MetaQuotes-Demo XAUUSD: raw epoch clock semantics and recurring two-hour tick/M1 absence

New comment
 
We are validating read-only research data from MetaQuotes-Demo, XAUUSD, terminal build 6230 and MetaTrader5 Python package 5.0.5735. No order submission is involved.

1. A locally observed quote on 1 October 2026 at 10:18:43.932533 UTC returned tick `time_msc=1790860723057` and `time=1790860723`. Ordinary UTC epoch decoding gives 13:18:43.057 UTC. Earlier XAUUSD and GBPUSD observations show the same three-hour discrepancy. The collector is not an independently synchronized server clock witness. Please confirm the intended raw tick/bar epoch semantics and this server's historical civil-time/UTC policy rather than assuming either clock is correct.
2. Historical range queries used raw integer seconds **1790809140–1790816460** and the equivalent timezone-aware datetime arguments **2026-09-30 22:59+00:00 to 2026-10-01 01:01+00:00**. Both forms returned identical 1,122 ticks and three M1 bars with `(1, Success)`. Using local half-open bins, raw label 22:59–23:00 contained 855 ticks, 23:00–00:00 contained none, 00:00–01:00 contained none, and 01:00–01:01 contained 267. The request endpoint's 01:01 bar is included by the native bar API and excluded from these bins. Exact representative raw values and hashes are attached in `SANITISED_SOURCE_EVIDENCE.json`.
3. The current research Specification quote table displays Monday–Friday **00:00–23:00**, weekends blank. Please provide the **complete historical quote-session list and time basis applicable Sep 23–Oct 1**, including the Oct 1 00:00–01:00 raw-labelled hour, holidays/DST/rollover and Friday/Sunday boundaries. A current weekly table alone cannot certify those historical hours.
4. Please establish whether successful empty tick/M1 ranges can reflect incomplete retained/synchronized history for this server. What authoritative source-side record or supported retrieval criterion distinguishes scheduled no quotes from missing history? The same two-hour pattern occurs across the affected H1 history. Earlier M1 evidence also has mapped-UTC Sep 24 12:45–12:49 and a longer Sep 7 absence under the provisional minus-three-hour mapping; please identify any applicable historical exceptions/retention limits. This is not a claim that missing ticks prove flat prices.

The local audit originally used an end-minus-one-millisecond bound that omitted 18 control ticks. This has been isolated and corrected by using whole-second covering bounds with local millisecond half-open filtering; both empty hours persist. No offset rotation or price fabrication was used.

Needed answer: source-backed epoch/time policy, complete applicable historical sessions and whether the recorded absent open-session observations are recoverable or known unavailable. No account/password/API key, configuration change, new account or alternative source is requested.
 
Ask your AI agent to search the answer before posting on the forum what was already asked and answered several times.
New comment