Build an Auto-Optimizing Trading System — Parameter Loop + Telegram Notifications

A strategy that never re-evaluates its parameters is a strategy that slowly dies. Markets shift, volatility changes, and what worked last month may not work today.

The solution? A workflow that periodically re-optimizes itself and sends you a Telegram notification with the results — all running automatically on ApudFlow.

The Complete Workflow

[Schedule Trigger] → [Fetch Prices] → [Swing Finder] → [Backtest] → [Parameter Loop]
                                                                              │
                                                                              ▼
                                                                      [Telegram Notify]

This workflow runs every day at market close, sweeps parameter combinations, finds the best one, and messages you the results on Telegram.

Why This Matters

Without automation: You manually run sweeps, compare results, update parameters, re-deploy. The whole cycle takes 30 minutes and you probably only do it once a week.

With this workflow: The system re-optimizes daily and notifies you. You just read the message and know your strategy is running with today's best parameters.

Step-by-Step Build

Step 1: Schedule Trigger

Add a Schedule Trigger that runs daily at a specific time:

{
  "everySeconds": 86400,
  "type": "interval",
  "runAt": "22:00 UTC"
}

This triggers the workflow once per day — after the markets close, when fresh OHLC data is available.

Step 2: Fetch Prices

Add Fetch Prices to pull the latest market data. Use enough bars for meaningful backtest results:

{
  "symbols": ["XAUUSD", "EURUSD"],
  "timeframe": "1h",
  "limit": 1000
}

Step 3: Swing Finder + Backtest

These are your strategy workers — exactly as you'd configure them for manual sweeps. The Swing Finder detects swing points, and the Backtest evaluates performance with the parameters that the Parameter Loop provides each iteration.

Step 4: Parameter Loop — The Optimizer

This is the heart of the system. Configure it to find the best parameters every time it runs.

Configuration:

{
  "sweepParameters": [
    {"name": "stop_loss",   "values": [1.0, 2.0, 3.0, 4.0, 5.0]},
    {"name": "take_profit", "values": [2.0, 3.0, 4.0, 5.0, 6.0]}
  ],
  "collectField": "result.sharpe_ratio",
  "rankingField": "sharpe_ratio",
  "rankingMode": "max",
  "prodMode": "always_run",
  "sessionLabel": "daily_optimization",
  "maxIterations": 0,
  "continueOnError": true
}

Key detail: prodMode is set to always_run. This means the full sweep runs every single time the workflow triggers — daily in this case. The sweep re-evaluates all parameter combinations against the latest market data.

Step 5: Telegram Notification

Add a Telegram Notify worker connected to after the Parameter Loop. This worker reads the sweep results and sends you a concise summary.

Configuration:

Message template:
📊 DAILY OPTIMIZATION COMPLETE
━━━━━━━━━━━━━━━━━━
Strategy: Gold & EURUSD Swing
Date: {{trigger_1.now}}

Best Parameters:
• Stop Loss: {{parameter_loop_1.bestParameters.stop_loss}}%
• Take Profit: {{parameter_loop_1.bestParameters.take_profit}}%

Results:
• Sharpe Ratio: {{parameter_loop_1.bestValue}}
• Total Iterations: {{parameter_loop_1.totalIterations}}
• Successful: {{parameter_loop_1.successfulIterations}}
• Failed: {{parameter_loop_1.failedIterations}}

Best iteration #{{parameter_loop_1.bestIteration}}

What the Telegram message looks like:

📊 DAILY OPTIMIZATION COMPLETE ━━━━━━━━━━━━━━━━━━━━ Strategy: Gold & EURUSD Swing Date: June 13, 2026

Best Parameters: • Stop Loss: 2.5% • Take Profit: 4.0%

Results: • Sharpe Ratio: 1.87 • Total Iterations: 25 • Successful: 23 • Failed: 2

Best iteration #7

Step 6: (Optional) Auto-Apply Best Parameters

If you want the workflow to automatically use the best parameters for the next run, the Parameter Loop's use_best mode combined with session saving handles this:

1. The sweep runs (prodMode: always_run)
2. Results save to session daily_optimization
3. On the next day's run, if you switch to prodMode: use_best, it loads the previous day's best without re-sweeping

But for true auto-optimization, keep always_run and let the system re-evaluate daily.

Advanced: Multiple Notification Levels

You can create a richer notification system by splitting the output:

[Parameter Loop] ─┬──▶ [Telegram: Summary]  (short, daily digest)
                   │
                   └──▶ [Condition] ──▶ [Telegram: Alert] (only if Sharpe dropped)

Condition Configuration

Check if the best Sharpe ratio fell below a threshold:

Condition: {{parameter_loop_1.bestValue}} < 1.0

If true → send an alert: "⚠️ Strategy degrading — Sharpe dropped below 1.0. Consider reviewing parameters."

If false → send the normal daily summary.

Real-World Scenario

Imagine it's June 2026. Gold has been trending strongly for weeks with low volatility. Your strategy was optimized with stop_loss: 2.5 and take_profit: 4.0.

Suddenly, volatility spikes. The Fed makes an unexpected announcement.

Your daily optimization run catches this immediately:

Day	Best SL	Best TP	Sharpe
June 10	2.5%	4.0%	1.87
June 11	3.5%	4.0%	1.65
June 12	4.0%	5.0%	1.42
June 13	4.0%	6.0%	1.38

Without daily optimization: Your strategy keeps using SL=2.5 in a high-volatility market, getting stopped out repeatedly.

With daily optimization: The system automatically widened the stop loss from 2.5% to 4.0%, adapting to the new regime. Your Telegram shows the progression daily.

Production Considerations

Rate Limiting

Daily sweeps (25 iterations) on 1h data consume minimal compute. If you're optimizing every hour, consider reducing the grid size or increasing maxIterations as a safety cap.

Session Storage

Each daily run saves a new session. After a year, you'll have 365 sessions — useful for trend analysis. You can clean old sessions from the Sessions tab.

Telegram Bot Setup

If you haven't set up Telegram notifications:

1. Create a bot via @BotFather on Telegram
2. Copy the bot token
3. Get your chat ID by sending a message to the bot and visiting https://api.telegram.org/bot<TOKEN>/getUpdates
4. Enter both in the Telegram Notify worker configuration

Summary

Component	Purpose
Schedule Trigger	Runs daily at market close
Fetch Prices	Gets fresh OHLC data
Swing Finder + Backtest	Evaluates strategy with current parameters
Parameter Loop	Sweeps all combos to find the best
Telegram Notify	Sends you the results in a formatted message

This workflow turns strategy optimization from a manual weekly chore into a fully automated daily process. Set it up once, and let the system keep your strategy aligned with changing market conditions.