Skip to content

Configuration — the .env file

Updated: June 2026. The .env is organized in three blocks: operational (API, risk, monitoring), ACTIVE STRATEGIES (validated parameters — do not touch without re-validating) and DISABLED STRATEGIES (rejected on the data, verdict in the comments). The bot reads the .env ONLY at startup: every change requires a restart.

Operational block

# Deribit
DERIBIT_API_KEY=...           # never commit; the dashboard never shows them
DERIBIT_API_SECRET=...
DERIBIT_ENV=test              # test | live

# Risk management (see 05_risk_sizing.md)
INITIAL_EQUITY=10000
BASE_RISK_PCT=0.01            # 1% risk per tactical trade
MAX_DAILY_LOSS_PCT=0.03       # kill switch after -3% daily
MAX_OPEN_TRADES=3             # max open positions on the venue
MAX_GROSS_EXPOSURE=1.5        # aggregate gross notional max 1.5x equity

# Monitoring
MONITORING_INTERVAL_MINUTES=15   # signal scan frequency
HEALTH_CHECK_INTERVAL_SEC=5
MAX_API_DOWN_SEC=30              # emergency close after 30s of API down
LOG_LEVEL=INFO

# Telegram (optional, empty = disabled)
TELEGRAM_BOT_TOKEN=
TELEGRAM_CHAT_ID=

Active strategies

Each strategy is enabled with TB_ENABLED / FS_ENABLED / MC_ENABLED and has its own parameter block in the .env (prefixes TB_, FS_, MC_). The defaults are the values validated over the 4 years: changing a parameter means leaving the validation → re-run the pipeline before trusting any result.

Multi-symbol

TB_SYMBOLS / FS_SYMBOLS / MC_SYMBOLS accept a list (BTCUSDT,ETHUSDT): the bot creates ONE INSTANCE per symbol with the derived Deribit instrument (ETHUSDT → ETH-PERPETUAL). Each side of each strategy is enabled only on the symbols where validation passed.

🔒 The complete strategy parameter tables (lookbacks, thresholds, multipliers and their validated ranges) are proprietary — available under the commercial license (contact: lantoniotrento@gmail.com). The dashboard editor still exposes them to the operator with the validated ranges as bounds.

Disabled strategies

All with *_ENABLED=false and the validation verdict in the .env comments. Do not re-enable without a full new validation.

VB_ENABLED=false        # Volume Breakout : PF 0.42-0.74 — no edge
MR_ENABLED=false        # Mean Reversion  : PF 0.28-0.53
LIQ_ENABLED=false       # Liq Squeeze     : PF 0.06-0.30
IS_ENABLED=false        # Imbalance Scalp : fee-bound, PF 0.18-0.50
BRINGS_ENABLED=false    # NY Brings       : PF 0.64, negative every year
WM_ENABLED=false        # W/M Formation   : non-structural edge
STRATEGY_SMART_MONEY_ENABLED=false
STRATEGY_IRON_CONDOR_ENABLED=false   # options, out of project scope

Notes

  • Config.load_strategies() is the single point that translates the .env into instances: the dashboard uses the same function, so it sees exactly the instances the bot builds.
  • Testnet: sizing equity is capped at $50k (testnet faucet funds inflate equity and would produce out-of-scale orders).
  • Config validation at startup: missing keys or an invalid DERIBIT_ENV block the bootstrap.