Extent of minute data

New comment
 

Hi All

i'm testing a strategy on the minute timeframe. I understand that on the graph, the minute data doesn't go back further than about 3 months, however on the strategy tester graph it shows minute data going back 6 months if i'm testing to 6 months back. I have also found that with some strategies, if i backtest last 3 months, it produces a profit, but if i go further back, it results in losses for various input variables. 

I want to understand how the strategy tester is showing minute data going further than 3 months back, and is this reliable data. 

Is some of my strategies bombing out due to lack of minute data? Some strategies show profits in 3 months back and 6 months back. 

should i put more weight on the last 3 months of testing? 

 
Laveshan Govender:

Hi All

i'm testing a strategy on the minute timeframe. I understand that on the graph, the minute data doesn't go back further than about 3 months, however on the strategy tester graph it shows minute data going back 6 months if i'm testing to 6 months back. I have also found that with some strategies, if i backtest last 3 months, it produces a profit, but if i go further back, it results in losses for various input variables. 

I want to understand how the strategy tester is showing minute data going further than 3 months back, and is this reliable data. 

Is some of my strategies bombing out due to lack of minute data? Some strategies show profits in 3 months back and 6 months back. 

should i put more weight on the last 3 months of testing? 

The strategy tester has different operation modes.

Ensure that you're using "Use real ticks" for realistic back tests.

It is common for a back test to be profitable over one period and bomb over a larger period.

Sometimes it means you just so happened to pick an easy period initially, and then you included challenging periods in the larger test.
New comment