Discovered an odd feature/bug when warmup period is requested - we get a duplicate datapoint at the beginning of our requested startDate. Run the attached backtest and verify duplicate from logs. Note if line #14 is commented out (warmup setup), it no longer reproduces; neither does it happen with warmup disabled and startDate moved back to 25th.

Looking at the fxcm eurusd minute data between 25th & 26th of June, no duplicates can be spotted.

Also the log confirms duplicate is at 1561507140 (which translates to Tue 25 Jun 2019 11:59:00 PM UTC); yet when we plot said data and consume from the ZMQ, the duplicate timestamp now reads 1561521600 (Wed 26 Jun 2019 04:00:00 AM UTC). All times should be UTC right, so where is this discrepancy creeping in from?

Author