Support Board
Date/Time: Thu, 06 Feb 2025 09:53:44 +0000
Problem with Backtesting - Big differnce between two instances of SC
View Count: 667
| [2020-04-22 14:05:10] |
| rainbow47 - Posts: 68 |
|
Hi, I am backtesting a S&P future strategy in two different insatnces of SC and get very different results (10 day testing with c.30k profit difference). I tested with "AutoTrade SSystem Replay Backtest" and "Replay Chart". The backtest runs over 10 days. I subscribed to the Denali Realtime feed and using a IB account. SC Instance 1 is for live trading and SC Instance 2 is for backtesting only. Therefore I am using different data connection settings (pictures attached). Any idea what could cause the dfifference? |
 Instance 1 - DataSettings.JPG / V - Attached On 2020-04-22 14:02:18 UTC - Size: 100.79 KB - 228 views Instance 2 - DataSettings.JPG / V - Attached On 2020-04-22 14:02:58 UTC - Size: 81.15 KB - 230 views |
| [2020-04-22 17:12:35] |
| rainbow47 - Posts: 68 |
|
I attached the activity log stats. Both Instances I use "Accurate Trading System Back Test" and testing them with "Auto Trade System Replay Back Test". I am using exactly the same study on both. How can the results be so different? Date Time Of Last Edit: 2020-04-22 18:07:34
|
| Activity Log Stats.JPG / V - Attached On 2020-04-22 18:07:31 UTC - Size: 591.29 KB - 220 views |
| [2020-04-22 18:41:59] |
| rainbow47 - Posts: 68 |
|
I did a bit more testing. It looks like Instance 1 is executing trades at different price than Inctance 2 when the signal gets triggered. See screenshots below. Based on my real life tests Instance 1 is fairly accurate while Instance 2 is totally off. Does anyone have an idea why prices for Instance 2 are off and how to fix it? Thanks Date Time Of Last Edit: 2020-04-22 19:12:20
|
| Instance 2 - TestTrade.JPG / V - Attached On 2020-04-22 18:42:17 UTC - Size: 202.16 KB - 224 views Instance 1 - TestTrade.JPG / V - Attached On 2020-04-22 18:45:22 UTC - Size: 206.11 KB - 229 views |
| [2020-04-24 07:17:33] |
| rainbow47 - Posts: 68 |
|
Worth noting that I use different Version of SierraChart: 1. 2072 (more precise and backtests are close to real tiem tests) 2. 2091 (completely off) Can someone help please? |
To post a message in this thread, you need to log in with your Sierra Chart account: