A trading strategy is a setup plus the rules that make it repeatable: which market and timeframe, what triggers entry, where the idea is wrong, how you exit, and how much you risk. A setup tells you what to look for. A strategy tells you what to do every time you find it.
That distinction is the whole of this page, and it is the thing most retail traders skip. Ask someone what their strategy is and you will usually get a pattern — a flag, a level, a moving-average cross. Ask what size they take it in, where they get out when it fails, and how many times they have traded it under those exact rules, and the answer thins out fast. A pattern you recognise is not a system you can run.
What separates a setup from a strategy
A setup is a recurring configuration of price that you can name and point at. There are hundreds of them, they are freely available, and none of them is proprietary. A strategy is the machinery you wrap around one of them so that two traders following it on the same day would place roughly the same orders.
The practical test is this: hand your strategy in writing to someone who trades the same market and leave the room. If they could take the next occurrence without phoning you, you have a strategy. If they would have to ask "how much?" or "where's the stop?", you have a setup and a hope.
The six parts every trading strategy needs
| Component | The question it answers | What happens without it |
|---|---|---|
| Market and timeframe | What am I trading, and on which chart? | You take the setup wherever you see it, on whatever chart is open |
| Trigger | What exact event turns the pattern into an order? | You enter early on anticipation, or late on regret |
| Invalidation | What price proves the idea wrong? | Losses have no defined size, which is how accounts end |
| Exit and targets | Where do I take profit, and in what portions? | Winners get cut early and losers held late |
| Size rule | How much do I risk on this occurrence? | Conviction sets size, and conviction is highest before losses |
| Do-not-trade conditions | When do I skip it entirely? | The strategy gets applied in the conditions it was never built for |
The last one is the least glamorous and the most protective. Every strategy has an environment where it works and an environment where it bleeds — a breakout method in a dead range, a fade in a trend day. Writing down the conditions under which you simply do not trade removes more bad trades than any refinement to the entry ever will. Market structure is the usual way to define them.
Notice that four of the six parts are about risk and exit, not entry. Entry is where attention naturally goes and where the smallest share of the outcome is decided. If you want the underlying logic, risk management is the pillar that carries it.
How to tell whether a strategy actually has an edge
An edge is not a feeling, a screenshot, or a run of good weeks. It is a positive expectancy measured over a sample large enough to be meaningful. Expectancy combines how often you win with how much you win when you do:
Expectancy = (win rate × average win) − (loss rate × average loss)
That formula is why win rate on its own is a poor measure. A strategy that wins 35 percent of the time with winners three times the size of losers is comfortably positive; one that wins 80 percent of the time with an occasional loss five times the average win is not. Run your own numbers through the expectancy calculator before drawing conclusions, and read trading expectancy for what the number does and does not tell you.
Sample size is the part traders wave away. Ten winners in a row feels like proof and is not: a coin-flip strategy produces a run of ten often enough that you should expect to see one eventually just by trading. Treat a hundred trades under identical rules as a working minimum before you believe your own results, and treat anything below thirty as a hint. The quickest honest route to a sample is a trading journal that records the rule version alongside every trade, so you know which strategy the numbers actually describe.
Why edges decay — and what to do about it
Strategies are not permanent, and the best public evidence comes from the closest analogue researchers can study. In Does Academic Research Destroy Stock Return Predictability? (Journal of Finance, 2016), R. David McLean and Jeffrey Pontiff examined 97 variables that published academic studies had shown to predict stock returns. They found portfolio returns were 26 percent lower out of sample and 58 percent lower after publication (McLean & Pontiff, 2016). Part of that decline is statistical bias in the original finding; part is real money arriving once the edge became known.
Those are equity anomalies over months, not intraday setups, and the numbers do not transfer to a five-minute chart. The mechanism does: an edge that is widely known attracts participants who trade against it, and an edge that looked strong in the data you selected it from usually looks weaker in the data you did not. The practical consequences are unglamorous and worth taking seriously.
- Keep measuring the strategy you actually trade. Decay shows up in a rolling expectancy long before it shows up as a bad month you can explain away.
- Be suspicious of a strategy that was optimised to fit history. The more parameters you tuned to make the backtest pretty, the less of the result belongs to the market.
- Expect the environment to matter more than the tweak. Most strategies that "stop working" have simply met a market regime they were never suited to.
- Retire deliberately, not emotionally. Decide in advance what level of measured decline means you stop trading it, so the decision is not made during a drawdown.
The main families of trading strategy
Almost every intraday strategy is a variation on one of four ideas. Knowing which family yours belongs to tells you which market conditions will hurt it.
- Breakout and continuation. You buy strength through a level and expect the move to extend. Works in trending, expanding markets; bleeds in chop, where every break reverses. The opening range breakout is the archetype, and what a breakout actually is covers the mechanics.
- Pullback and retest. You wait for price to come back to a broken level and continue from there. Fewer trades, better locations, more missed moves. See what a retest is.
- Mean reversion and fade. You sell extension and buy weakness back toward a reference such as VWAP. Works in ranges; produces its worst losses on exactly the trend days when it feels most obviously right.
- Reversal at a level. You trade an exhaustion or failure at a significant price — a prior day high, a supply zone, a failed breakout. The highest reward-to-risk of the four and the lowest hit rate, so it punishes sizing mistakes hardest.
None of these is better than the others in the abstract. They are better or worse in a given market on a given day, which is why the do-not-trade condition is a component and not an afterthought.
How many strategies should you run?
One, until it is automatic. The argument for a single strategy is arithmetic rather than philosophy: every strategy you add divides your sample. Five strategies that each trigger once a week will take years to give you a hundred trades of any one of them, which means you will never know which of the five is carrying you and which is quietly bleeding.
There is a second cost. Running several strategies means choosing between them in real time, and that choice is made under exactly the pressure that produces the worst decisions. A trader with one strategy and a written do-not-trade list spends the session waiting. A trader with five spends it deciding. Once the first is genuinely automatic — you can state it from memory and your journal shows a hundred trades taken to the rule — a second, ideally one that works in the conditions the first sits out, is a reasonable addition.
Frequently Asked Questions
What is the difference between a trading setup and a trading strategy?
A setup is a pattern you recognise on a chart. A strategy is that setup plus every rule needed to trade it the same way twice: the instrument and timeframe, the trigger that turns a pattern into an order, the price that says you were wrong, how you exit, how much you risk, and the conditions under which you do not take it at all. Most traders own a folder of setups and believe they own a strategy.
How many trades do you need to know if a strategy works?
More than most traders use, and the honest answer is that no fixed number settles it. A run of ten winners tells you almost nothing, because a strategy with a 50 percent win rate produces ten in a row often enough to be unremarkable across a career. Judge on expectancy across a sample of a hundred trades or more, taken under the same rules, and treat anything smaller as a hint rather than evidence.
How many trading strategies should one trader run?
One, until it is genuinely automatic, and rarely more than two or three after that. Every additional strategy divides your sample, so five strategies traded once a week each will take years to produce enough data to evaluate any of them. It also multiplies the decisions you make under pressure, which is where most traders lose far more than they lose to a weak edge.
Do trading strategies stop working?
Many do, and there is published evidence for it in the closest analogue we have. McLean and Pontiff studied 97 stock-return predictors from the academic literature and found portfolio returns 26 percent lower out of sample and 58 percent lower after the findings were published. The practical response is not to abandon strategies but to keep measuring the one you trade, so a decay shows up in your own numbers rather than in your account balance.
Bottom line
Write down six things and you have a strategy: market and timeframe, trigger, invalidation, exit, size rule, and the conditions under which you stand aside. Trade it a hundred times without changing it, measure expectancy rather than win rate, and let the number — not a good week — decide whether it stays. Everything else on this site is a component of that structure rather than a substitute for it, and the fastest way to turn the list into something you can actually run is building a trading plan around a single setup you already understand.