Support Board

Date/Time: Sun, 19 May 2024 00:29:23 +0000

Post From: Huge Differences in backtesting between

[2019-12-22 02:26:09] Link To Thread - Link To Post [Go To First Post]
Furinkazan - Posts: 19
In my backtesting of a strategy, I use both "CQG Web API" and "SC Data - All Services" data feeds. I found a huge discrepancy in results between the 2 data feeds. In a backtest of 1000 days of data, "SC Data" yields $200K of profits while "CQG Web API" yields only $5K! The strategy uses 1 minute chart. The "Intraday Data Storage Time Unit" for both data services is 1 Tick. I did a trade by trade comparison for many records and I found that trades almost always get a better fill using "SC Data". Even though the difference of individual trades is pretty small, the results could be significantly different even within a day, not to mention over 1000 days. The better fills with "SC Data" applies to both entries and exits. I've seen many cases where using CQG data would get me the worst possible fills on an exit. For example, it tends to exit on the extremes (high or low of the bar whichever is more unfavorable). I used Auto Trade System Replay Back Test with the following settings: Replay Mode: Accurate Trading System Back Test Mode Jump Method: To Next Bar Am I using the correct settings for getting the most accurate testing results? Is "SC Data" more accurate than the CQG data at the tick level?