A Portfolio of Strategies — Diversify Instead of Hunting One Grail

The typical trader's path looks like this: find a strategy, test it, watch the strategy run out of steam, drop it, and go looking for the next one — a better one, "the real one" this time. Repeat that five times and you'll notice a pattern: every strategy has a regime where it makes money, and a regime where it gives it back. Trend following shines in long trends and bleeds in consolidation. Mean reversion is the opposite. Breakout strategies feed on volatility and starve on quiet. Looking for a strategy that always works is looking for the Grail — and it ends the way it does in the legends: nobody ever brings it back.

Professional funds solved this problem differently. They don't look for one perfect strategy — they build a portfolio of strategies: a collection of approaches with different mechanisms, whose combined result is more stable than the result of any single component. This article shows how that logic works at the biggest players (drawing on Man Group's material on multi-strategy portfolios) and what a retail trader can realistically take from it.

Educational disclaimer: this material is for educational purposes only and is not investment advice. Diversification reduces the swings in results but doesn't eliminate the risk of loss — and if miscalculated, it can create a false sense of safety. Past results don't guarantee future ones.

Why One Strategy Isn't Enough

Not because it's a bad strategy — because it's a single one. Even a validated strategy with a real edge has two properties you can't engineer away:

First, regime-dependence. A strategy's edge is usually one market mechanism: continuation, mean reversion, a volatility premium. The market switches between states where that mechanism works and doesn't — and it switches without warning. A trend strategy can give back capital for many months of consolidation before a single large trend pays for all of it. Whoever runs only that strategy is betting their entire survival on patience through a drawdown.

Second, mortality. Edges fade — because the market changes, or because too many players discover the same edge. With a single strategy, the edge fading means the end of the operation; with a portfolio, it means losing one engine out of several.

The math favors the portfolio. If two strategies are of similar quality but their results are weakly correlated, combining them produces a similar return with smaller swings — that is, a better ratio of profit to pain. It's the same idea Harry Markowitz won a Nobel Prize for, just applied not to assets but to strategy result streams. Smaller drawdowns aren't cosmetic: they mean less risk you'll abandon a good system at the worst possible moment, and less temptation toward desperate decisions.

Strategy Correlation — the One Number That Actually Matters Here

The entire value of a strategy portfolio rests on one condition: the components have to be genuinely different. And "different" doesn't mean "differently named" — it means low correlation in results.

This is where the main trap lurks, and it's one Man Group solves internally through what's called strategy grouping: before allocating risk, the fund clusters strategies with similar drivers into groups (by style, by correlation, by clustering), so it doesn't accidentally build a portfolio where "four different strategies" turns out to be the same exposure four times over. Their example: two value strategies plus an event-driven strategy plus a trend strategy, each at 25% of risk — and suddenly half the portfolio's risk sits in value. The retail equivalent: breakouts, momentum and trend following on the same market are, in practice, one giant "let the move keep going" position. When consolidation arrives, all three bleed at once — and the trader is left wondering why "diversification didn't work."

How to calculate this yourself, without a quant desk: gather daily or weekly results for each strategy (from backtests, then ongoing from your trade journal) and calculate a simple correlation in a spreadsheet. Rough interpretation: below ~0.3 is real diversification; around 0.5-0.7 is partial; above that, it's one risk in two wrappers. The classic pair with naturally low or negative correlation is trend plus mean reversion: one makes money on continuation, the other on its absence, so by definition they favor opposite market states.

How the Biggest Players Do It — Three Pillars From Multi-Strategy Practice

There are three concepts worth borrowing from Man Group's material on building multi-strategy portfolios — all of them scale down to a retail account:

1. Diversify through risk allocation, not prediction. A fund doesn't ask "which strategy will make money this quarter" (return forecasts are noisy) — instead it spreads risk across groups of strategies so that none of them dominates. That's a fundamental shift in the question, from "what's going to happen?" to "how do I need to be positioned, whatever happens?".
2. Risk targeting. A portfolio of low-correlation strategies naturally has low volatility — sometimes low enough that it won't hit its target. Funds then scale exposure up to the planned risk level. The retail takeaway is more cautious: the size of each position follows from planned risk, and the portfolio's total risk is a designed quantity, not an accidental sum.
3. Risk management as a separate layer. Strategies can behave differently than the backtest promised. Funds run an independent mechanism that cuts a strategy's risk during an excessive drawdown and watches whether positions from different strategies are quietly stacking into one giant exposure. For you, that role is played by rules written down in advance: a per-strategy loss limit, a total limit, a shutdown criterion.

How to Build Your Own Mini Strategy Portfolio

1. Start with one validated strategy. A portfolio of junk is still junk. Each component has to stand on its own: its own backtest, positive expectancy after costs, a mechanism you understand (risk-reward calculated, not just sensed).
2. Pick the second component for a different mechanism. Not "another trend strategy, but on a different indicator" — a different driver entirely: you have trend → add mean reversion; you have a continuous strategy → add a selective one that only plays a handful of times a month, in the best conditions. Selectivity is itself a form of diversification — a strategy that can choose NOT to play doesn't bleed capital in a regime it doesn't understand.
3. Split risk deliberately. Start simple, with parity: each strategy gets a similar risk budget (e.g., a maximum 3-4% monthly drawdown per strategy), not a similar amount of capital — a more volatile strategy gets smaller positions.
4. Measure correlation on an ongoing basis. A journal with per-strategy results, plus a quarterly correlation recalculation. If two components converge above ~0.7, you have a duplicate — cut it.
5. Set shutdown criteria in advance. For each strategy: at what drawdown (e.g., 1.5× the historical maximum) does it move into observation mode with no money behind it. A portfolio makes this decision psychologically easier — you're benching one engine, not the whole operation.

Numerical example (illustrative): strategy A (trend) and strategy B (mean reversion) each have about 15% annual volatility of results. Traded individually, each can show a drawdown of about 20%. At near-zero correlation, a 50/50 portfolio has volatility not of 15% but of roughly 10-11%, and shallower drawdowns — at a similar average return. That difference feels cosmetic until you live through a drawdown: -12% is something you can wait out according to plan; at -25% most people break their own rules. Strategy diversification is, in practice, buying discipline at the price of a bit of profit in the best-case scenario.

How Strefa Thinks About This — Process, Not Prediction

At Strefa we repeat this until it's boring: you can't consistently predict the market — you can consistently run a process. A portfolio of strategies is the institutional version of the same idea. Instead of asking "what's bitcoin going to do?", you ask: which mechanisms am I trading, how do I know they work, how much risk do I assign each one, and how will I know when one has stopped working. You replace forecasting with measurement and allocation. It's less exciting than "the big trade of the year" — and that's exactly why it works longer than a single season.

When It Doesn't Work and the Most Common Traps

• Pseudo-diversification. Three momentum strategies on three coins are one strategy. You can't fake correlation with names — calculate it.
• Correlations rise in a crisis. In a panic, "everything correlates with everything": strategies that looked independent in calm data can lose money together. A portfolio smooths out ordinary times; for the tails you need risk limits, not hope.
• Reallocating after results. Adding risk to a strategy after its best quarter and cutting it after its worst is buying the top of your own portfolio. Change allocation by rule, not by emotion.
• Too many components. Every strategy costs attention, monitoring and commissions. Two or three well-understood strategies beat ten mediocre ones — excess is dilution of edge and a rise in cost, exactly as with over-diversification in classic portfolios.
• A portfolio as an excuse. "I'm diversified" doesn't replace validating the components. The order is non-negotiable: first one strategy with an edge, then a second, then a portfolio.

The final lesson: the Grail exists, it's just not a strategy. It's a process — a set of independently validated edges, deliberately split risk, and rules for shutting down whatever stopped working. Individual strategies will come and go; a well-run portfolio of strategies is resistant to that by design, not by luck.

FAQ

Rafał — Krypto Bez Ściemy

Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.

kryptobezsciemy.plInstagramYouTubeKlub Zero