A historically profitable strategy can fail on its first operating day because a feature arrives late, an order is duplicated or the paper fill model assumes liquidity that never existed. Staged deployment connects backtest evaluation to prospective evidence while limiting the capital exposed to unresolved assumptions. This lesson defines the distinct claims tested by replay, shadow operation, paper trading and limited live execution, then builds a reconciliation and promotion policy around those claims.
Learning objectives. By the end of this note you should be able to:
- distinguish prospective signals from simulated and actual executions;
- freeze a reproducible research version before collecting promotion evidence;
- reconcile differences through data, signal, portfolio, order, fill and cash layers;
- write observable stage gates and rollback conditions;
- explain why a short successful trial does not prove either capacity or durable alpha.
In this lesson
- Separate the claims being tested
- Freeze what the trial means
- Find the earliest point of divergence
- Reconcile economic shortfall
- Write gates that can fail
- Scaling and rollback change the risk profile
Before you start
The full lesson, with worked examples and exercise solutions, is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Quantitative Research, Data and Backtesting syllabus.