Diversification in copy trading: how many traders and how much each

One allocation is a bet on one person's risk management, one strategy, and one period of market history. Adding a second name does not automatically make you safer - it makes you safer only if the second bet is not the first bet.

The single-name failure mode

The most common way to lose money in copy trading is not picking a bad trader. It is picking one good trader and being present for their worst month, with all of your capital. When a grid unwinds, or a leverage day arrives, one account carries 100% of your loss - and no amount of that trader's quality changes the outcome.

Diversification is not about return. It is about making sure that the account you are wrong about is not the whole portfolio.

How many names

The practical range is 3 to 6. Below three, one mistake dominates. Above six, you spread across so many that a single bad pick costs little - but you also lose the ability to notice that something is wrong, and you end up with a portfolio of accounts you have not read.

Bybit allows a limited number of simultaneous subscriptions per account; the current limit is 10, which is enough for this range and not enough to make "many small bets" a strategy on its own.

How to split

Two rules do most of the work:

1. Cap any single name at 10–15% of the copy allocation. Above that, one drawdown becomes your problem rather than your data.
2. Weight by risk, not by return. A 12% drawdown account and a 45% drawdown account that show the same ROI should not get the same size.

A worked example with 5,000 USDT of copy capital and a pool of six names:

Trader	Risk profile	Share	USDT
A	low drawdown, 14% max DD	20%	1,000
B	low drawdown, 18% max DD	20%	1,000
C	mid, 28% max DD	20%	1,000
D	mid, 32% max DD, high Sharpe	15%	750
E	aggressive, 55% max DD	15%	750
F	backup slot, currently subscribed	10%	500

The aggressive slot is capped at 15% not because it is bad, but because its outcome range is wider. The backup slot exists so you always have capacity to move into a better name without waiting for something to fail.

Styles, not rankings

Correlated names are not diversification. Three accounts that all trade the same token with similar leverage will fail together. The grouping that actually helps:

• one low-volatility swing account - the ballast,
• one trend follower - profits in sustained moves,
• one mean-reversion account - profits in ranges, bleeds in trends,
• one longer-horizon account - provides the idle-time diversifier,
• one small aggressive slot - optional, sized as an option rather than a core.

Our low-risk screens (/low-drawdown/, /risk-adjusted/) make this easier: pick from accounts whose volatility and drawdown are already in a range you are comfortable with, then differentiate by style.

What diversification does not fix

• Systemic moves. A liquidation cascade hits correlated accounts together. Diversification reduces idiosyncratic loss, not market loss.
• Copy-trading friction. Every name you add multiplies your own operational load: PCSL settings, withdrawals, and reviewing eight pages instead of three.
• The mistake of averaging in. Adding a fifth name to a portfolio that is already down 20% is not diversification.

How to review the portfolio

Once a quarter, look at the portfolio as one account:

• recompute the combined drawdown - it is deeper than the average of the parts,
• check whether the names are still doing what they did when you picked them (style drift is the most common portfolio failure),
• re-run the low-risk screens: accounts qualify into a screen; they do not stay in it automatically,
• withdraw profit on schedule regardless of how the portfolio is doing.

A portfolio of copy traders behaves like a small leveraged fund with no risk process. The only process it has is the one you impose on it - which is why the allocation table above has numbers in it rather than adjectives.