Support Board
Date/Time: Mon, 22 Dec 2025 09:26:58 +0000
Post From: Explaining the Differences in Entry DateTime Between Accurate Replay and Forward Testing
| [2025-12-21 21:44:11] |
| Apollonia - Posts: 30 |
|
Hi SC Engineering, I noticed that there was a consistent ~2 second difference between equivalent transactions whenever I forward tested and replayed over the same period. I wanted to confirm what the differences in times were between them both. When I looked at the trade activity I saw that the "Bar start date-time: xxxx" was the same entryTime in the accurate replay (~2 seconds faster). My questions: 1. Why is the accurate replay's entryTime equivalent to the "Bar Start date-time"? My study conditions are not bar-based and when I used GetCurrentDateTime() to separately measure the time (after order was accepted) it matched. Is this just misleading naming? 2. Is the difference attributable to real processing time for simulated orders? 3. Is there a way to bridge this gap in replay so it's more realistic? Partially because the ~2 second difference can affect certain parts of the logic, like on edge cases where a trade may be blocked in one instance vs not blocked in another due to the slight timing differences. Some timing examples for your reference: Forward tested | Accurate Replay 2025-12-18 19:31:12.519 BP | 2025-12-18 19:31:10.952 BP 2025-12-18 19:43:32.037 BP | 2025-12-18 19:43:30.487 BP 2025-12-18 20:54:38.928 BP | 2025-12-18 20:54:37.231 BP Date Time Of Last Edit: 2025-12-21 21:51:48
|