Glossary term
Gate
Definition
A process check that can reject an idea before it becomes a fill — hours, stale data, cash floor, daily-loss brake, position cap. About process, not whether a name is “good.”
How it works
A gate is a process check that can stop an idea before it becomes a fill, whatever the analysis concluded. The course teaches eight: a kill switch, market hours, a confidence floor, a position cap, an order-rate cap, a cash floor, a daily-loss brake and data freshness. Gates are combined with AND, so every one must be quiet for anything to proceed.
Most gates are about circumstances and the book, not about a security. Hours, freshness and rate are about when and how often; the cap and the cash floor are about the shape of the book; the loss brake is about the recent past. Only the confidence floor touches the analysis, and it asks whether enough was said rather than what was said. A gate that fires says the process stops here for a reason that would apply to any name.
Worked example
Because gates multiply, a working system can refuse often. If each of the eight is independently quiet 95% of the time, all eight are quiet 0.95⁸ = 0.663 of the time. Refusals occur about 33.7% of the time, roughly 85 of 252 trading days.
If each gate is quiet 99% of the time, 0.99⁸ = 0.923, so something refuses on about 7.7% of days, around 19. At 90% each, 0.90⁸ = 0.43, and the process refuses more often than it proceeds with every component working correctly. In practice gates are not independent, so refusals cluster on bad days.
Hypothetical, for education only. Not a recommendation and not a prediction.
Common mistake
The mistake is reading a refusal as a verdict on the name. A failed cap is a size rule, a stale tape is a data problem, and neither is a grade. A second error is overriding a gate on the occasion that feels special: those occasions cluster where conviction is highest and estimates are most likely inflated, so a rule that steps aside then binds only when it is not needed.