Hi,

Check out http://www.cubesteak.net/mt4-multi-b...-data-project/

Hi,

Check out http://www.cubesteak.net/mt4-multi-b...-data-project/

I leave my PC connected 24/5 to the internet, all the time presumably collecting demo tick data from different brokers. Now I would have thought that were I to backtest pairs that I know are being fed tick data, from for example Jan to Feb, I would get results around the 99% quality range. I don't! Anyone know why???

Did you reimport and "PeriodConvert" your new tick data ? Maybe this link will help:
http://www.forex-tsd.com/metatrader-...highlight=tick

I'm aware of how to convert M1 data to tick etc. My point is that I should already have 'real' tick data on my computer, downloaded over the past several months from being permanently online. I recall reading quite some time ago about someone using this data and receiving good results....

So what's the problem ? is it that after converting your tick data it still doesn't show as 99% ? Did you period convert each timeframe from the tick data ? Best regards, Lee

Properly converted M1 data will never show more than 90% data quality - why? because all points in between any 2 price bars are interpolated. Whereas tick data that is being streamed to the computer real-time is 'real' price data which is why the data quality number as defined by the MT backtester 'should' be higher. Many scalping strategies when tested with 90% data are losers but prove profitable when tested when the data quality is closer to 99%, so data quality clearly matters. I was just thinking that as I should have months of 'real' price data on my computer (after months of real-time streaming) that the data quality number shown by MT on backtests for this period should be higher. But it isn't.........

thats a very good point, I'm sure I have seen strategy tester reports with 99% quality though - I think many people feel let down by the quality of the backtester, I have seen BIG changes in backtest results since upgrading from v198 the backtests seem completly different ! How is this possible

Simple really. They've constantly making changes to make the whole thing closer to reality. And thus many people have seen their EAs performing worse.

Or would you rather prefer seeing great results in the tester and bad on a real account?



Cheers,
Diam0nd




I LOVE

__________________
- MetaTrader v4.00 Build 217 (02 Jul 2008)
- MetaTrader MultiTerminal v4.00 Build 213 (20 Mar 2008)
- Always use the latest (^ABOVE^) MetaQuotes products, saves a lot of nerves. What's new.
"METAQUOTES FOR LIFE!"

Yes, so have I but these are using 'real' tick data - you will not see 99% from converted M1 data. 'Tross' has a thread going where he tests EA's with such a setup with some interesting results.