You built the rule. You validated it, backtested it, read the assumptions, saved it with proof attached, and armed it.
Your phone buzzes. The condition is true.
This moment — between the notification and any action — is where rule-based trading usually breaks. Not in the logic, not in the backtest. In the ninety seconds after the alert, when a person decides whether to follow their own rule.
First, be clear what just happened
An alert is a notification that a condition became true. That is all.
- It is not an order.
- It is not a position.
- It did not reserve margin or move money.
- Nothing is pending anywhere.
Cash strategy arming evaluates closed bars and creates alerts only. Live strategy order execution is excluded from the first release. If a trade is going to exist, you are going to place it, through the order ticket, with route and margin review.
This boundary is deliberate and worth internalising rather than merely knowing. Smart alerts and the execution boundary covers it properly.
The practical consequence: you have time. Nothing is running away from you. The most damaging thing you can do is treat the alert as urgent and act inside five seconds.
Second, know that the bar has closed
Conditions evaluate on closed bars. The alert arrives after the bar completes, not at the instant price touched a level during it.
Two things follow.
The price has already moved. On daily bars, the signal fires after the close and you act at the next open. Some of the move is behind you. This is not a defect — it is what makes the signal reliable rather than a level that was touched and rejected. But it must be part of your expectation.
Do not chase. If the price has run well past your entry reference by the time you can act, the trade you are considering is not the trade the backtest evaluated. The disciplined response is to skip it. A missed signal costs nothing. A chased entry with a stop now far away costs real money.
Third, run the checks
Before placing anything, four questions. This takes under a minute and prevents most bad executions.
Is the instrument tradeable right now? Check the spread and the depth. A signal on an instrument you cannot transact in at a sensible price is not actionable, whatever the rule says.
Where does the stop go, and can I afford it? The stop comes from the timeframe's normal range, not from what you wish to risk. Calculate the distance, then size the position so the loss at that stop is within your per-trade limit. If the required size is uncomfortably small, take the small size — do not widen the stop to justify a bigger one.
Has anything material changed? A rule cannot see a results announcement tomorrow or a policy decision this afternoon. Check the market news. This is the one legitimate override, and it should be rare.
Do I already hold correlated exposure? Three signals in the same sector are one bet. Position-level correlation is covered in risk management in derivatives.
Fourth, place it or skip it — and record which
Then act. Either place the order with the stop calculated above, or decline it.
Write down which you did, and why. This log is the most valuable thing you will build, because it is the only way to find out whether your overrides help.
Most traders discover the same thing: the signals they skipped performed roughly like the ones they took, and their skipping was noise. A few discover a real filter. Neither is knowable without the record.
The four ways this goes wrong
Skipping after losses. Three losing signals in a row and the fourth gets ignored. It is invariably the one that works — not through malice, but because a strategy's results are a sequence and removing arbitrary entries from a sequence produces a different, worse strategy.
If you cannot follow the rule through its losing runs, the position is too large. Reduce size until you can. That is the actual fix.
Taking signals the rule did not give. The mirror image. The condition is nearly true, the chart looks compelling, and a position appears. You are now trading discretion with a strategy attached for reassurance, and you will never learn whether the rule worked.
Chasing. Covered above, and worth repeating because it is the most common. Late entry, distant stop, wrong risk. Skip instead.
Sizing by conviction. Doubling up on signals that feel strong. Your confidence is not a variable in the strategy and has no established relationship to outcomes. Fixed sizing, or sizing derived from stop distance. Not from feeling.
Choosing which alerts to receive
Alert fatigue is a real mechanism of failure. Twenty notifications a day and you stop reading them — including the one that mattered.
Set alerts you will actually respond to. If a signal arrives at a time you cannot act, receiving it achieves nothing except training you to ignore the channel. That is a timeframe problem, and choosing a timeframe is where to fix it.
Fewer, more meaningful alerts beat comprehensive coverage you have learned to dismiss.
When to override
Overrides should be rare and reasoned. Legitimate ones:
- A scheduled event the rule cannot see — results, a policy decision.
- The instrument is not sensibly tradeable — spread too wide, book too thin.
- The position would breach a limit you set in advance — correlation, daily loss cap, margin headroom.
Not legitimate: it feels wrong, the last two lost, the chart looks toppy, someone on the internet disagrees.
The distinction is simple. A good override cites a fact the rule could not access. A bad one cites a feeling the rule was specifically built to remove.
The sixty-second procedure
- Recognise the state. An alert fired. No order exists. There is no rush.
- Check the bar closed and the price has not run. If it has, skip.
- Check tradeability. Spread, depth.
- Calculate the stop, then the size. In that order, never reversed.
- Check for news and correlated exposure.
- Place it or skip it.
- Log which, and why.
Seven steps, under a minute, and it converts an alert into either a properly sized position or a deliberate decline.
The strategy finished its work when it notified you. What determines your results is whether the procedure after the buzz is as considered as the rule that produced it — and for most traders, it is not.