The Illusion of Three Profitable Bots
Most traders think running three profitable bots means 3x the profit. It doesn't. Most of the time, it means one bot with 3x the risk—and all three fail at once.
You've tested all three separately. Each one returns 8-15% on their strategy. So why, when you run them together on a live account, does a single market shock wipe out weeks of gains in hours?
That's correlation.
What Correlation Actually Is (and Why It Breaks Portfolios)
Correlation is the degree to which two assets or strategies move together. If all three bots respond to the same market condition—say, a sudden spike in volatility or a Fed announcement—they're correlated.
Here's the problem: your backtest tested each bot independently. It never tested what happens when all three respond to the same catalyst at the same time.
Example. Bot A: trend-following on EURUSD. Bot B: trend-following on GBPUSD. Bot C: trend-following on AUDUSD. All three catch big USD moves. When the dollar reverses, all three exit in the same direction. That's not diversification. That's the same bot running three times with different currency pairs.
The math is brutal. Instead of distributing risk, you've stacked it.
Why Backtests Hide Correlation
Your backtest shows three bots doing well. Each bot's equity curve is smooth. No drawdown above 12%. So you deploy.
Then March 2020 hits. Or August 2024. Or the next black swan.
The backtest didn't test correlation because it tested each bot in isolation against historical data. History doesn't repeat volatility spikes. Live markets do.
When a real shock hits, all three bots respond to the same fear. They liquidate the same positions. They trigger the same stop-losses. The 3x returns you expected become 3x losses in minutes.
This is why pros don't test strategies in isolation. They test the portfolio as a system.
The Dollar Cost of Correlation
Let's put a number on it. Say you run three bots on a $30,000 account. $10,000 per bot. Each bot is designed to risk 2% per trade.
When correlation hits, all three don't risk 2% each. They risk 6% at once—simultaneously. That's not three 2% risks. That's one 6% risk. On a $30,000 account, one 6% move drops you to $28,200. Two correlated moves in a row (they tend to cascade) and you're at $26,400. Four moves and your account is halfway gone.
The traders who survive understand this: correlation multiplies your losses faster than it multiplies your wins.
How Professionals Build Multi-Bot Portfolios
Pros don't diversify by running similar strategies on different assets. They diversify across three dimensions:
- Asset class—don't run all three on currencies. Run one on forex, one on commodities, one on indices. These don't move together in a crisis.
- Time frame—don't run all three on 4-hour charts. Run one on daily, one on 4-hour, one on 15-minute. Different timeframes capture different moves.
- Strategy logic—don't run trend-following on all three. Mix trend, mean-reversion, and breakout. When trends fail, mean-reversion often thrives.
Now correlation is actually low. When crypto crashes, your forex bot might trend up. When indices pull back, your commodities bot catches a fear bid. They're not moving together—they're offsetting.
When Correlation Kills: Real-World Examples
March 2020: Every risk asset crashed together. Stocks, forex, commodities all correlated at 0.95 (nearly perfect). Traders who thought they were diversified across three asset classes got destroyed because the correlation breakdown happened in a black swan, not in backtest.
August 2024: BoJ surprise tightening tanked the yen carry trade. Every bot that relied on JPY weakness failed at the same moment. Traders with three "different" strategies on JPY pairs all blew up on the same day.
The pattern: stress events expose hidden correlation. Your backtest never tested a stress event because backtesting data doesn't include the emotional panic that drives real crashes.
Three Signals of a Truly Uncorrelated Portfolio
Before you deploy a multi-bot system, check these:
- Win/loss timing is offset—if Bot A wins on Monday and Bot B loses on Tuesday (not both Monday), correlation is low. If all three win or lose on the same days, correlation is high.
- Drawdowns don't compound—run a stress scenario where the market moves 5% against all three. Does one bot handle it while the others stay flat, or do all three bleed? Safe systems have staggered drawdowns.
- Liquidation paths diverge—ask: if volatility spikes 50%, do all three exit the same position, or do they exit different ones? Same exit = high correlation. Different exits = low correlation.
If you can't answer these three questions about your multi-bot system, correlation risk is hiding in your portfolio.
How Alorny Handles Multi-Bot Correlation
This is exactly what separates custom-built portfolios from template bots. When we build multi-bot systems, we test correlation across live market conditions, not just historical backtest data.
We run stress scenarios where all three strategies face the same market shock—and we rebuild them until they respond differently. One sells, one holds, one buys. That's uncorrelated.
Each bot in a portfolio gets its own asset class, its own time frame, and its own strategy logic. No cut-and-paste. No hoping correlation stays low. We build for stress.
A custom multi-bot portfolio starts at $300 per bot (simple divergence strategies) to $500+ per bot (AI-based uncorrelated systems). Most traders spend way more than that discovering correlation the hard way—on a live account.
Message us your three bots and we'll audit them for correlation risk. We'll show you the exact portfolio structure that survives the next market shock.