The Four Losses That Made Me Rebuild a System That Was Working
Four consecutive stops. All within the maximum acceptable loss. All correctly executed — the stops triggered at the right level, the exits were taken immediately, the journal entries were written. By every objective measure, the system had worked correctly on all four trades. The positions had not been qualified for entry, moved against me, and been exited within the defined risk parameters. That is the system functioning.
But four in a row felt different from one, or two, or even three. By the fourth stopped trade, the emotional weight of the sequence had convinced me that the entry criteria were too loose, that the scoring system needed recalibration, that the stop placement was systematically wrong. I spent two weekends rebuilding the scoring rubric, tightening the entry criteria, and adjusting the stop placement methodology. I made the system more restrictive and changed three specific parameters that had felt like the source of the losses.
The fifth trade was entered under the new criteria. It advanced 22% over seven weeks. So did the sixth. And the seventh. When I looked at whether the four stopped trades would have been excluded by the new criteria, the answer was no — all four would still have qualified. The streak had not been caused by a system flaw. It had been caused by the normal randomness that any probabilistic system produces. I had rebuilt a working system because four consecutive losses had felt like a pattern when they were actually just variance.
The new system performed roughly the same as the old one. The two weeks of rebuilding work had produced nothing of value except the psychological relief of feeling like I had done something in response to the losses.
The Diagnostic — Is This a System Problem or Normal Variance?
Review each stopped trade against the entry criteria: five-pillar score above the minimum threshold, six confirmation checks all passed, market environment GREEN or YELLOW, position size calculated correctly from the maximum loss budget. If any trade was entered outside the qualifying process — a boredom entry, a FOMO entry, a mid-week impulse — the losing streak includes at least one unqualified trade. That trade is not evidence of a system flaw. It is evidence of a process breakdown. The system did not fail. The system was bypassed.
Review each stop placement against the Support Level of the base. Was the stop just below the Support Level as defined? Or was it placed arbitrarily — at a round number, or moved after the entry because the original stop felt too tight? Stops that were not placed at the correct structural level are a process error, not a system error. The fix is not to change the stop methodology. It is to apply the existing methodology correctly on the next entry.
Check whether the Market Pulse at the time of each entry was GREEN, YELLOW, or RED. Multiple stopped trades entered during a period when the market environment was YELLOW or transitioning to RED are not evidence of a stock selection problem. They are evidence that the environment was unfavourable and the entries should not have been made at full size — or possibly at all. The market environment is the context for every individual trade. Four stops in a YELLOW-to-RED transition are expected. Four stops in a confirmed GREEN environment are worth reviewing more carefully.
If every stopped trade produced a loss within the maximum acceptable loss calculated before entry, the risk management system worked correctly. The trades lost money. The system did its job. The total drawdown from four consecutive maximum acceptable losses is a known, predictable outcome — it is the worst-case scenario that was calculated when the system was designed. If the losses exceeded the maximum acceptable loss on any trade, the stop was either not placed correctly or not honoured. That is a process execution problem, not a system design problem.
Imagine a fair coin — one with a genuine 50% chance of heads on every flip. Flip it twenty times and there will almost certainly be a sequence of four, five, or even six consecutive tails somewhere in those twenty flips. The sequence of consecutive tails is not evidence that the coin is biased. It is a natural, mathematically expected feature of any probabilistic sequence. An investor who flipped that coin twenty times, saw five consecutive tails, and concluded the coin was biased would be making a reasoning error. The correct response to five consecutive tails from a fair coin is to continue flipping, because the expected value of each subsequent flip has not changed. A trading system with a 60% win rate will produce streaks of consecutive losses. The length of those streaks is not evidence of system failure. It is a natural feature of any probabilistic system operating over a finite number of trials. The response to a losing streak is the diagnostic, not a system rebuild.
Same four losses. Same outcomes on the subsequent trades. The difference is two weeks of rebuilding work that produced nothing — and a damaged relationship with the original system that was already working correctly. For illustrative purposes only. Past performance does not guarantee future results.
The urge to change the system during a losing streak is the losing streak's most expensive cost. The losses themselves are within budget. The system change that follows them often is not.
What to Do Instead of Changing the System
Run the diagnostic. If all four questions above have acceptable answers — every trade was qualified, stops were correctly placed, environment was supportive, losses were within budget — the correct response is to continue the system unchanged. Record the streak in the journal. Note that four consecutive losses occurred and that the diagnostic confirmed system adherence. Return to the Friday review. The streak is over when it is over. Its end is not predictable, and the next trade's probability of success is not changed by how many consecutive losses preceded it.
If the diagnostic reveals a process error — an unqualified entry, a stop placed incorrectly, an entry made in a deteriorating environment — identify and correct the specific process error. Do not rebuild the system. Fix the execution of the specific criterion that was violated. One fix, precisely targeted at the actual error. Not a comprehensive system overhaul driven by the emotional weight of four losses in sequence.
Reducing position size temporarily after a losing streak — entering the next qualifying trade at 50% of the normal maximum loss rather than 100% — is a reasonable psychological adjustment that some investors find helpful. It does not improve the system's expected outcome, but it reduces the emotional stakes of the next trade at a moment when confidence is lower than usual. The reduction should be temporary and pre-defined: return to full sizing after the first profitable trade, not after an arbitrary number of weeks. Past performance does not guarantee future results.
→ Why Conviction Without a System Is Just Stubbornness
→ What to Do When Your Stop Is Hit
→ How to Build Patience as a Trading Skill
Every Friday — System Consistent. No Changes Between Issues.
The Friday Flash applies the same qualifying criteria every week — regardless of recent results. The system does not change because of a losing streak. Free. No card needed.
Send Me the Friday FlashFrequently Asked Questions
System changes are appropriate when a pattern of process errors has been identified through the journal — not through the emotional experience of a losing streak. If the journal review shows that over a full year, entries made in YELLOW environments consistently underperformed entries made in GREEN environments, that is a data-supported reason to raise the entry threshold for YELLOW conditions. If the journal shows that a specific type of base pattern consistently failed while other patterns succeeded, that is a data-supported reason to tighten the base quality criteria. Changes made from journal data after a sufficient sample size — at minimum thirty to fifty trades — are evidence-based improvements. Changes made after four consecutive losses are emotional responses. The timing is the signal. Data, after a large sample, justifies changes. Losses, in a small recent sequence, do not.
Mathematically, a system with a 60% win rate will produce a streak of four consecutive losses approximately 2.56% of the time in any given sequence of four trades. In a year of twenty-four trades, encountering at least one streak of four consecutive losses has a probability of roughly 45% — meaning nearly half of all years will include a four-trade losing streak even when the system is working correctly. A streak of five consecutive losses has a probability of approximately 1% per five-trade sequence. These sequences feel significant when experienced, but are entirely consistent with a system operating at its expected performance level. The question is never whether a losing streak indicates failure. The question is whether the trades in the streak were executed correctly. Past performance does not guarantee future results.
A brief pause to run the diagnostic — reviewing the four questions in this article against the journal records of the stopped trades — is appropriate and takes one to two hours. An extended pause of weeks or months in response to a losing streak that passed the diagnostic is not appropriate. The market does not wait for the investor's confidence to recover, and qualifying setups that arise during a self-imposed pause are missed. The investor who pauses for six weeks after a losing streak that was within normal variance returns to trading at the same point in the market cycle but has missed whatever opportunities arose during the pause. The diagnostic, completed and confirming system adherence, is the only response needed. Continue the Friday review. Continue the watchlist maintenance. Enter the next qualifying trade with the same process as the previous ones.
Four consecutive stops. Two weekends rebuilding a system that was working. The fifth trade under the new system advanced 22%. So did the sixth and seventh. The four trades I had attempted to fix by rebuilding would all have qualified under the original criteria as well. The streak had been variance. The rebuild had been a response to the feeling of the variance rather than to any actual system flaw.
The correct response — run the diagnostic, confirm system adherence, continue unchanged — would have cost two hours instead of two weekends. The outcome of the subsequent trades would have been identical. The relationship with the original system would have been maintained intact rather than damaged by the implication that it needed to be rebuilt.
A losing streak feels like a pattern. A diagnostic reveals whether it is one. Most of the time it is not. Most of the time it is four flips of a fair coin that happened to come up tails.
Amateurs change the system after a losing streak because the losses feel like evidence. The process-driven investor runs the diagnostic first — because evidence requires data, and four consecutive losses in a probabilistic system is not data. It is the normal distribution doing what normal distributions do.Every Friday — Same Criteria. Same Process. System Never Changes Mid-Streak.
The Friday Report applies the same qualifying process every week. The criteria do not change because of recent results. Five stocks. Every Friday conditions allow.
See How The Friday Report Works →