Summary

After a suspected platform update in early-to-mid December 2025, my algorithm's warm-up process is being skipped or completed almost instantly, causing no trades in live deployment while backtests work correctly.

Timeline

  • October 2025 - Early December 2025: Algorithm ran live continuously with normal warm-up behavior (3-5 minutes, progress bar showing 0% → 100% gradually)
  • Mid-December 2025 (exact date unknown): Warm-up started completing in under 1 minute, progress bar now jumps from some random percentage (e.g., 54%) directly to 100%

 

Symptoms

Aspect Before (Normal)After (Broken)Warm-up duration 3-5 minutes< 1 minute Progress bar Gradual 0% → 100% Jumps X% → 100% instantly Live trading Trades generated No trades at all Backtesting Trades generated Trades generated (works fine)

armitage_1766625293.jpgarmitage_1766625586.jpg

Critical Evidence

  1. I have an old project【P2】 that still warms up correctly (Stop → Redeploy → still works)
  2. I cloned this exact project【P1】 (identical code, same account, same settings)
  3. The cloned project exhibits the broken behavior - warm-up completes instantly

This proves the issue is NOT in my code, but something at the platform/project level that Clone does not preserve.

 

My Warm-up Logic

  1. # Two-loop warm-up pattern (worked for months): # Loop 1: Fill RollingWindow with 2880 bars history = self.History[TradeBar](symbol, 2880, Resolution.Minute) for bar in history: symbol_data.price_window.Add(bar.Close) # RollingWindow[float](2880) # Loop 2: Calculate indicators for i in range(2880): self.indicator_engine.calculate_all_indicators(symbol_data)

 

One More Important Point:

This issue started appearing around December 9th, 2025. I strongly suspect a recent platform update caused this.

Before Dec 9: All my algorithms using this manual History warmup pattern worked perfectly ”indicators warmed up correctly, live trades generated normally.