In this lesson
- learn to state a trading idea as a hypothesis
- turn it into exact rules with no room for "by eye"
- find out why fewer rules are often more
"I buy when it looks strong and sell when it stops." That is how many people trade – and they never find out whether it works, because "looks strong" means something different every day. You can only test what is written so precisely that a computer could follow it.
1. Hypothesis: why it should work#
Every strategy starts with the sentence "the market tends to…". For example: "When Bitcoin is in a long-term uptrend, it tends to continue in it." And right after: why? Because people react to news slowly, trends attract more buyers, big players buy gradually.
A hypothesis without an explanation is dangerous. If you find a pattern that makes no sense ("it rises on Tuesdays after a full moon"), you most likely found chance that will not repeat.
2. Rules: exact and complete#
Now you turn the hypothesis into trading rules. They must answer everything that is decided in a trade:
| Question | Example of an exact rule |
|---|---|
| What do I trade? | BTC/USDT, spot |
| On which timeframe? | daily candles, decision after the day closes |
| When do I enter? | closing price above SMA 200 |
| When do I exit? | closing price below SMA 200 |
| Where is the stop-loss? | 2 ATR below the entry |
| How big a position? | 1% of the account at risk |
| What does a trade cost? | a 0.1% fee on both buying and selling |
The test of rules: give them to two people and let them go through the same chart. If they make different trades, the rules are not precise enough.
3. Less is more#
Every extra rule and every number that can be tuned (an average length, an RSI threshold, a stop-loss percentage) gives the strategy more chances to "learn the past by heart". In the previous course we called that overfitting.
- Start with as few rules as possible that still make sense.
- Prefer round, common numbers (20, 50, 200) to "discoveries" like 137.
- Every extra rule needs a reason, not just a better test result.
Rules are not a promise
Exact rules do not guarantee a profit. They only guarantee that you can find out honestly whether the strategy worked in the past – and that you know exactly what you are doing when you trade.
Common myths#
- "I have the rules in my head." Then they cannot be tested and they change with your mood.
- "I will add a filter and the results will improve." In a test almost always. In reality often not.
- "Why it works does not matter, as long as it works." Without a reason you will not notice when it stops working.
Summary
- A strategy starts with a hypothesis: what the market does and why.
- The rules must say exactly what, when, how much and at what cost – so that two people apply them the same way.
- Fewer rules and common numbers mean less risk of overfitting.
- Exact rules are a condition for testing, not a guarantee of profit.