Why Uncorrelated Markets Cut Your Risk! (The Math No One Explains)

New comment
 
Most traders chase a better entry. Fewer think about something that matters far more for survival: how their positions relate to each other. If everything you trade tends to lose at the same time, you don't have a portfolio — you have one big bet wearing several costumes.

Here's the idea that changed how I build systems.

— Correlation is the hidden risk multiplier —

Imagine two markets, each with a solid edge, but they move together. When one is in a losing stretch, the other usually is too. Their drawdowns stack on top of each other, and your account feels the full weight of both at once. Adding the second market barely helped and you just doubled the size of the same risk.

Now imagine two markets with the same edge, but they move independently. When one is having a bad month, the other is often flat or up. Their losing stretches land at different times and partly cancel out. You keep most of the combined return, but the combined drawdown is much smaller than either one alone.

That's not an opinion. It's just how variance works when you add independent streams together.

— The part people miss —

When you combine uncorrelated return streams, the returns add up faster than the risk does. Return grows roughly in proportion to how many streams you add. Risk (drawdown, volatility) grows much more slowly, because the losses don't line up. The result is a smoother equity curve and a higher "return per unit of drawdown" , the number that actually decides whether you can hold a system for years.

A single market can't give you that at any risk setting. You can only get it by combining things that genuinely move differently.

— Why "genuinely" is the hard part —

Here's the catch:
most markets are more correlated than they look, especially when it matters. In a panic, "different" assets often crash together! Correlations snap to 1 exactly when you needed them low. So the diversification has to be real and structural, not just a list of different symbols!!

That usually means mixing things driven by different forces : A metal , an equity index , a currency cross , maybe a crypto — instruments that respond to different fears and different flows. The goal isn't more markets for the sake of it. It's markets whose bad days don't share a calendar.

— Why this matters even more for prop accounts —

On a funded or prop account, a smoother curve isn't just nicer. It's the difference between passing and blowing the drawdown limit. A 'single market' system depends entirely on the mood of one market during your evaluation window. Spread the same edge across uncorrelated markets and the odds of a clustered drawdown — the kind that trips a daily or total limit — drop sharply.

— The takeaway —

Don't just ask "does this market have an edge?" Ask "when this one is bleeding, what else am I holding, and is it bleeding too?"
The traders who survive aren't the ones with the single best market. They're the ones whose markets don't all fall apart on the same day.

Diversification isn't a slogan. Done honestly, it's the closest thing trading has to a free lunch.
 

This is an age-old debate─trading a single instrument versus trading a basket of uncorrelated instruments that are indirectly hedged.

Comprehensive exit logic makes all of the difference in the world when trading a single instrument (this assumes that we're not dealing with any Martingale/direct hedging nonsense). Even though I generally implement a dynamic exit, it's always backed up by a fixed stop placed a reasonable distance away from my entry. I refer to that as my fail-safe stop, or my "oh $@#%!" stop. Of course, nothing is truly fail-safe. The idea is that the dynamic exit (a market order) and the fixed stop (a working/pending order) complement each other. Even in a super hot market where fixed stop execution could get jumped, the market order is resent. As a result, I have black swan event mitigation.

With a basket trading strategy, you really have to implement similar "master" exit logic─close everything. I've been trading long enough to know that a black swan event can cause everything to crap the bed at the same time, and in the blink of an eye. You could say that, in that moment, the uncorrelated becomes correlated. Correlation ratios remain fairly steady, until they don't. Without that master exit logic, a basket could turn into a basket case.

The main benefit of the aforementioned single instrument trading is the fact that the strategy can be nimble enough to turn on dime and enter in the opposite direction (in a detrended strategy, that is). At that point, the black swan turns rainbow colored. Jumping right back into basket trading immediately following, or during, a black swan event is more problematic.
 
Yes, studying correlation first before trades is important when basket trading. Especially keeping an eye on currency strengths of all relevant pairs. A basket trades setup can turn sour pretty fast if we don't keep watching.
 

Mathematically speaking, diversifying between different instruments gives equal or less risk than the individual risk. I demonstrated this some years ago at uni. Markowitz theory talks a lot about this. Perhaps looking at it could help, but truly understanding it takes a lot of deep applied math understanding: multi-objective optimization, calculus, etc.

Just a simple inequality.

This is the basis for the theory, the sum of weighted variances of each instrument is greater than or equal to the portfolio variance.

 
Ryan L Johnson #:

This is an age-old debate─trading a single instrument versus trading a basket of uncorrelated instruments that are indirectly hedged.

Comprehensive exit logic makes all of the difference in the world when trading a single instrument (this assumes that we're not dealing with any Martingale/direct hedging nonsense). Even though I generally implement a dynamic exit, it's always backed up by a fixed stop placed a reasonable distance away from my entry. I refer to that as my fail-safe stop, or my "oh $@#%!" stop. Of course, nothing is truly fail-safe. The idea is that the dynamic exit (a market order) and the fixed stop (a working/pending order) complement each other. Even in a super hot market where fixed stop execution could get jumped, the market order is resent. As a result, I have black swan event mitigation.

With a basket trading strategy, you really have to implement similar "master" exit logic─close everything. I've been trading long enough to know that a black swan event can cause everything to crap the bed at the same time, and in the blink of an eye. You could say that, in that moment, the uncorrelated becomes correlated. Correlation ratios remain fairly steady, until they don't. Without that master exit logic, a basket could turn into a basket case.

The main benefit of the aforementioned single instrument trading is the fact that the strategy can be nimble enough to turn on dime and enter in the opposite direction (in a detrended strategy, that is). At that point, the black swan turns rainbow colored. Jumping right back into basket trading immediately following, or during, a black swan event is more problematic.
Really well put and you're pointing at the exact failure mode that matters. You're right that "uncorrelated becomes correlated" precisely when it hurts most. correlation ratios hold steady until a black swan snaps them to 1 in the blink of an eye. Diversification smooths the normal times, not the crash.
That's why I'd argue a basket approach is only as safe as its master exit. For me that means an account level "kill switch" that watches total equity, not individual positions. if aggregate drawdown hits the limit, everything gets flattened at once, regardless of what any single instrument is "supposed" to do. Close everything, ask questions later. Without that, you're right, a basket can quietly turn into a basket case.
But more importantly :
A "close everything" switch is only worth having if the underlying edge is real, with a recovery factor and Sharpe strong enough to climb back out after the hit. Otherwise the kill switch just fires over and over, and you bleed the account to death by a thousand "safe" exits!
The exit protects you from the catastrophe. only a genuine, "regime tested" edge brings the account back afterward. One without the other is a trap.
I love your point about the fail safe stop backing up the dynamic exit . the market order resend for when a resting stop gets jumped in a hot market is exactly the kind of black-swan mitigation people skip until it burns them once. The dynamic plus fixed pairing is underrated.
The single instrument nimbleness point is fair too. A basket trades resilience for agility. it survives better but pivots slower. Different tools for different temperaments. Appreciate you taking the time to write this out. 🙏🏼
 
DonEps #:
Yes, studying correlation first before trades is important when basket trading. Especially keeping an eye on currency strengths of all relevant pairs. A basket trades setup can turn sour pretty fast if we don't keep watching.
Absolutely. The one nuance I'd add is :
Orrelations aren't static. So instead of trying to monitor currency strength and pair relationships in real time, I prefer to pick markets that are structurally uncorrelated to begin with . You know? different asset classes driven by different forces, not just different pairs of the same currencies. That way I'm not depending on a relationship staying favorable; the independence is baked in. Still needs watching, but there's less that can quietly drift against you. Good point though.. A basket you set and forget is a basket that eventually surprises you 😁
 
Isaac Uriel Arenas Caldera #:

Mathematically speaking, diversifying between different instruments gives equal or less risk than the individual risk. I demonstrated this some years ago at uni. Markowitz theory talks a lot about this. Perhaps looking at it could help, but truly understanding it takes a lot of deep applied math understanding: multi-objective optimization, calculus, etc.

Just a simple inequality.

This is the basis for the theory, the sum of weighted variances of each instrument is greater than or equal to the portfolio variance.

Exactly right.
This is Markowitz / Modern Portfolio Theory at its core, and that inequality (σp ≤ Σ Wi·σi) is the whole engine behind the article.
I kept the piece deliberately intuitive because most traders will never sit through the applied-math derivation multi-objective optimization and calculus scare off the very people who'd benefit most but you're pointing at the rigorous foundation underneath it :
The portfolio's risk is strictly less than the weighted sum of the parts, and the gap widens as correlation drops (equality only holds at perfect +1 correlation).
The one practical footnote I'd add to the theory :
Markowitz assumes the covariance structure is stable. In live markets it isn't correlations drift, and in a crisis they collapse toward 1, which is exactly when the diversification term in that inequality vanishes and σp approaches the weighted sum. So the math gives you the edge in normal regimes, and a hard account-level risk limit has to cover the moments the assumptions break.
Beautiful theory but it needs a seatbelt in practice. Great comment, thanks for adding the rigor.
New comment