Reactor for the trader with a defined method

You can start from one part of the method — detection, the entry decision or position management — and connect the steps later. You can use ready-made prompt examples and adapt them to your approach.

Start from the way you examine an opportunity: what must be true, what strengthens the idea, what rules it out, and in which situations you prefer to wait.

In the instructions you can separate binding limits from considerations the Agent should weigh. For example, a risk limit can be fixed, while the quality of the entry is assessed through several data points.

You can set up a detection Agent that scans the market and returns the cases that match your method's conditions. Each result comes with the information needed for the next step.

You can save the instructions and run the scan at set times. That creates both a list of trading candidates and a foundation for collecting research data.

Start by collecting occurrences of the defined pattern. With the candle tools you can replay price movement after each signal: the move in the trade's favor and against it, the time to target, and the behavior after the entry point.

The research lets you examine the opportunity itself before adding an Agent that decides which cases to buy.

In a real example, a Mapper detected Bullish Engulfing every four hours. Checking the movement after the signals led to choosing a 3% stop-loss and a 4.5% profit target. Then a Buyer that buys all the signals was run, to collect results.

As of preparing this example, by the measurement provided, 800 trades closed with a 65% win rate. The win rate is one part of the evaluation, alongside the size of wins and losses, costs, and the period measured.

You can give a decision Agent the information from detection and instruct it to assess the quality of the trade: market context, entry conditions, risk, and factors that contradict the idea.

The Agent can decide to buy, wait or skip. You can compare different approaches or different models on the same cases.

Reactor lets you translate the trade plan into orders. You can set a single or staged entry, different allocations, loss protection and profit targets, depending on what the account and the asset support.

In the relevant execution path you can review a preview before the orders are sent.

You can give position management a role of its own. An Agent can check protections, consider partial profit-taking, lock in profit, or leave part of the position for further movement.

Using the trade records and price replay, you can research how management affected the result: what was taken, what stayed open, and what happened after the exit.

You can connect detection, decision, execution, management and reporting. Each step's output feeds the next, and each role gets its own instructions and permissions.

The records let you go back to a specific run and check which model worked, which tools were called, what came back, what failed, and how much time and money it took. Alongside that, you look at trading performance.