How Do I Know If I'm a Good Trader?

Why can't the account balance answer this?

Because a run of results is a sample, and a sample small enough to accumulate in a year is mostly noise. The width of the uncertainty is computable, which makes this a question with an arithmetic answer rather than a philosophical one.

The 95% confidence interval on a win rate is p ± 1.96 x sqrt(p x (1-p) / n). Run it at a true 55% win rate:

Trades	95% interval on the win rate	Contains a coin flip?
30	37.2% – 72.8%	Yes
100	45.2% – 64.8%	Yes
200	48.1% – 61.9%	Yes
381	50.0% – 60.0%	Only just not

So 381 trades is where a genuine five-percentage-point edge stops being confusable with luck. At two trades a session that is about three-quarters of a trading year; at two a week it is closer to four years.

The uncomfortable implication is not that you are bad. It is that for most of the first year, the data cannot tell you either way, and any confidence you have in that period came from somewhere other than the evidence.

It is worth knowing the base rate you are measuring against. Using the complete trading history of every investor in Taiwan, the aggregate individual portfolio suffered an annual performance penalty of 3.8 percentage points, with losses equal to 2.2% of the country's GDP.^[1] A trader who is genuinely flat is already unusual.

What should you measure instead?

Process, because process converges far faster than outcome. These are all measurable inside 30 trades, which is where you actually are:

• Plan adherence. In what fraction of trades was the stop where you said it would be when you entered? This is a percentage you can compute from your own record, and it does not need a large sample to be meaningful — it is a count, not an estimate.
• Expectancy, not win rate. (win% x avgWin) - (loss% x avgLoss). A 40% win rate at 2R is 0.4 x 2 - 0.6 x 1 = 0.20R per trade; a 70% win rate at 0.4R is 0.7 x 0.4 - 0.3 x 1 = -0.02R. The higher win rate is the losing method.
• Whether the trades were recorded before the fact. A record written after the close cannot distinguish a plan you followed from a plan you reconstructed, and your memory will side with the version that reads better.
• Size discipline. The ratio of your largest risk to your median risk. If it is above about 3, one decision is carrying more weight than your edge is.

What does a good trader actually look like at 50 trades?

Not profitable, necessarily. Consistent in a specific, checkable way: the stops sat where they were stated, the sizes clustered rather than spiked, the losers were taken at the stated level, and the record contains all of them. That description can be true of a trader who is down for the period, and it is a much better predictor than the balance is.

Conversely, a trader up 40% over 30 trades with three of those trades doing all the work and no stated stops has produced no evidence about themselves at all. At a true 55% win rate the 30-trade interval runs from 37.2% to 72.8%; a 70% result sits inside it comfortably.

How does a record change the answer?

It makes the process metrics computable by somebody other than you, which is the part memory cannot supply. Adherence measured against a plan you can still edit is adherence to whatever the plan says now.

Of Taiwanese day traders followed from 1992 to 2006, under 1% predictably earned positive abnormal returns net of fees.^[2] And in the published finance literature, Harvey, Liu and Zhu argued that so many predictors had been tested that a new claim should clear a t-statistic above 3.0 rather than 2.0 to be believed.^[3] Both say the same thing about your own sample: the bar is higher than intuition sets it.

kappi is a trade recorder: you commit a trade before the fact, it is sealed on your device for a time-capsuled delay you choose, then kappi publishes it on a Merkle-anchored log. The seal is the whole mechanism — once a stop, a target and a thesis are sealed, revising them is no longer a private edit. $15/month, no free tier.