Page 1 of 1

Ticks modelled vs Bars in Test

Posted: Thu Aug 04, 2016 8:08 am
by keleshia
I have one more question, i tried Backtesting, and why do i get different Ticks Modelled and Bars in Test for the same pair and different TF?

May i know how do i ensure whether which is the right one and closes? The most ticks modelled and most bars in test?

Re: Ticks modelled vs Bars in Test

Posted: Fri Aug 05, 2016 8:59 am
by tickstory
Hi Keleshia,

The number of bars in the rest relates to the number of bars tested in that timeframe, so it will differ. The number of ticks modelled represent the actual price movements that occurred.

Hope this helps.

Re: Ticks modelled vs Bars in Test

Posted: Wed Aug 10, 2016 8:10 am
by keleshia
Hi

the thing is, i backtested it on the same Timeframe, same EA, but with different Datas that I have exported.
For example, SQ Tick Downloader and conversion via Tickstory, and another via conversion via Birt's TDS.

So i'm not sure why they vary greatly. Any idea why? How do i know which is the correct one? More Bars? More models?

Re: Ticks modelled vs Bars in Test

Posted: Thu Aug 11, 2016 9:00 am
by tickstory
Hi Keleshia,

We can only really comment on Tickstory: Assuming you are using the default MT4 Export settings, the number of ticks modelled is 1-to-1 with the raw tick data that Dukascopy has available. The data is utilised as-is except for any timezone adjustments you have specified. If you enable options such as "Filter duplicate ticks" then the number of ticks modelled will obviously be adjusted down accordingly.

If you are comparing with other products, the simplest option may be to export a small subset (say 1 day) of tick data to CSV and see what differences there are.

Hope this helps.