MQL5 Cloud Network: inconsistent OnTick stream under real-tick modeling
Technical reproduction report  |  14 October 2026  |  MetaTrader 5 Strategy Tester
1. Issue summary
A non-trading diagnostic Expert Advisor (TickParityDiagnostic_V4.mq5) produces materially different OnTick event sequences in local Strategy Tester optimization and MQL5 Cloud Network optimization when configured for “Every tick based on real ticks”. The CopyTicksRange tick-history metrics and CopyRates(M1) bar metrics match exactly between environments in the same fixed one-hour window. The discrepancy changes optimization outcomes and prevents reliable local reproduction of Cloud-optimized trading strategies.
Observed OnTick events: local 45,665 versus Cloud 28,758 (16,907 fewer; approximately 37.0% below local). This is an observed discrepancy, not proof that the Cloud deliberately uses generated ticks.
2. Reproduction configuration
* Broker/server: FxPro-MT5; symbol GOLD; chart timeframe M5.
* Modeling option selected: Every tick based on real ticks (Portuguese UI: “Cada tick baseado em tick real”).
* Test interval: 2026-06-22 00:00 through 2026-06-27 00:00 (server time).
* DiagnosticHour = 2026.06.22 10:00:00 (one-hour measurement window: 10:00–11:00, server time).
* Optimization: complete optimization, Custom max; DiagnosticVariant = 1..22 with step 1; CacheBuster kept fixed for comparable runs (change only to avoid reuse of cached results).
* EA performs no trading. Each optimization pass reports one diagnostic metric via OnTester().
* Cloud run: 22 tasks (100%) processed on Cloud agents, previously reported agent build 6251. Local run: agent build 6230. Updated portable local run: terminal and agents build 6259, 10 local tasks (100%).
3. Metric-by-metric results
Values below are taken from exported optimization results. The portable build 6259 run measured variants 1–9 only. The full local and Cloud runs measured all 22 variants.
Variant / metric
Local full
Cloud full
Local 6259
Cloud vs local
1. OnTick count
45665
28758
45665
DIFF
2. OnTick timestamp hash
514014352
1491913586
514014352
DIFF
3. OnTick Bid hash
1608073743
339140397
1608073743
DIFF
4. OnTick Ask hash
2761616376
2513685456
2761616376
DIFF
5. OnTick combined hash
4060870639
32894255
4060870639
DIFF
6. OnTick Bid changes
13311
28723
13311
DIFF
7. OnTick Ask changes
13313
28723
13313
DIFF
8. OnTick first time_msc
1782122400020
1782122400000
1782122400020
DIFF
9. OnTick last time_msc
1782125999870
1782125999751
1782125999870
DIFF
10. CopyTicksRange count
45665
45665
45665
MATCH
11. CopyTicksRange timestamp hash
514014352
514014352
—
MATCH
12. CopyTicksRange Bid hash
1608073743
1608073743
—
MATCH
13. CopyTicksRange Ask hash
2761616376
2761616376
—
MATCH
14. CopyTicksRange combined hash
4060870639
4060870639
—
MATCH
15. CopyTicksRange Bid changes
13311
13311
—
MATCH
16. CopyTicksRange Ask changes
13313
13313
—
MATCH
17. CopyTicksRange first time_msc
1782122400020
1782122400020
—
MATCH
18. CopyTicksRange last time_msc
1782125999870
1782125999870
—
MATCH
19. M1 bar count
60
60
—
MATCH
20. M1 bar hash
2084135571
2084135571
—
MATCH
21. M1 summed tick volume
30916
30916
—
MATCH
22. M1 missing minutes
0
0
—
MATCH
4. Key controls and observations
* Local real-tick OnTick count equals the CopyTicksRange count (45,665). Cloud OnTick count is 28,758, whereas its CopyTicksRange count is also 45,665.
* All nine CopyTicksRange metrics match local versus Cloud, including hashes and first/last millisecond timestamps.
* All four M1 metrics match local versus Cloud: 60 bars; bar hash 2,084,135,571; summed tick volume 30,916; zero missing minutes.
* OnTick Bid/Ask changes: local 13,311 / 13,313; Cloud 28,723 / 28,723.
* OnTick first and last tick timestamps (milliseconds): local 1782122400020 / 1782125999870; Cloud 1782122400000 / 1782125999751.
* The local portable installation downloaded fresh GOLD M1 history and tick history, including approximately 290.86 MB of preliminary tick downloads, and still reproduced the earlier local OnTick results exactly on build 6259.
* The portable log explicitly confirms “authorized (agent build 6259)” and “local 10 tasks (100%), remote 0 tasks (0%), cloud 0 tasks (0%)”.
5. Requested investigation
Please investigate why Cloud optimization agents generate or deliver a different OnTick event stream from local Strategy Tester agents for the same broker, symbol, time interval, and selected real-tick modeling mode, despite identical historical tick and M1 queries. In particular:
* Does the Cloud use a different tick modeling, validation, filtering, or fallback path under “Every tick based on real ticks”?
* Could agent build differences, environment restrictions, or tick preparation explain this reproducible mismatch?
* How can a user verify from agent logs which tick stream was actually supplied to OnTick during a Cloud optimization?
* Is there a documented configuration or corrective procedure to ensure Cloud and local Strategy Tester tick-stream parity?
6. Reproduction steps
1. Compile the attached TickParityDiagnostic_V4.mq5 in MetaEditor.
2. Select GOLD / M5, the stated dates, and “Every tick based on real ticks”.
3. Set DiagnosticHour and optimize DiagnosticVariant from 1 to 22 with Custom max.
4. Run once with local agents only and once with MQL5 Cloud Network only, with identical parameters and broker data.
5. Export optimization results; compare variants 1–9 (OnTick), 10–18 (CopyTicksRange), and 19–22 (M1).
6. Confirm task allocation and agent builds in the Tester Journal; avoid optimization cache reuse.
7. Scope and limitations
The evidence establishes a repeatable discrepancy in these test environments, but does not establish its internal cause. CopyTicksRange agreement does not prove that the same ticks were delivered as OnTick events. The portable build 6259 run was local only and is not a same-build Cloud comparison. The Cloud build 6251 was observed in prior Cloud journal evidence; the attached 20261014.log documents the later portable local run. No claim is made that the Cloud intentionally substitutes generated ticks.
8. Attached evidence
TickParityDiagnostic_V4.mq5 (source); Novo Teste Local(1).xlsx (full local 22-variant results); Novo Teste Cloud Network(1).xlsx (full Cloud 22-variant results); Teste Local Portátil.xlsx (build 6259 local variants); 20261014.log (portable Tester journal).
Prepared as a minimal, independent reproduction. No trading-account identifiers or proprietary strategy source are needed.