The Time Travel Workflow

Time Travel is QuanterLab's feature for doing rigorous walk-forward research without leaking the future into your decisions. It is the platform's answer to the single hardest discipline in quantitative research: separating "what I know now" from "what I would have known then."

The Problem It Solves

Every researcher carries the same invisible burden: you already know what happened. You know AAPL recovered after the 2018 selloff. You know which strategies worked in 2020 and which collapsed in 2022. That knowledge contaminates every choice you make — which timeframe to test, which tickers to include, which indicators to favor. The contamination is unconscious and unavoidable.

The result is a subtle, pervasive form of look-ahead bias: even strategies you believe were tested cleanly are quietly fitted to outcomes you already know.

How It Works

Click the Time Travel ribbon button and pick a date — say, 2022-06-01. The platform now treats that date as "today":

• All charts end at the chosen date.
• All scanners and screeners only see data up to that date.
• All backtests use only data prior to that date.
• Your saved configs and projects are tagged with the virtual-today date, so you can audit later what you decided when.

You research, build, and save strategies in this state. When you exit Time Travel, you have a dated, immutable record of what you would have committed to without seeing the future.

The Disciplined Workflow

1. Pick a virtual date. Far enough in the past that genuine forward data exists (≥ 6 months ago is a good minimum).
2. Enter Time Travel mode. Build, sweep, and save your strategy as if today were that date.
3. Lock and save. Once you are happy, save the config. The platform timestamps it with the virtual date.
4. Exit Time Travel. The forward data — from the virtual date to actual today — is now your genuine, untouched out-of-sample.
5. Run the saved config on the forward data. The result is the closest thing to live trading you can get from historical data.

Multi-Date Time Travel

The most powerful use of Time Travel is to repeat the workflow at several historical dates. Build a strategy "on" 2020-01-01, again "on" 2021-06-01, again "on" 2023-01-01. Each instance has its own clean OOS forward window. Across the three, you get a quasi-walk-forward view that includes the human element — your choices, not just optimized parameters.

If your strategies look very different at different virtual dates, your research process is regime-sensitive in a way pure walk-forward cannot reveal.

What Time Travel Cannot Fix

• Survivorship bias in your data sources. If your ticker list reflects today's composition (e.g., today's S&P 500), Time Travel still sees only those names. Use point-in-time index data when possible.
• Selection bias in what you test. If you only run Time Travel on strategies you already think will work, you are still cherry-picking. Time Travel is a tool, not a moral.
• Personal memory. If you genuinely remember what happened on the chosen date, the discipline is partial. Choose dates you don't have a strong memory of, or commit to acting only on platform output.

The Bottom Line

Time Travel is the closest a researcher can get to running a real-time experiment without waiting for real time. Used systematically, it converts subtle look-ahead bias into something you can audit, date, and trust. Used occasionally as a sanity check, it catches the kinds of overconfidence that quietly destroy edge.