Why Statistics Before Strategy
Every trading result is a mix of edge and luck. Over 20 trades, luck dominates. Over 500 trades, edge dominates. A trader who cannot tell the two apart will abandon a good system during a normal losing streak, or double down on a bad one after a lucky month.
This guide gives you the five numbers that describe any strategy, and the one question to ask before believing any of them: how many trades is this based on?
Flip a fair coin 10 times: 7 or more heads happens about 17% of the time. That does not make the coin special. A trading record of 10 trades is the same - a 70% win rate over 10 trades tells you almost nothing. Only a large number of trades separates a real edge from a lucky streak.
The Five Numbers
Measure everything in R - multiples of the amount you risked. A trade that risked $100 and made $180 is +1.8R. Working in R makes results comparable across account sizes.
| Number | Formula | What it tells you |
|---|---|---|
| Win rate | wins / total trades | How often you are right - on its own, almost meaningless |
| Average win / average loss | mean R of winners, mean R of losers | How much you make when right versus lose when wrong |
| Expectancy | (win rate x avg win) - (loss rate x avg loss) | Average R per trade - the edge itself |
| Profit factor | gross profit / gross loss | Above 1.0 = profitable; 1.3-2.0 is realistic for a robust system |
| Maximum drawdown | largest fall from a peak to a later low | The pain you must survive to collect the edge |
How Many Trades Before It Means Anything
A win rate measured on a small sample has a wide margin of error. The standard error of a win rate p measured over n trades is sqrt(p x (1 - p) / n). Roughly 95% of the time the true value lies within two standard errors of what you measured.
| Measured win rate | Trades | 95% range of the TRUE win rate |
|---|---|---|
| 55% | 40 | about 39% to 71% |
| 55% | 100 | about 45% to 65% |
| 55% | 400 | about 50% to 60% |
After 40 trades, a "55% win rate" cannot be told apart from a coin flip - and even after 100 the range still reaches down to 45%. That is why every serious backtest in this course asks for at least 100 trades as a floor, not as proof, and why a live track record needs months, not weeks. The line to clear is the strategy's own break-even win rate, not 50%: a system whose winners are twice its losers breaks even at 33%, so a 55% measured over 100 trades is already well clear of it.
Losing Streaks Are Normal
Even a good system produces long losing streaks. For a win rate p and loss rate q = 1 - p over N trades, the typical longest losing streak is roughly log(N x p) / log(1/q) - and the real one is often longer. Counted exactly:
| Win rate | Trades | Typical longest losing streak | Longer than this less than 1 time in 10 |
|---|---|---|---|
| 50% | 100 | 6 | 8 |
| 40% | 100 | 7 | 11 |
| 40% | 500 | 11 | 14 |
At 1% risk per trade, 11 losses in a row is a drawdown of about 10.5%. If that would make you quit, the problem is your risk per trade, not your strategy.
- Order A100 trades40% win rate2.2R avg win1R avg loss+0.28R expectancy+28R total12.6R max drawdown9 losses in a row (max)
- Order B100 trades40% win rate2.2R avg win1R avg loss+0.28R expectancy+28R total7R max drawdown5 losses in a row (max)
- Both lists hold exactly the same trades: 40 winners averaging 2.2R, 60 losers of 1R, expectancy +0.28R, +28R at the end.
- Only the order differs. Order B's worst run is 5 losses and its deepest drawdown 7R. Order A meets 9 losses in a row and a 12.6R drawdown - with the same edge.
- You do not choose the order. Size your risk so the order A picture is one you can sit through.
Drawdown and Recovery
Losses and gains are not symmetric. A 10% loss needs an 11% gain to recover; 25% needs 33%; 50% needs 100%. Keep risk per trade small enough that the worst expected streak leaves you able to recover.
This guide is education, not financial advice. Most retail CFD accounts lose money; statistics help you find out honestly whether yours is one of them.