Designing a Strategy

2.1 Start with a reason

Mint wanted to build a system and started by trying combinations of indicators: RSI below 30 with a MACD crossover, then Bollinger Bands with volume spikes, then three moving averages at once.

"I'll just keep trying combinations until something works," said Mint.

"That's the fastest way to fool yourself," said Hash. "If you try enough combinations, something will always look like it worked in the past, purely by chance. Start with a reason why a strategy should work. Then test whether it did."

2.2 Hypothesis first

A good strategy begins with a hypothesis: a clear explanation of why it should make money.

Weak starting point	Strong starting point
"This indicator combination looked good on a chart"	"Crypto trends persist because of momentum and crowd behaviour, so following trends should earn returns"
"I'll test everything and keep the best"	"Funding rates become extreme when positioning is crowded, so fading extremes should work in ranges"

A hypothesis tells you what to test, which conditions should favour the strategy, and what would mean it has stopped working.

2.3 Common strategy types

Strategy type	Hypothesis	Works best in	Struggles in	Related module
Trend following	Prices that are rising tend to keep rising for a while	Strong trends	Choppy ranges (Module 4, Chapter 8)	Module 4
Mean reversion	Prices that move too far tend to snap back	Ranges	Strong trends	Module 4, Chapters 9 and 11
Breakout	Moves out of tight ranges tend to continue	Volatility expansion after compression	False breakouts (Module 4, Chapter 5)	Module 4
Carry and funding	Earn funding or basis while staying delta-neutral	Positive funding markets	Funding reversals, exchange failure	Module 7, Chapter 12
Arbitrage	Exploit price differences between venues	Fragmented markets	Competitive markets where gaps close fast	Module 1, Chapter 6
Rebalancing and accumulation	Buy and rebalance systematically over time	Long-term rising markets	Long declines	Module 12, Chapter 5

2.4 Where edges come from

An edge is a reason a strategy should earn more than random trading after costs. In crypto, possible sources include:

Source of edge	Example
Behavioural	Crowds overreact, chase trends, and panic (Module 12, Chapter 9)
Structural	Funding rates, liquidations, and unlocks create predictable flows (Modules 5 and 7)
Risk premium	Being paid for taking risks others avoid, like selling insurance (Module 8, Chapter 6)
Speed or access	Faster execution or better data; hard for individuals to compete here
Patience	Holding through volatility that others can't tolerate

For most individuals, edges from speed are unrealistic: professionals and bots dominate there (Module 1, Chapter 9). Edges from patience, behaviour, and structural flows on longer timeframes are more accessible.

2.5 Writing precise rules

Rules must be precise enough to test. Compare:

Vague rule	Precise rule
"Buy when the trend is up"	"Buy at the daily close when the close is above the 50-day SMA and the 50-day SMA is above its value 10 days ago"
"Sell if it looks weak"	"Sell at the daily close if the close is below the 50-day SMA"
"Use a sensible stop"	"Initial stop at entry minus 2 × 14-day ATR"
"Size it reasonably"	"Size so that the initial stop risks 1% of the account"

If you can't write a rule precisely, you can't test it, and you can't automate it.

2.6 Keep it simple

Simpler strategies tend to be more robust: they have fewer parts that could have fitted random patterns in past data.

Simple strategy	Complex strategy
Few rules and parameters	Many indicators and conditions
Easy to understand why it works or fails	Hard to know which part matters
Less likely to be overfitted (Chapter 4)	More likely to look great historically and fail live

"So the best strategy might be boring," said Mint.

"Usually," said Hash. "If a strategy needs eight indicators to work, it probably doesn't work."

2.7 Mint's first strategy

Component	Rule
Hypothesis	Bitcoin trends persist on daily timeframes due to momentum and crowd behaviour
Market	BTC spot
Timeframe	Daily, using UTC closes (Module 4, Chapter 2)
Entry	Buy at the close when price closes above the 50-day SMA, and the 50-day SMA is rising
Exit	Sell at the close when price closes below the 50-day SMA
Initial stop	Entry minus 2 × 14-day ATR
Sizing	Risk 1% per trade based on the initial stop; no leverage
Costs	0.1% per trade in fees and slippage
Expected weakness	Many small losses in ranging markets

2.8 Try it yourself

Write a hypothesis for a strategy you'd like to test, explaining in two sentences why it should work. Choose a strategy type from section 2.3, and write precise rules for every component in Mint's table. Identify the market conditions where you expect it to struggle.

Key takeaways from this chapter

1. Start with a hypothesis explaining why a strategy should work, not by trying random indicator combinations.
2. Common strategy types include trend following, mean reversion, breakout, carry, arbitrage, and systematic accumulation.
3. Edges come from behaviour, structure, risk premiums, speed, or patience; speed edges are unrealistic for most individuals.
4. Rules must be precise enough for two people, or a computer, to apply them identically.
5. Simpler strategies tend to be more robust than complex ones.
6. Know in advance which market conditions your strategy should struggle in.