Why sub-accounts matter for risk management
Isolated balance
Per-strategy P&L clarity
Experimental safety
One venue, multiple bots
When to use sub-accounts
Testing a new mode
Testing a new mode
- Create a sub-account on Binance.
- Generate API key for the sub-account (same security disciplines).
- Spin up a second TradingBot container with the sub-account credentials.
- Allocate
$1,500–$3,000to the sub-account. - Run Tsl2Sell on the sub-account for 4+ weeks.
- If it works, scale up. If not, only the small sub-account capital was at risk.
Testing a new symbol
Testing a new symbol
Testing a SignalEditor recipe
Testing a SignalEditor recipe
Strategy diversification within one venue
Strategy diversification within one venue
Family / partner account separation
Family / partner account separation
When NOT to use sub-accounts
❌ For first-month operators
❌ For first-month operators
❌ When you don't have enough capital to fragment
❌ When you don't have enough capital to fragment
$10,000 into 4 sub-accounts of $2,500 each — most modes won’t operate effectively at that size.Consolidate at lower capital; fragment at higher capital.❌ As substitute for proper backtesting
❌ As substitute for proper backtesting
How sub-accounts work in unCoded’s architecture
One TradingBot per sub-account
One TradingBot per sub-account
Per-sub-account API keys
Per-sub-account API keys
- Spot trading + read only.
- Withdrawals NEVER enabled.
- IP allowlisted to your VPS.
- 2FA on the master account (sub-accounts don’t have separate 2FA on most venues).
- Quarterly rotation across all sub-account keys.
Per-sub-account TelegramBot (optional)
Per-sub-account TelegramBot (optional)
Setting up sub-accounts
Create a sub-account at the venue
experimental-tsl2sell).Fund the sub-account
$1,500–$3,000 for experimental strategies.Generate API key for the sub-account
- Spot trading + read.
- No withdrawals.
- No internal transfers.
- IP allowlisted to your VPS.
Spin up a second TradingBot container
EXCHANGE_ID configuration.Configure the experimental setup in the Dashboard
Run for the validation window
4 weeks minimum for most experiments. Watch behavior daily for the first week, weekly thereafter.