Hybrid Trading Bots (MT5 + PocketOption)

This workspace now contains two independent bot paths:

• mt5_signal_bot.py: MT5 EUR/USD signal bot
• pocket_signal_bot/: PocketOption hybrid signal bot (paper + demo)

PocketOption Hybrid Signal Bot

The PocketOption bot is designed around:

• EMA + RSI signal engine (CALL / PUT / NO_TRADE)
• Risk controls (daily stops, max losses, max trades/day, min payout)
• Hybrid adapters:
  • Unofficial API adapter (primary)
  • Browser adapter fallback
• Structured event logs with timing fields

Files

• pocket_signal_bot/config.py
• pocket_signal_bot/strategy.py
• pocket_signal_bot/risk.py
• pocket_signal_bot/paper_simulator.py
• pocket_signal_bot/adapters/api_adapter.py
• pocket_signal_bot/adapters/browser_adapter.py
• pocket_signal_bot/runner.py
• pocket_signal_bot/logger.py

Install

pip install -r requirements.txt

If using browser fallback, also install browser runtime:

python -m playwright install chromium

Demo mode: fix “Could not read current price from UI”

1. Set in .env: PO_HEADLESS=false and PO_BROWSER_STARTUP_WAIT_SEC=15 (or higher on slow PCs).
2. Run python -m pocket_signal_bot.runner. A Chromium window opens on the quick-trading URL.
3. Log in if you see a login page; wait until the live quote is visible.
4. In normal Chrome on the same page: F12 → Elements → select tool → click the price number → right‑click element → Copy → Copy selector.
5. Put that string in .env as PO_PRICE_SELECTOR=... (use | between multiple selectors to try in order).
6. Run the bot again. If API still fails, browser candles will work once the selector matches.

The bot no longer exits the whole process when browser price read fails once; it logs data_error and retries on the next poll.

If the chart price is drawn on a canvas (no numeric <span>), keep PO_USE_WS_QUOTES=true (default). The browser adapter listens to the page WebSocket (e.g. update_quotes) and extracts the last price for PO_SYMBOL without needing PO_PRICE_SELECTOR.

Environment file (optional)

Copy the sample and edit:

copy .env.example .env

pocket_signal_bot/config.py loads .env automatically when python-dotenv is installed.

PocketOption credentials (demo first)

Set environment variables before running (or put them in .env):

• PO_SESSION -> session token from PocketOption web auth flow
• PO_UID -> your account uid
• PO_IS_DEMO=true
• PO_MODE=paper | demo | live (live = real money; requires PO_LIVE_CONFIRMED=true)
• Optional symbol/time controls:
  • PO_SYMBOL=EURUSD_otc
  • PO_TIMEFRAME_SEC=60
  • PO_EXPIRY_SEC=60

Run

Paper mode (safe):

set PO_MODE=paper
python -m pocket_signal_bot.runner

Demo mode (PocketOption demo account):

set PO_MODE=demo
set PO_SESSION=your_session
set PO_UID=your_uid
python -m pocket_signal_bot.runner

Live mode (real money — same code path as demo, but api_is_demo=false and real cabinet URL for browser):

set PO_MODE=live
set PO_LIVE_CONFIRMED=true
set PO_SESSION=your_real_session
set PO_UID=your_real_uid
set PO_IS_DEMO=false
python -m pocket_signal_bot.runner

The bot refuses to start PO_MODE=live unless PO_LIVE_CONFIRMED=true.

Event output goes to the terminal when PO_CONSOLE_LOG=true (no log files are written).

Trade history & PNG charts

Each run stores trades in data/trading_history.json. After every settled trade (if PO_CHARTS_AUTO=true), PNG charts are written to data/charts/:

File	Content
latest_pnl.png	Cumulative PnL + per-trade bars (current session)
latest_winrate.png	Win rate over trade number
compare_setups.png	Compare 15/15, 30/30, 60/60, M5/M3 sessions
dashboard.png	All sessions’ PnL curves overlaid

A/B test workflow — change Pocket Option + .env for each setup, run demo, stop bot (Ctrl+C), repeat:

# Example: 15s candle + 15s bet
PO_TIMEFRAME_SEC=15
PO_EXPIRY_SEC=15

Regenerate charts anytime:

python -m pocket_signal_bot.charts

Optional: PO_EXPERIMENT_LABEL=15/15 overrides the auto label used in comparison charts.

Readiness checklist before live

• At least 200+ demo trades logged
• Positive expectancy with realistic payout conditions
• No recurring adapter disconnect loop
• Latency and failover behavior verified in logs
• Risk limits tested (daily stop, max losses, max trades)

Important

No strategy guarantees profit. Treat this as research/automation infrastructure and validate on demo first.