Why they fell short
A candidate can appear in more than one failed-gate group. These groups overlap and must not be added together.
RESEARCH JOURNAL · NOT BET RECOMMENDATIONS
Near Misses
Fully audited sides, totals and player props that narrowly missed our BET thresholds. These are hypothetical research observations, not model-recommended picks or a suggestion to wager.
We record eligible entries before kickoff, preserve the original line and price, and track every outcome—including losses. Official BET results stay separate.
Prospective research record
What counts as close?
The first qualifying observation for each event, market and selection is retained. Refreshes cannot replace its price. Missing audits, stale quotes and failed price limits are excluded. Later promotion to an approved BET does not erase the original research entry; the two cohorts can overlap and must not be combined.
Still awaiting evidence
These analyses are incomplete, so we cannot call them near misses. They are excluded from the research results above.
Learning without hindsight
Research returns assume one hypothetical unit risked at the recorded price. They are not real wagers or proven profitability. Small samples, correlated games, model changes and testing many thresholds can mislead. We will use these observations to propose changes, then test those changes prospectively rather than retune the rules to past winners.