The Probation Period a New Rule Serves

New rules are supposed to be adopted or rejected. Most of them deserve neither for a while, and the useful middle state is a rule that is recorded every session and acted on in none of them. The breakdown sitting on orb trading review consultoriainnova lists the stages in order, since a strategy for the opening range is easiest to damage in the weeks just after a promising idea arrives, and probation is how the idea gets tested without the account paying for it.

What Probation Actually Means

A professional examines financial graphs on a laptop, indicating market trends.

The candidate rule is written down with the same precision as a live rule, applied to every session on paper, and its outcome logged beside what the live system did. Nothing about position sizing or entries changes. At the end of the period there are two parallel records covering identical conditions, which is a far better comparison than anything a backtest of the same idea would produce.

Set the Length Before Starting

A person in a home office analyzing stock market data on a laptop with a coffee cup beside them.

Thirty occurrences of the rule firing is a reasonable minimum, and it may take two months to accumulate. Setting the length in advance prevents the version of this that always happens otherwise, where the rule is promoted early because its first six results were good. Six results are noise, and promoting on them converts a testing procedure into a decision made on a small sample size.

Log the Non Events

The sessions where the candidate rule would have blocked a trade matter as much as the ones where it would have taken one. A filter is judged mostly on what it removes, so the record has to include every entry it would have prevented and what that entry actually did. Recording only the trades it would have added produces a flattering and meaningless result.

Judge It on More Than the Total

A rule that improves expectancy while doubling the number of sessions with no trade at all has a real cost. So does one that improves the win rate by cutting winners short. The comparison should cover frequency, average result, worst outcome and how much extra attention the rule requires during the first hour, since a rule nobody can execute reliably under time pressure is not an improvement.

Promotion and Retirement

A rule that passes gets a version number and joins the live set, alone, with no other change in the same interval. A rule that fails goes into the archive with its full parallel record rather than being deleted, because the same idea will occur again in a year and the record is the only thing that will stop it costing another two months.