Self-Trade Prevention (STP) & Wash Trading Filters: Avoiding API Account Bans

Wash trading—the illegal practice of simultaneously buying and selling the same financial instrument to create a misleading appearance of active trading volume—is strictly prohibited across all regulated and semi-regulated financial markets. However, in the world of automated crypto trading, wash trades frequently happen entirely by accident.

Consider a trader who deploys two independent bots on Bybit or Binance: Bot 1 is a Long Futures Grid trading SOL/USDT between $130 and $160; Bot 2 is a Short Futures Grid trading SOL/USDT between $145 and $175. Between $145 and $160, their trading corridors overlap completely.

If the price whipsaws rapidly across the $150 level, Bot 1 may place a limit sell order at $150.20 while Bot 2 simultaneously submits a limit buy order at $150.20. When the exchange matching engine processes these orders under the same Master Account UID, the buy order fills directly against the sell order. You have just executed a wash trade with yourself, paying double taker fees while triggering exchange compliance alarms.

Exchanges like Binance, Bybit, OKX, and Coinbase are under intense scrutiny from global regulators to eliminate fake volume and market manipulation. Their surveillance departments deploy automated algorithms that flag accounts generating high ratios of self-executed volume.

The consequences of self-trade violations are severe:

• Fee Penalties: Self-trades typically incur full Taker fees on both sides of the transaction, resulting in immediate financial loss.
• Maker Rebate Forfeiture: If you operate on an institutional VIP tier with Maker fee rebates, self-trades are disqualified from rebate calculations and can trigger tier downgrades.
• API Blacklisting: Repeated violations trigger automated security bans on your API keys or complete account termination for suspected market manipulation.

To protect legitimate quantitative traders from accidental collisions, modern exchange APIs support Self-Trade Prevention (STP) parameters (e.g., stpMode in Bybit V5, selfTradePreventionMode in Binance). Traders specify how the matching engine should behave when an incoming order would trade against an existing resting order from the same account:

• Cancel Maker (CM / EXPIRE_MAKER): The existing resting limit order on the book is cancelled immediately, and the incoming aggressive order is allowed to execute against remaining public liquidity.
• Cancel Taker (CT / EXPIRE_TAKER): The incoming aggressive order is rejected or cancelled, preserving the resting limit order on the book. (This is the preferred mode for grid bots).
• Cancel Both (CB / EXPIRE_BOTH): Both the resting limit order and the incoming order are cancelled simultaneously.

While exchange-side STP provides a critical safety net, institutional algorithmic infrastructure enforces Client-Side Collision Prevention before transmitting packets over the wire:

1. Unified Local Order Book Cache: Maintain a local in-memory ledger tracking all active resting bids and asks across all running bot instances on that account.
2. Pre-Flight Cross Inspection: Before submitting a new buy order at price P_{buy}, check if P_{buy} \ge \min(P_{sell}) among active sell orders. If a cross is detected, refuse submission and alert the strategy manager.
3. Sub-Account Partitioning: Run opposing strategies (e.g., Long Grid and Short Grid) on distinct sub-accounts with separate API keys to mathematically eliminate internal cross-matching.

Follow these implementation standards across your automated trading codebase:

• Bybit V5 API: When submitting orders via POST /v5/order/create, always include the parameter "smpType": "CancelTaker" (Self-Match Prevention).
• Binance Spot & Futures API: Pass "selfTradePreventionMode": "EXPIRE_TAKER" in your order creation payload.
• OKX Unified API: Include "stpMode": "cancel_taker" in order parameters.
• Continuous Audit Logging: Log every STP cancellation event in your monitoring dashboard. A spike in STP triggers indicates faulty grid boundary calibration that requires immediate human attention.