Every strategy has conditions it needs. The obvious response is to run several so that something is always working — which is correct in principle and, done carelessly, produces one blurred strategy that nobody can evaluate.
Selling umbrellas and sunglasses from the same cart is genuinely sensible — one of them sells whatever the weather does. Selling umbrellas and raincoats is not diversification; it is one weather bet with two products.
A trend system and a mean-reversion system are the umbrella and the sunglasses. A breakout system and a momentum system are the umbrella and the raincoat, and a range-bound quarter takes both.
What actually diversifies
| Combination | Correlated? | Verdict |
|---|---|---|
| Trend following + mean reversion | Low — they need opposite conditions | Genuine diversification |
| Breakout + momentum | High — both need trending markets | One strategy, double size |
| Equity long + pairs trading | Low — one is directional, one is relative | Genuine diversification |
| Same system on two timeframes | Moderate to high | Partial at best |
| Same system on two sectors | High in a market-wide fall | Barely diversification at all |
Total open risk across systems is what actually binds. Two strategies at 1% each is 2% at risk when they happen to agree — which is exactly when they usually do.
Allocating capital between them
- 1Fixed allocations, decided in advance
Say 60% to the primary system and 40% to the secondary. Reviewed annually, not after a bad month — otherwise you are simply chasing whichever worked recently.
- 2One total risk budget
Cap combined open risk across all systems, not per system. Two systems fully deployed simultaneously must still respect your overall limit.
- 3Separate the records completely
Each system needs its own trade log and its own equity curve. Blended results make it impossible to tell which one is carrying the account.
- 4Never let one borrow from the other
Taking extra size in system A because system B is doing badly is how two disciplined systems become one undisciplined one.
How many is too many
Each additional system needs monitoring, execution attention and enough trades to be evaluated. Beyond two or three, most individual traders are managing systems rather than trading them.
- One primary system you know intimately
- One secondary that works in the opposite regime
- Separate logs, fixed allocations
- Enough trades in each to evaluate annually
- Five systems, none with enough trades to judge
- Allocations shifting after every bad month
- A blended log with no attribution
- Adding a system after reading about it, not testing it
You run a breakout system and a momentum system. Both had their worst months in the same three periods. What do you actually have?
Ek thele pe chhata aur dhoop ka chashma bech rahe ho — mausam kuch bhi ho, kuch na kuch bikta hai. Par chhata aur raincoat bechna? Woh do cheezein nahi, ek hi mausam ka daanv hai. Do system tabhi alag hain jab unka bura mahina alag-alag ho — warna woh ek hi system hai double size mein.
- Diversification between systems means their drawdowns do not coincide.
- Breakout plus momentum is one strategy; trend plus mean reversion is two.
- Cap combined open risk across systems, not per system.
- Keep separate logs — without attribution you cannot tell which one to keep.
- Two or three is the practical limit; beyond that you are managing rather than trading.
Mark it done to track your progress through the curriculum.
Common questions
Short, direct answers to what people ask about this topic.
- strategy correlation meaning in trading
- Strategy correlation describes whether two trading systems make and lose money at the same time. The test is not whether their rules look different but whether their drawdowns land in the same months — if they do, you are running one exposure at twice the size you intended, with extra bookkeeping on top. Low correlation is the entire reason two systems can be better than one.
- two systems whose drawdowns arrive in the same months are
- Effectively one strategy at double size. Diversification between systems is defined by whether the losing periods coincide, not by how different the entry rules read — a breakout system and a momentum system both need trending markets, so a range-bound quarter takes both of them. Plotting the monthly returns of the two against each other settles the question before any capital is committed.
- how many trading systems should one person run
- Two or three is the practical ceiling for most individual traders, because every extra system needs monitoring, execution attention and enough trades of its own to be judged. Past that you are managing systems rather than trading them, and none of them ever accumulates a sample big enough to evaluate. A workable shape is one primary system you know intimately plus one that earns in the opposite regime.
- why keep a separate trade log for each strategy
- Because a blended equity curve makes it impossible to tell which system is carrying the account and which is causing the damage, and that eventually leads to retiring the one that was quietly working. This is the attribution problem, and it is not administrative tidiness — it is what makes any future decision about either system possible. Give each system its own trade log and its own equity curve from the first trade.
- do I cap risk per system or across all systems
- Across all of them — total open risk is what actually binds, so two systems risking 1% each put 2% at stake whenever both happen to be in the market, which is usually the moment they agree with each other. Fixed allocations decided in advance, say 60% to the primary system and 40% to the secondary and reviewed annually rather than after a bad month, stop the split drifting toward whatever worked recently.