Contents (10)
- Why your balance cannot tell you
- The five numbers that can
- Expectancy is the number that decides it
- Five winning trades that proved nothing
- How many trades before you can trust the answer
- Is it broken, or is this normal?
- Signs you never actually tested it
- What backtesting is worth
- Common questions
- The short version
Most people answer this question by looking at their account balance. The trouble is that the balance is the one thing that cannot answer it. A good strategy can lose for weeks. A bad one can make money for weeks. While it is happening, you cannot tell which is which.
There is a better way. It takes five numbers, and more trades than you would like.
Why your balance cannot tell you
One trade tells you almost nothing. Twenty trades still tell you almost nothing. That is not a motivational line, it is arithmetic, and it is covered in full in why trading is so hard.
To be reasonably sure that a strategy winning 55% of the time is not just a coin flip, you need somewhere around 785 trades. So if you are judging a strategy on last month, you are guessing.
What you can do sooner is measure the things that settle down long before the balance does.
The five numbers that can
All five come out of your trade record. You do not need software, and none of them is the balance.
First, one word you will need: R. R is the amount you risk on one trade. If you risk $100 and win $200, that is a 2R win. If you lose the $100, that is a 1R loss. Thinking in R lets you compare trades of different sizes.
| Number | How you get it | What it tells you |
|---|---|---|
| How many trades | Count them. Only the trades of this exact strategy | Whether the other four mean anything yet |
| Win rate | Winners divided by all trades | Very little on its own. It only matters next to the next row |
| Average win vs average loss | Average winner divided by average loser, in R | Whether the strategy pays enough when it is right |
| Expectancy | See below | Whether the strategy makes or loses money |
| Worst losing streak | The longest run of losses so far | Whether your position size can survive it |
People stare at the win rate and ignore expectancy. A 70% win rate is worthless if the losses are four times the size of the wins. A 35% win rate is excellent if the winners are big enough.
Expectancy is the number that decides it
Expectancy is what one trade is worth on average. The formula is short:
Expectancy = (win rate × average win) − (loss rate × average loss)
Say you win 40% of the time, your average win is $200 and your average loss is $100. That gives (0.40 × 200) − (0.60 × 100), which is 80 − 60, so $20 per trade. Over a hundred trades, that is $2,000. Not from being clever. From having an edge and letting it run.
The same idea is easier to read in R. Here is what one trade is worth for the combinations people actually run:
| Reward per winner | Win rate 40% | Win rate 50% | Win rate 60% |
|---|---|---|---|
| 1R | −0.20R | 0.00R | +0.20R |
| 1.5R | 0.00R | +0.25R | +0.50R |
| 2R | +0.20R | +0.50R | +0.80R |
| 3R | +0.60R | +1.00R | +1.40R |
Two things stand out. If your winners are the same size as your losers (the 1R row), you have to win more than half the time just to break even, before costs. That is why scalping for small targets is harder than it looks. And a strategy that wins 2R at a 40% win rate makes money while being wrong three times out of five.
Anything at or below zero is not a strategy having a bad month. It loses money by design, and no amount of discipline will fix it.
Five winning trades that proved nothing
One strategy looked perfect because the first five trades all won. Five in a row is enough to feel invincible and nowhere near enough to know anything. I did not understand that at the time, so I raised my size and started making plans.
Then the market changed a little and the whole thing fell apart. The same rules that had looked flawless started producing loss after loss. What I had read as proof was a run of luck.
It took close to a hundred trades before the numbers said anything I could lean on. Nobody impatient wants to hear that. It is still the most useful thing I know about testing a strategy.
How many trades before you can trust the answer
Expectancy from fifteen trades is a rumour. Roughly, this is what each number of trades can tell you:
| Trades logged | What you can say |
|---|---|
| Under 20 | Nothing. Change nothing |
| 20 to 50 | Whether you followed the plan. Not whether the plan works |
| 50 to 100 | A first read on expectancy and your worst streak. Treat it as provisional |
| 100 to 300 | A usable estimate, if the market did not change under you |
| Several hundred | Enough to tell a small edge from luck |
The awkward truth is that most retail traders never log enough trades of one approach to know whether it worked. They switch after every bad run, which resets the count to zero. Then they do it again for years.
Is it broken, or is this normal?
This is the question people actually ask, usually in the middle of a drawdown. Here is how to tell the two apart.
| What you see | Probably normal | Probably broken |
|---|---|---|
| Five or six losses in a row | Yes. At a 45% win rate almost everyone sees this inside 100 trades | Only if the losses are bigger than you planned |
| A drawdown deeper than anything before | Possibly, if your record is short | Worth a look once it is well past your worst streak |
| Win rate fell, R did not | Likely variance | If it lasts 100+ trades, the setup may have stopped working |
| Losses bigger than planned | Not variance | An execution problem, not a strategy problem |
| Setups are getting hard to find | The market changed | The strategy may suit conditions that have ended |
The row that matters most is the fourth. If your losses are bigger than you planned, you have not tested the strategy at all, because what you ran was not the strategy.
Signs you never actually tested it
Before you decide a strategy does not work, rule these out. All four are far more common than a broken method.
- You changed the rules during the test. Then the sample is of nothing.
- You did not follow the rules. If the “did I follow my plan” column in your journal is mostly N, you tested your discipline, not your strategy.
- The rules were too vague to break. “Buy the pullback” is not a rule. “Buy the first close above the 20 EMA after two touches of the level” is.
- Your position size kept changing. If risk varied from trade to trade, the results cannot be compared and expectancy in R means nothing.
What backtesting is worth
Testing a strategy on past data is useful, and it is oversold. Here is the honest split.
What it gives you: a big sample fast, an idea of the worst drawdown the approach would have had, and a quick way to throw out ideas that never worked at all.
What it cannot give you: honesty about the past. You already know how it ended, and every choice you make while testing is shaped by that. Add enough rules and any period in history can be made profitable. That tells you nothing about the next one. And you cannot backtest your own behaviour, which is usually the real limit.
The practical middle: use backtesting to reject ideas, not to confirm them. Then forward test the survivors at the smallest size your broker allows. A strategy that works on paper and falls apart in a live account is the normal outcome, not the exception.
Common questions
How do I know if my trading strategy works?
Work out expectancy from at least a hundred logged trades where you followed the rules and kept risk the same. If it is clearly above zero after costs, it works. If you cannot work it out because the record does not exist, that is your answer to a more urgent question.
What win rate should a good strategy have?
There is no such number. Win rate only means something next to your reward to risk. 35% is excellent at 3R and fatal at 1R.
How long should I test a strategy?
Count trades, not weeks. A hundred is a reasonable minimum for a first verdict. Decide the number before you start. If you decide it during a drawdown, you will stop at the worst moment.
Should I run more than one strategy?
Not until the first one has a few hundred logged trades. Two strategies halve the sample for each and double the chance you learn nothing about either.
The short version
Five numbers: how many trades, win rate, average win against average loss, expectancy, worst streak. Expectancy is the verdict. The trade count decides whether the verdict means anything. The balance is not on the list.
Most strategies that get abandoned were never tested. They were interrupted.