Futures / Original case research
Futures Replay, Queue Position and Fill Evidence
Use replay to inspect a rule without presenting historical touches or simulated orders as subscriber execution.

A replay answers a bounded software question
Replay can help inspect when a documented condition is evaluated against retained data. It cannot by itself prove how a live subscriber's order was routed or filled. NinjaTrader's alert guidance notes that alert behaviour in accelerated playback has limitations. That is a platform-specific observation, not a universal speed threshold for every replay engine. Preserve mode, data granularity, time speed, rule version and settings before comparing outputs.
A chart touching a limit price does not establish that the account had a queued order or sufficient executable quantity. Trade events and quotes supply different evidence. If a simulator makes a fill assumption, record it explicitly. Avoid silently improving it after seeing the favourable exit. A result may legitimately be called a simulated or research result; the label must remain visible wherever the number travels, including machine-readable records.
Historical and live states remain separate
A platform may initialise strategy objects differently in historical and real-time modes. Retain the transition and do not treat historical order identifiers as proof of a later actual order. For each account claim, request actual acknowledgements and executions. Unknown queue position cannot be reconstructed from a single candle. A replay can still expose duplicate conditions or configuration problems even when fill verification remains outside its evidence.
Freeze the evaluation convention
Specify available data, latency assumptions, touch or trade-through treatment, quantity and costs before inspecting outcomes. Keep unsuccessful, cancelled and unfilled cases. A perfect hypothetical entry selected from a completed window should not be renamed a first-available broker fill. Historical research conventions deserve their own identity and can be studied without weakening them. The key requirement is that their best-price or other conventions remain distinct from actual customer transactions.
What to publish after a replay review
Report the conditions reproduced, settings that mattered, missing records and the simulation rules. Do not append a live-performance badge merely because software tests passed. A feature regression test is not an independent strategy-validation sample. If an author later supplies a consecutive real-time observation ledger, evaluate it as another evidence object with its own period. The original replay remains useful for diagnostics rather than retroactively becoming a real account history.
Completed fictional record
A fictional candle spans 100 to 105 and a simulator awards entry at 100 and exit at 105. The underlying sequence, queue, order submission and execution receipts are absent. The completed review can reproduce the candle-based convention but cannot establish a five-point customer gain. A second convention that requires a documented order and fill would leave that trade unresolved, not automatically losing. The two conventions answer different questions. The reviewer preserves the research reference, lists the missing execution fields and makes no claim that either platform or author committed fraud.
| Recorded object | Value or state | Interpretation |
|---|---|---|
| Candle convention | Five-point hypothetical move | Research result |
| Order acknowledgement | Not provided | Unknown |
| Actual fills | Not provided | No subscriber return |
Primary references and scope
Public documents inspected 13 October 2026. Cases here are original fiction; sources support bounded mechanisms or product functions, not an account test.