Best Intraday Trading Strategies

Any search for the best intraday trading strategies returns entry rules. Almost none of them include the information required to judge whether the rule is worth using, which is why traders accumulate methods without ever establishing that any of them works.

The useful question is not which strategy is best but how to evaluate one. The criteria below apply to any method, including ones you devise yourself, and applying them eliminates most of what circulates.

Expectancy, Not Accuracy

Win rate is the most quoted and least informative statistic in trading. A method winning most of the time with small gains and occasional large losses loses money; one winning less than half the time with gains larger than its losses makes money.

What matters is the combination of average gain, average loss and frequency, taken together and after costs. Any presentation that leads with accuracy while omitting average loss size is either uninformed or selective, and both warrant the same scepticism.

Sample Size Determines What You Can Conclude

A dozen trades tell you nothing. Short sequences are dominated by variance, and both a good method and a poor one can produce almost any short-run result.

Judging requires a sequence long enough for variance to average out, spanning different conditions rather than a single favourable stretch. Traders routinely abandon sound methods after a handful of losses and adopt poor ones after a handful of wins, which is a sample-size error rather than a judgement failure.

Costs Must Be Inside the Calculation

A strategy evaluated on price movement alone is not being evaluated at all. Brokerage, exchange charges, statutory levies and the spread recur on every round trip and scale with frequency, while the edge does not.

Include the full round-trip cost at your actual size in every figure. Methods that look profitable in analysis frequently fail in practice for this reason alone, and the higher the frequency the more decisive it becomes, as set out in intraday trading strategies.

Regime Dependence Is Not Optional

Every method depends on conditions. Breakout approaches need directional sessions; mean reversion needs ranges to hold; momentum needs participation. A method tested only in the conditions that suit it has not been tested.

Examine performance separately across quiet and volatile periods, and across trending and rangebound ones. A method that earns in one regime and gives it back in another is not a strategy but a bet on conditions persisting.

Overfitting Looks Exactly Like Skill

Adding conditions improves past results and reduces trade count. Beyond a small number of parameters, the rules are describing the specific sequence they were derived from rather than any repeatable behaviour.

Two guards are practical. Prefer methods with few conditions that can be stated in a sentence, and require any rule to make sense as a description of participant behaviour rather than only as a pattern in data. A rule with no explanation for why it should work usually has none.

Drawdown Is the Constraint That Binds

A method’s worst losing sequence matters more than its average result, because that sequence determines whether it can be followed. A strategy producing a run of losses larger than the trader’s tolerance will be abandoned partway, at which point its long-run statistics are irrelevant.

Estimate the worst run the method has produced and ask honestly whether you would continue through it. If the answer is no, the method is unsuitable regardless of its overall figures, and reducing size until the answer becomes yes is the correct adjustment.

Forward Testing Before Committing

Historical results are constructed with knowledge of what happened next, however carefully one avoids it. Forward testing — applying the method to unseen conditions, on paper or at minimal size — removes that advantage.

It also tests something historical analysis cannot: whether you can execute the method under live conditions. Many strategies fail not because the rules are wrong but because the trader cannot follow them when money is moving.

Execution Assumptions Must Be Realistic

Analysis usually assumes entry and exit at the price on the chart. Live trading involves spreads, slippage and partial fills, and in fast conditions the difference can exceed the expected gain per trade.

Test the assumption at your intended size in the instruments you actually trade. A method viable in a liquid index instrument can be unviable in a thinner one purely through execution, as covered in equity intraday tips.

Whether the Method Suits You

A strategy requiring constant attention is unusable for someone who cannot watch continuously. One producing long inactive periods is unusable for someone who cannot tolerate them. One demanding rapid decisions is unusable for someone who deliberates.

Fit is not a secondary consideration. The best method you cannot execute is worse than an average one you can, and honest assessment here saves considerable expense.

What Published Results Omit

Presented results are usually selected. The strategies discontinued do not appear, the start date is frequently favourable, costs are often excluded and position sizing is left unstated.

Ask for the four qualifications that make any figure interpretable: period, comparison, costs included, and whether all trades are shown. A figure lacking them is not evidence, and the same scrutiny applies to any signal service presenting a record.

Building Rather Than Collecting

Traders accumulate strategies instead of developing one. Each new method resets the sample size, so nothing ever reaches the length at which a conclusion becomes possible.

Choose one approach suited to your instruments and availability, apply it consistently, record every trade and evaluate it after enough of them. The routine that supports this is in the intraday trading guide, with instrument considerations in intraday tips and a starting path in intraday tips for beginners.

Consistency of Execution Is Part of the Method

A strategy’s recorded results assume it was applied the same way every time. Where entries are sometimes taken early, stops occasionally widened and positions sized by mood, the results describe a mixture of approaches rather than the method being evaluated.

This is why the record should include whether the plan was followed on each trade. Separating the trades executed as designed from the rest frequently reveals that the method is sound and the execution is not, which is a completely different problem with a completely different remedy.

Beware Comparing Against the Wrong Benchmark

An intraday method should be judged against what the capital could have done elsewhere with comparable risk, and against the effort it consumes. A method producing modest returns for constant attention and substantial stress may be underperforming a far simpler alternative.

This comparison is rarely made because it is uncomfortable, and it is the most useful one available. The relevant alternative for long-horizon capital is described under investment advisory, and the two should not be funded from the same pool.

Judging a Method You Did Not Build

The same criteria apply to any approach obtained elsewhere, with one addition: establish what conditions it was developed in and whether those conditions currently exist. A method derived during a strongly directional period will disappoint in a quiet one.

Ask what the method’s worst run has been and how many trades the record covers. A presentation that leads with accuracy and omits average loss, sample size and costs is selecting the flattering figure, and the same scrutiny applies to any record shown for a method you are considering adopting.

Records Are the Only Evidence You Control

Published results describe someone else’s execution. Your own log, kept from the first trade, describes what the method produces in your hands at your size in your instruments, including slippage and missed entries.

That is the figure that decides whether to continue, scale up or stop. Everything else is an indication, and traders who rely on external records rather than their own routinely persist with methods that have never worked for them specifically.

FAQs

Is a high win rate a good sign?

Not on its own. A method winning often with small gains and occasional large losses loses money. Average gain, average loss and frequency must be assessed together.

How many trades before I can judge a method?

Enough for variance to average out, across different conditions rather than one favourable stretch. Short sequences are dominated by luck in both directions.

What is overfitting in practice?

Adding conditions until past results look excellent. The rules then describe the period they came from rather than repeatable behaviour, and they fail on new data.

Why does drawdown matter more than average return?

Because it determines whether the method can be followed. A losing run beyond your tolerance will be abandoned partway, making the long-run figures irrelevant.

Should costs be included in evaluation?

Always, at your actual size. Costs recur on every round trip and scale with frequency, while the edge does not, so excluding them invalidates the whole assessment.

Is paper trading useful?

Yes, for testing rules on unseen conditions and for checking whether you can execute them. It understates the emotional difficulty, so treat it as necessary rather than sufficient.

What qualifications should accompany any published result?

Period, comparison, costs included and whether all trades are shown. Without these four, a figure cannot be interpreted and is not evidence.