Good review output
A useful review produces a rule change, a watchlist change, a sizing change, or a decision to collect more paper-trading evidence before changing anything.
Workflow
Paper-trading risk review workflow
Use this workflow to review whether a paper-trading agent is following its rules, respecting limits, and producing decisions you can explain.
| Step | Question | Trading Boy surface |
|---|---|---|
| 1. Define limits | What is the maximum size, loss, frequency, and correlated exposure? | Persona rules and agent creation. |
| 2. Run paper mode | Did the agent act inside those limits? | Paper agent and Telegram workflow alerts. |
| 3. Review decisions | Can the decision be explained from context, thesis, and rules? | Decision history and journal review. |
| 4. Measure behavior | Are drawdown, sample size, and frequency acceptable for the thesis? | Leaderboard context and risk metrics. |
| 5. Refine rules | Should the rule change, or is the sample too small? | Daily review, behavioral analysis, and audit trail. |
Bad review output
A weak review jumps from one lucky or unlucky paper result to a live-capital conclusion. Trading Boy should keep the process evidence-first.
Safety boundary
This workflow is educational and operational. It is not investment advice and does not recommend any live trade.
What a risk review should prove
A paper-trading risk review should prove that the workflow can be explained. It does not need to prove that the trade was profitable. It needs to show whether the agent respected the written plan, whether the risk was sized before the decision, and whether the result changed anything meaningful about the rule.
Before a simulated entry, the review should confirm the planned size, the maximum paper loss, the invalidation level, and the reason the trade belongs in the agent persona. If those details are missing, the paper trade may still be recorded, but it should be tagged as incomplete process. That tag matters because it prevents a lucky outcome from hiding a weak workflow.
After the trade closes, the review should separate market movement from rule behavior. A loss caused by a valid invalidation is different from a loss caused by an oversized paper position, repeated entries, or an ignored stop. Trading Boy pages link this workflow to the position size calculator, maximum drawdown calculator, and journal pages so each review can point to the exact constraint that needs attention.
The best risk reviews produce a concrete next action. That action may be to keep collecting paper data, lower the allowed size, reduce trade frequency, tighten the setup filter, or retire a rule that keeps creating unreviewable decisions. If the review produces only a general feeling, the agent needs more structure.
Example risk review
Paper result: An AI paper trading agent records four simulated trades in one session. Two are profitable, one is flat, and one creates a paper loss that is larger than the intended single-trade risk.
Risk finding: The losing trade had a clear thesis, but the agent opened it after two similar positions were already active. The issue is not just the loss. The issue is correlated exposure and frequency. The review tags the session as a rule-fit problem even though the total session result was positive.
Next action: The trader adds a rule that blocks new paper entries when similar exposure is already open, then watches the next sample. The old sample remains useful because it explains why the rule changed.
Risk checks before entry
Confirm position size, invalidation, maximum paper loss, daily drawdown allowance, correlated exposure, event risk, and whether the setup belongs to the agent's written role.
Risk checks after exit
Compare planned risk with realized paper loss, review whether the agent followed the exit rule, and decide whether the rule needs a focused change or more sample size.
How to decide from the review
A risk review should end with a narrow decision. If the paper trade respected the plan, the best next action may be to keep collecting evidence. If the paper trade broke the plan, the next action should name the exact constraint that failed.
Do not let one review rewrite every part of the workflow. If the problem was correlated exposure, change the exposure rule. If the problem was late exits, change the exit rule. Keeping the decision narrow makes the next paper sample a cleaner test of whether the risk control improved.
Paper-trading risk review FAQ
What is a paper-trading risk review?
It is a structured check that asks whether a simulated trade or agent decision followed written size, drawdown, frequency, invalidation, and rule-fit limits.
When should I run a risk review?
Run one before entries, after exits, and whenever an agent changes behavior, increases frequency, or creates drawdown that needs explanation.
Does risk review make a strategy safe for live trading?
No. It improves review quality for paper trading, but it does not remove market risk, predict future returns, or provide financial advice.