Three losing trades can make the next decision feel different, even when the written rules have not changed. You might enter early to recover faster, increase size or decide that the next valid signal will fail too. Each response changes what you are actually testing.
This Read the Reaction lesson asks a narrower question than whether a strategy is profitable: did you carry out the decision you had specified? Answering it helps identify what needs investigation. It cannot establish that a system deserves more capital.
Seykota’s lesson starts with the operator
In an interview hosted on his website, Ed Seykota recalls testing Donchian’s rules on brokerage mainframes in the punch-card era. The context is systematic technical trend following rooted in traditional futures markets, not a crypto strategy.
In his November 28, 2019 FAQ reply, he challenges the idea that a trading bot removes emotions: a drawdown may instead prompt its owner to tinker with it.
That observation gives us a useful starting point. An automated entry does not automate the owner’s choice to switch the system off, change its settings or override its size. Those choices belong in the review too.
Review execution separately from outcomes
Seykota’s June 18 and June 27, 2020 replies connect theoretical-versus-actual results with willingness to follow rules, and warn that disappointed expectations can lead to abandoning a system.
Our practical interpretation is to keep two questions separate: what did the trade earn or lose, and did the actions match the plan? A profitable deviation remains a deviation. A correctly executed loss remains a loss; calling its execution consistent does not make the strategy good.
A losing streak counts consecutive losing trades. A drawdown measures the fall from an earlier equity peak. They describe different aspects of a record, and neither alone diagnoses the cause of poor results. See backtests and live results for the measurement context.
Freeze the exercise rules before the next trade
Consider an imaginary trader reviewing the next opportunity after three consecutive losses. The following is an original paper-trading exercise, not Seykota’s rule set, an observed crypto incident or evidence of an edge.
For a fictional spot token, the trader has written these instructions before any of the outcomes below:
- Enter only after a completed one-hour candle closes above a previously recorded breakout level.
- Use the position quantity recorded in the plan; do not increase it to recover earlier losses.
- Follow the recorded exit instructions; do not move the protective stop farther away to avoid realizing a loss.
- If order acknowledgments and the position record disagree, pause new entries and reconcile the position under the emergency procedure.
These are selected rules for classifying behavior, not a complete tradable system. The level, size, exit logic and emergency procedure would need full specifications and testing. Our Turtle lesson explains why an entry rule alone is incomplete.
Four cases: the result does not rewrite the action
Treat these as four alternative scenarios for that next opportunity, not a sequence of trades or an estimate of outcome frequencies. Profit and loss mean the hypothetical result after costs.
- Rules followed; loss. The candle confirms, the planned quantity enters and the recorded exit executes. Record the loss and consistent execution. Review the strategy’s evidence separately; do not label the entry a violation merely because it lost.
- Rules followed; profit. The same requirements are respected and the exit produces a gain. Record consistent execution and the profit. One winner does not validate the method.
- Rules broken; profit. The trader buys before the candle closes, hoping to recover sooner, and later exits profitably. The premature entry still violates the timing rule. The gain cannot retroactively supply confirmation.
- Rules broken; loss. The candle confirms, but the trader doubles the planned quantity to recover previous losses. Record the sizing violation and actual loss separately. This outcome cannot fairly represent the unchanged system’s position sizing.
Before revealing an outcome in your own exercise, classify the action using only the rule and information available at the decision time. Then reveal the result. The classification should survive that reveal.
Following a system includes knowing when to pause
Consistency is not a reason to continue a defective or unaffordable process. Write operational pause conditions and research review criteria before a stressful episode. A missing order acknowledgment is a concrete execution problem; it does not require waiting for a longer losing streak.
Other review triggers can include a predeclared loss limit, persistent costs beyond the tested assumptions or evidence that live decisions differ from the tested implementation. Specify the measurement window and response in advance. There is no universal number of losses that proves a system has failed—or that a winner is due.
Pausing new entries does not automatically close existing positions. Their management needs its own procedure. Likewise, a planned stop is not a guaranteed fill or loss cap; the risk guide explains the distinction.
If you revise the method, record a new version and test the change. Keep results from different versions identifiable instead of presenting the combined history as one unchanged strategy.
Try a two-line decision review
Choose a completed trade and write:
- Process: the rule version, required action, actual action and evidence available then.
- Outcome: the result after costs, any execution discrepancy and whether a predefined pause or review condition was reached.
If a field is unknown, mark it unknown. Reconstructing missing information from the final chart can make an improvised decision look planned.
The aim is to identify the next question accurately: repair execution, investigate the strategy, reconsider affordable risk or follow the existing review procedure. Discipline supplies a clearer record for that investigation. It does not supply a profitable edge.
Educational material, not financial advice. All scenarios are hypothetical. Neither historical trading experience nor consistent rule-following guarantees future results.
