I'm running a dip buyer strategy and I recently just ran the backtest for 24 months with 10 minute granular scale and results were horrible, way worse than even the worst Monte Carlo simulation using same date scramble (which I thought was the preferred method for a dip buyer). I'm using EODHD for my intraday data and Norgate SP 600 current and past for the main data set. Any advice or observations?
Rename
The topic is currently discussed in this thread:
https://www.wealth-lab.com/Discussion/Difference-between-backtesting-and-live-trading-results-10705
https://www.wealth-lab.com/Discussion/Difference-between-backtesting-and-live-trading-results-10705
Should we merge this post with the topic you referred to, @DrKoch?
...albeit these are slightly different questions:
* Mismatch between live trading and backtest
* In a backtest with EOD data: Mismatch between raw backtest and backtest with Granular processing
* Backtest with small position size (no NSF positions) and backtest with large position size (many NSF positions): Significant mismatch in Avg. Profit per Trade.
* Mismatch between live trading and backtest
* In a backtest with EOD data: Mismatch between raw backtest and backtest with Granular processing
* Backtest with small position size (no NSF positions) and backtest with large position size (many NSF positions): Significant mismatch in Avg. Profit per Trade.
QUOTE:
Should we merge
I think it would be best to have two separate discussions:
1. Real trading vs. backtest
2. Problems with backtests: Granular Processing and NSF positions
Unfortunately Discussion 10705 is already a mixture of both (I messed this up... :( )
I just had a chance to read the other thread...In my opinion they can be combined. They pretty much deal with the same exact thing...mine just adds another layer to the issue which is the granular backtest which appears to be much closer to real life performance.
Your Response
Post
Edit Post
Login is required