Your first thirty days running automation

Week one: verify, do not evaluate

Nothing meaningful about performance can be learned in week one. What can be learned is whether your installation is correct, whether your broker executes acceptably, and whether the trades being taken match the specification. Spend the week auditing mechanics and deliberately ignoring the balance.

Weeks two and three: leave it alone

This is the hardest instruction we give and the one most often disregarded. The urge to optimise arrives early, usually after the first losing trade, and it is almost always destructive. You do not yet have enough data to distinguish a problem from a Tuesday.

• —Do not change lot size after a loss. That is the impulse the entire architecture exists to prevent.
• —Do not close a position manually because it looks uncomfortable. You are overriding a tested rule with an untested feeling.
• —Do not add a second strategy to the same account to 'diversify'. You are compounding two sets of unknowns.
• —Do write down each urge and the date. Reviewing that list in month three is more educational than any course.

Week four: the first honest review

Now pull a full statement and review process rather than outcome. Was drawdown inside tolerance? Was slippage stable? Did the trade count sit in the expected range? A month is still a small sample for judging returns, but it is a perfectly good sample for judging whether the machine is running correctly.

Month one is about whether the machine is installed correctly. It is not about whether you are right.

What week one should and should not contain

The first week is an operations test, not a performance test. Nothing about profit or loss over five sessions carries information, and treating it as though it does is how new operators talk themselves into changing settings before the system has taken a representative sample of trades. The correct question for week one is whether the machinery works.

The week one checklist

• —Hosting stayed connected through every session without manual intervention.
• —Trades opened with the expected lot size for your configured risk.
• —Stops were placed at structural levels, not at broker minimums.
• —The terminal recovered correctly from at least one deliberate restart.

Building the habits that survive month three

The operators who do well after a year are not the ones who watched most closely in month one — they are the ones who established a light, repeatable routine and then left the system alone between checkpoints. Constant observation creates the illusion of control and reliably produces interventions that cost money.

A sustainable routine

The discipline is not in watching more carefully. It is in deciding in advance how little watching is enough.