A trading strategy is not a set of indicator settings. It is a written answer to five questions, specific enough that another person could follow it and reach the same decisions you would. Most traders who believe they have a strategy have never written it down — and discover, when they try, that it was never finished.
This is the first lesson in our Trading Strategies section. Before covering scalping, intraday, swing and the rest individually, it is worth establishing what actually makes a strategy work, why most fail, and how to test one honestly.
The five questions
Every complete strategy answers these. If any answer is vague, that is where your losses will concentrate.
1. What do you trade? Which instruments, on which timeframe, during which sessions. “Forex” is not an answer. “EUR/USD and XAU/USD on the 4-hour chart, during London and the overlap” is.
2. When do you enter? The specific conditions that must be present. Not “when it looks good” — a list of criteria that are either met or not met.
3. Where is your stop? Decided before entry, based on where the trade idea is proven wrong.
4. Where do you exit with profit? A defined target, or a defined rule for trailing.
5. How much do you risk? A fixed percentage, applied identically regardless of how good the setup appears.
Write your current strategy as answers to those five questions. Most traders find they can answer one, three and five clearly, but that question two — entry conditions — turns into a paragraph of qualifications. That vagueness is not a detail. It is the gap where emotional decisions enter, and it is why identical setups produce different actions on different days.
Why strategies fail
In twelve years I have seen far more strategies abandoned than tested. The failures cluster into four patterns.
The strategy was never actually defined
A trader has a general approach — buy support in an uptrend, sell resistance in a downtrend — but no specific criteria. On Monday a level qualifies; on Thursday a similar level does not, because the trader felt differently that day.
Results from an undefined strategy cannot be assessed, because there was no consistent input to assess.
The sample was too small
A trader takes six trades, loses four, and concludes the strategy does not work. A method winning 50% of the time produces that outcome routinely. Four losses in six is entirely ordinary variance.
Judging a strategy on fewer than fifty trades is judging noise.
Position sizing undermined it
The signals were fine. The trader risked 8% on some trades and 1% on others, depending on confidence. A losing streak landed on the large positions, and the account could not recover.
This is the most common cause of failure I encounter, and the strategy is usually blamed for it.
It was abandoned mid-drawdown
Every method has losing periods. The trader who switches strategies during one never completes a full cycle, and adopts a new method just in time for its own drawdown.
Win rate is not the goal
Beginners optimise for win rate, which is intuitive and misleading.
A strategy winning 40% of trades can be highly profitable. A strategy winning 80% can lose money. What matters is the relationship between win rate and the size of wins versus losses.
Strategy A — 40% win rate, 1:3 risk-reward
10 trades: 4 wins × 3R = +12R
6 losses × 1R = −6R
Net: +6R
Strategy B — 80% win rate, 1:0.2 risk-reward
10 trades: 8 wins × 0.2R = +1.6R
2 losses × 1R = −2R
Net: −0.4R
Strategy B wins twice as often and loses money. This is exactly the profile of strategies that take small profits quickly and hold losers — which is what most untrained traders do naturally, because it feels better.
R here means risk units. If you risk 1% per trade, one R is 1% of your account. Thinking in R rather than dollars removes account size from the conversation and makes strategies comparable.
Expectancy — the number that matters
Expectancy tells you what an average trade is worth over many trades.
Expectancy = (Win% × Average win)
− (Loss% × Average loss)
Example: 45% win rate
Average win 2R, average loss 1R
(0.45 × 2) − (0.55 × 1) = 0.35R
Positive expectancy means the strategy makes money over a large sample. That is the only definition of “works” that means anything.
An expectancy of 0.35R means each trade is worth 0.35% of your account on average, if you risk 1% per trade. Across two hundred trades a year, that is meaningful. Across ten trades, it means nothing — variance dominates at small sample sizes.
Matching a strategy to your life
This is where most strategy selection goes wrong, and it has nothing to do with the market.
| Style | Hold time | Screen time | Trades/month |
|---|---|---|---|
| Scalping | Seconds–minutes | Constant | 100–400 |
| Intraday | Hours | 2–4 hrs daily | 20–60 |
| Swing | Days–weeks | 20 min daily | 5–15 |
| Position | Weeks–months | Weekly review | 1–5 |
A trader with a full-time job who selects scalping has chosen a method they cannot execute. They will take trades at the wrong times, miss setups, and rush decisions — then conclude the strategy does not work.
The strategy was fine. The match was not.
Start from your available hours, not from which style sounds most appealing. Scalping looks exciting and is the hardest to execute well. Swing trading looks slow and suits most people with jobs considerably better.
Discretionary versus mechanical
Two broad approaches, with a real trade-off.
Mechanical strategies specify conditions precisely enough that judgement plays no role. If the conditions are met, you take the trade. These are testable, consistent, and easier to follow under pressure — but they cannot adapt to context.
Discretionary strategies use rules as a framework while the trader judges specific situations. These can adapt, but they are difficult to test honestly and open a door for emotional decisions dressed as judgement.
For beginners, mechanical is considerably safer. Not because it produces better signals, but because it removes the decision point where inexperience does the most damage. Discretion is a skill built on top of a mechanical foundation, not a substitute for having one.
Testing a strategy honestly
Three stages, in order.
Backtesting. Reviewing historical charts to see how the rules would have performed. Useful for checking the logic holds, but vulnerable to hindsight bias — it is very easy to see a setup clearly when you already know what happened next.
Forward testing on demo. Trading the rules in live conditions with virtual money. This removes hindsight bias. Aim for at least fifty trades.
Live at minimum size. As covered in the Beginner’s Guide, this is where you discover whether you can actually follow the rules when money is real. Many strategies that worked perfectly on demo fail here, and the strategy is not the reason.
Change one thing at a time, and only after completing a full sample. Traders who adjust the entry criteria, the stop distance and the timeframe simultaneously after fifteen trades have learned nothing — they cannot attribute any change in results to any specific adjustment. This is the difference between testing and fiddling.
The journal
A strategy without records is untestable.
Record for each trade: date, instrument, direction, entry, stop, target, position size, the reason you entered, and the outcome. The reason matters most — broker statements record what happened, not why.
After fifty trades, patterns appear that are invisible day to day. You may find that your losses cluster in one session, or on one instrument, or on trades taken outside your written criteria. That last category is usually larger than traders expect.
What a realistic strategy looks like
Complete, written, and unremarkable:
Instruments: EUR/USD, GBP/USD
Timeframe: 4-hour chart
Sessions: London and overlap only
Entry: Price above 200 EMA (uptrend)
Pullback to 20 EMA
RSI between 40–55, turning up
Bullish reversal candle at the level
Stop: 2 × ATR below entry
Target: 2 × the stop distance (1:2)
Risk: 1% of account per trade
Exit rule: One target, one stop. Full close
at whichever is reached first.
Notice the final line. A single stop and a single target, closing the full position, is deliberately simple — and that simplicity is a genuine advantage for anyone still learning. Partial closes and multiple targets introduce decisions mid-trade, which is exactly when judgement is least reliable.
This is the structure we use for our own published signals, for the same reason: one entry, one stop, one target, full close. It makes the outcome unambiguous and the position sizing straightforward.
What to do next
The lessons ahead cover each major style in detail — what it requires, who it suits, and where it fails.
Start with scalping, which is both the most popular choice among beginners and the least suitable for them. Understanding why is useful even if you never trade it.
Frequently asked questions
What makes a trading strategy work?
Positive expectancy over a large sample, combined with position sizing small enough to survive losing streaks. The entry criteria matter less than most traders assume — consistency of execution and risk control determine outcomes far more.
What is a good win rate for forex?
Win rate alone means nothing without the risk-reward ratio. A 40% win rate at 1:3 is highly profitable; an 80% win rate at 1:0.2 loses money. Focus on expectancy, which combines both figures into one number.
How many trades before I know if a strategy works?
At least fifty, and a hundred is better. Below that, variance dominates — even a profitable strategy routinely produces stretches of five or six consecutive losses that mean nothing.
Should I use a mechanical or discretionary strategy?
Mechanical, especially early on. Precisely defined rules remove the decision point where inexperience causes the most damage, and they can actually be tested. Discretion is a skill built on top of a mechanical foundation.
Can I trade profitably with a job?
Yes, with swing trading. Positions held for days require around twenty minutes of daily review and orders placed in advance. Many consistently profitable traders work this way permanently — it is not a compromise.
Why do most trading strategies fail?
Four reasons dominate: the strategy was never defined precisely enough to follow consistently, the sample was too small to judge, position sizing was inconsistent, or it was abandoned during an ordinary drawdown. The signals themselves are rarely the constraint.