Support Board
Date/Time: Sun, 24 Nov 2024 02:29:56 +0000
Post From: Amazing backtesting performance. Sanity check.
| [2024-07-08 10:22:15] |
| RobJ - Posts: 4 |
|
Hi guys, I'm doing automated backtesting with my own ASCIL studies on a single chart using the accurate trading system backtest mode. All involved studies are optimized for speed (0 ms while running backtests), are intraday and replay is done with the chart hidden (no drawing) with chart update interval at 10 ms. I've done extensive testing with various replay speeds ranging from 1 - 4800 during regular trading hours (ES). I'm happy to report I'm getting the same results / trades even on the highest speeds. Only at speeds higher then 3600 performance gains start to diminish which is due to my hardware. 1. Well done SC! This piece of software is capable of pretty much everything you can think off using ASCIL/C++. For example, my studies use a logging framework and MySQL for getting/storing results. 2. My backtest speeds are much higher then most other users report. Some of them seem to have problems running speeds higher then 100. As a sanity check, is there any reason why I shouldn't be able to get these results. In the documentation I can't find any specific limitation but I just want to make sure I haven't missed anything before spending days/weeks backtesting strategies at 3600x. Thanks! |