Does price action work? How to assess a trading rule

“Price action” is too broad to prove or disprove as one strategy. It can mean many different ways of reading candles, structure and levels. The testable question is: does a specific rule, defined before the outcome, show a positive net result on data that did not help create it? This website does not establish such an edge.

Turn a chart opinion into a rule

“Buy strong support” is not reproducible. A test needs the market and timeframe, how support is identified using information available then, an entry trigger, exit rule, position calculation, costs and conditions for skipping. Write how to handle gaps, missing candles and a candle that contains both stop and target.

Two readers should be able to identify the same eligible decisions. If they cannot, first test the clarity of the definition. Knowing patterns but not when to enter gives a decision exercise for that stage.

A test record should answer these questions

Question	Evidence to keep
Did I select a favourable period?	The starting rule and all eligible periods
Did I change rules after losses?	Every version and every parameter tried
Did future information leak in?	Data available at each decision, including higher-timeframe closes
Does the result survive costs?	Net returns under stated fill, spread and fee assumptions
Does it hold outside development data?	Untouched later periods or a predeclared holdout
How uncertain is the result?	Sample size, dispersion, drawdowns and clustered exposures

A larger sample does not fix future leakage or cherry-picking. If you keep consulting the holdout and changing the rule, it becomes development data too.

A profitable-looking sample is not enough

Hypothetical arithmetic: 10 trades, each winner +2R and each loser −1R. Four wins and six losses produce 4 × 2 − 6 = +2R gross. If the assumed total cost is 0.25R per trade, the cost is 2.5R and net result is −0.5R.

Even a positive net result would describe that sample, not guarantee an underlying edge. The win-rate and costs guide explains the assumptions. The simulation limitations identified in 17 CFR 4.41, including hindsight and nonexecution, are reasons to keep hypothetical results distinct from live records.

What can our cases establish?

Our 110 touch labels describe a selected BTC/ETH/SOL educational bank. They include neither filled entries nor a tested entry rule. They do not establish that confirmation improves returns or supply a strategy win rate. Inspect the methodology and public-data limits.

What should I do next?

If you cannot yet define a rule, practise stating decisions with the future hidden. If the rule is already precise, log all eligible trades, preserve a holdout and evaluate the assumptions before optimising it.

Price Action Trading Simulator supports the first task and manual historical replay on iPhone and iPad. It does not run coded tests across many symbols or reproduce broker fills. One five-case session is free; backtest starts need Pro. For the difference between a replay result and live execution, read profitable backtest, losing live.