Why is a winning streak not proof of skill?

Because streaks occur on their own at any positive expectancy and are therefore a property of random sequences. An example with round numbers: at a hit rate of 50 percent, the probability that five consecutive trades all win is one in thirty two. Across a hundred trades such a run is therefore not remarkable but expected, and at a higher hit rate longer runs occur correspondingly more often. It follows that a good week is entirely compatible with unchanged skill. The better test is the reverse: anyone who does not accept a losing streak as proof of incompetence cannot count a winning streak as proof of the opposite, since both are the same effect with the sign flipped.

How do I spot overconfidence in my journal?

Through three traces, all calculable from existing data. First position size: take risk per trade as a percentage of your account and plot it as a time series across your trades, because a rising percentage after good stretches at an unchanged rule is the clearest trace. Measuring against the account matters, since a growing account permits larger amounts under the same rule. Second, the time between closing one trade and opening the next, compared after wins against after losses. Third, the share of rule breaking trades over time together with their results, which is the most important of the three and also the one most rarely tracked.

What helps against position sizes creeping upward?

Three measures, all of which take position size out of the moment. First, position size as a formula rather than a decision, meaning risk amount divided by stop distance gives the quantity, because with the route fixed there is no point at which a judgement can slip in. Second, updating the reference figure only on fixed dates: the account balance you derive your risk from gets rolled forward monthly or quarterly rather than after every good week. That is the single most effective measure, because it interrupts the reinforcing mechanism directly. Third, a cap on increases per update, so one strong stretch cannot permanently shift your size.

When may I increase my position size?

When your metrics have improved across enough trades, not when your recent trades went well. This test works: has your expectancy improved across at least thirty to fifty trades under unchanged rules, and has the share of rule breaking trades not risen in that time? With both conditions met, an increase is justified. Timing matters just as much: the decision has to happen as planned rather than following a streak. An increase on the day after your fifth win is the wrong sequence regardless of the numbers, because it establishes exactly the link between success and size that you are trying to avoid.

Overconfidence in Trading: Three Traces in Your Journal

Q: Why is a rule breaking trade that wins especially dangerous?

Because it confirms the wrong lesson, and does so with the most persuasive argument available. It appears to demonstrate that the rule was dispensable. The effect is amplified by the fact that rule breaking trades win more often than usual in friendly market phases, because almost everything works. The deviation therefore gets rewarded precisely when it costs least, and the bill only arrives once the environment turns. The countermeasure is a recording rule rather than a matter of will: mark every trade for rule compliance independently of its outcome and analyse that column separately, because without the separation a winning breach disappears into the total.

In many journals the most expensive stretches sit conspicuously close behind the best ones. That is not coincidence but a pattern with three measurable traces: creeping position size, shorter gaps before the next entry, and a rising share of rule breaking trades that happen to be winning. Unlike tilt or FOMO this pattern has no unpleasant trigger but a pleasant one, and that is exactly what makes it invisible.

Why is a winning streak not proof of anything?

Because streaks occur on their own at any positive expectancy. They are not a distinction but a property of random sequences, and the arithmetic is short.

An example with round numbers: at a hit rate of 50 percent, the probability that five consecutive trades all win is one in thirty two. Across a hundred trades such a run is therefore not remarkable but expected. At a higher hit rate, longer runs occur correspondingly more often.

A sober conclusion follows: a good week is entirely compatible with unchanged skill. It contains no information about whether you have improved, and that information gets extracted from it regardless.

The reverse holds too and makes the better test: if you do not accept a losing streak as proof of incompetence, a winning streak cannot count as proof of the opposite. Both are the same effect with the sign flipped.

Which three traces does your journal show?

All three can be calculated from existing data, provided timestamps and risk per trade are recorded.

Position size creeps upward

Take risk per trade as a percentage of your account and plot it as a time series across your trades. If it rises after good stretches while your rule is unchanged, you have the clearest trace available.

What matters is measuring against the account rather than in currency. A growing account permits larger amounts at an unchanged rule, and that is arithmetic rather than a pattern. Only a rising percentage is notable.

The gap before the next entry shortens

Measure the time between closing one trade and opening the next, and compare that gap after wins against the gap after losses.

This figure needs only the timestamp you already have. If gaps after wins are systematically shorter, you are not waiting for a signal but for an opportunity.

Rule breaking trades increase and win

The third trace is the most important. Plot the share of rule breaking trades over time and put their results alongside. If the share rises in good stretches while those trades are also winning, you have the most dangerous constellation there is.

Why is the third trace the most dangerous?

Because a rule breaking trade that wins confirms the wrong lesson. It appears to demonstrate that the rule was dispensable, and it does so with the most persuasive argument available, namely a profit.

In a friendly market phase rule breaking trades win more often than usual, because almost everything works. The deviation therefore gets rewarded precisely when it costs least, and the bill only arrives once the environment turns.

The countermeasure is a recording rule rather than a matter of will: mark every trade for rule compliance independently of its outcome, and analyse that column separately. Without the separation, a winning rule breach disappears into the total. How to set up such fields without over engineering them is covered under loss aversion, where the same separation is needed in the opposite direction.

How does this differ from overtrading and tilt?

Through the trigger and through duration. Tilt reacts to a loss and fades. Overtrading is a permanent state with no trigger. Overconfidence reacts to success and fades too, it simply does not feel like a problem.

It follows that the usual countermeasures do not apply here. A mandatory break after a losing streak, which works against tilt, never triggers during a winning streak. A weekly cap on trades does bite, but it addresses count rather than size.

What is missing is a rule that engages after good stretches. Almost nobody builds one, because good stretches do not feel like they call for a rule.

How much of your success belongs to the environment?

Some of it, and that share cannot be read off your own result. In a broadly rising phase many setups work at once, poor ones included, and your account does not distinguish between a good trade and a friendly market.

That is where the real confusion comes from. It is not chance being mistaken for skill but environment, and that case is harder to spot, because the environment genuinely is a cause, just not yours.

One field makes it checkable: record per trade which market environment it took place in, from a short fixed pick list, assigned before entry. After enough trades you can then split your result by environment.

If your result drops sharply in weak environments while looking good in strong ones, the good stretch describes the market more than it describes you. That is not bad news but the reason not to size up further during strong phases.

Which countermeasures actually work?

Three, and all three take position size out of the moment.

Position size as a formula rather than a decision. Risk amount divided by stop distance gives the quantity. With that route fixed, there is no point at which a judgement can slip in.
Update the reference figure only on fixed dates. The account balance you derive risk per trade from gets rolled forward monthly or quarterly rather than after every good week. This is the single most effective measure, because it interrupts the reinforcing mechanism directly.
A cap on increases per update. Even when the account has grown sharply, the reference figure rises only by a share fixed in advance. That stops one strong stretch from permanently shifting your size.

Order helps alongside them: the reference figure gets adjusted outside trading hours, noted with a date, and never on a day when a signal is currently on the screen.

When may position size legitimately rise?

When your metrics have improved across enough trades, not when your recent trades went well. The difference lies in the sample and in when the decision is made.

This test works: has your expectancy improved across at least thirty to fifty trades under unchanged rules, and has the share of rule breaking trades not risen in that time? With both conditions met, an increase is justified.

What matters just as much is that the decision happens as planned rather than following a streak. An increase on the day after your fifth win is the wrong sequence regardless of the numbers, because it establishes the link between success and size that you are trying to avoid.

How do you check this in hindsight?

Through a single analysis you can set up in minutes. Sort your trading weeks by result and look at the following week in each case.

Then compare two figures: the average result of the following week after your best weeks and after your average ones. If it sits systematically lower after the best weeks, you have the effect in your own numbers and need no further argument.

Beware the sample: weeks are few, and few weeks are noise. This analysis only becomes meaningful after a year or more, and until then the three traces above are the more reliable indicators. A properly kept trading journal supplies all four figures without extra effort once risk per trade and rule compliance are recorded.

What is the core of it?

That success and skill are indistinguishable in day to day trading, at least across a small number of trades. This cluster otherwise deals with reactions to losses, because those hurt and therefore get noticed. This reaction does not hurt, and it only surfaces in the numbers once the enlarged position meets a poorer stretch. The groundwork for separating outcome from decision sits in trading psychology and discipline.

Overconfidence After a Winning Streak: Trading's Priciest Moment