Print-resolution charts (bucket analysis, cross-version scatter, regression face-off)
  • Robustness appendix — bootstrap CIs, sample-size convergence, version cross-validation
  • Glossary + references to academic literature

Data used

Every field that goes into the analysis. Open methodology — you can reproduce this paper from any pick history with these columns.

FieldTypeSourceWhat it means
pick_idBIGSERIALnba_predictions_backtestUnique pick identifier.
dateDATEnba_predictions_backtestGame date.
efficiency_edgeNUMERIC(8,5)nba_predictions_backtestComponent 1: offensive/defensive efficiency edge in points.
schedule_edgeNUMERIC(8,5)nba_predictions_backtestComponent 2: rest, travel, and B2B edge in points.
context_edgeNUMERIC(8,5)nba_predictions_backtestComponent 3: matchup-specific edge (lineup, injury, motivational).
compositeNUMERIC(8,5)nba_predictions_backtestSum of the three components — the model’s headline edge.
components_agreeingINTnba_predictions_backtestHow many components (0–3) point the same direction. The paper’s key cohort variable.
decisionTEXTnba_predictions_backtestBET / LEAN / LEAN+ / PASS — confidence label assigned at pick time.
pickTEXTnba_predictions_backtestTeam selected.
pick_oddsINTnba_predictions_backtestAmerican odds at entry.
correctBOOLEANnba_predictions_backtestDid the pick win?
units_returnedNUMERIC(7,3)nba_predictions_backtestNet units after grading.
roi_per_pickNUMERIC(6,4)derivedunits_returned / units_staked.