Yes that good in itself, but until atleast MY broker (and preferably more of them) offer MT5 I won't even think of rewriting all my code.
In the mean time I'll get the job done by the +4-fold increase in speed in W7/64, combined with the 10-fold increase from tick skipping.... combined we get some 40-fold increase in speed.
You code does skip ticks, but it does that in a way possible important informations might get lost.
You code does skip ticks, but it does that in a way possible important informations might get lost.
Use code like this:
to skip ticks without loosing important informations.
All Bar openings, as well as new highs and new lows are not skipped.
I don't know if this would be an equivalent. However, one could remove the Volume information while importing data into mt4. By default, it'll use the minimum O-H-L-C or 4-Ticks, this helps speed up the back-testing.
Along those lines, have anyone tried entering Volume(s) of say 100 during a M1-data import?. My guess is that this should be a viable alternative to Tick-Data if it works. I planed to test it but never gotten around to it :)
My broker has now enabled MT5 on Live accounts and that makes it less academic. I wrote my first MQL5 indicator (to display the spread) last night and this evening I had a go at converting the speed test EA from the first page of this thread. Here it is ...
Now the idea was to compare speeds in the strategy tester doing like-for-like trades. Of course I forget the most fundamental change in MQL5. You can't hedge. In fact you can only ever have one position open per symbol by the looks of it. That means multi-strategy, separated by MAGIC_NUMBERS, seems to be impossible.
My test EA failed miserably as it just places a trade every hour to see how long it takes. Doing a like-for-like comparison this way is not possible :-(
Yes that good in itself, but until atleast MY broker (and preferably more of them) offer MT5 I won't even think of rewriting all my code.
In the mean time I'll get the job done by the +4-fold increase in speed in W7/64, combined with the 10-fold increase from tick skipping.... combined we get some 40-fold increase in speed.
You code does skip ticks, but it does that in a way possible important informations might get lost.
Use code like this:
to skip ticks without loosing important informations.
All Bar openings, as well as new highs and new lows are not skipped.
You code does skip ticks, but it does that in a way possible important informations might get lost.
Use code like this:
to skip ticks without loosing important informations.
All Bar openings, as well as new highs and new lows are not skipped.
I don't know if this would be an equivalent. However, one could remove the Volume information while importing data into mt4. By default, it'll use the minimum O-H-L-C or 4-Ticks, this helps speed up the back-testing.
Along those lines, have anyone tried entering Volume(s) of say 100 during a M1-data import?. My guess is that this should be a viable alternative to Tick-Data if it works. I planed to test it but never gotten around to it :)
All Bar openings, as well as new highs and new lows are not skipped.
My broker has now enabled MT5 on Live accounts and that makes it less academic. I wrote my first MQL5 indicator (to display the spread) last night and this evening I had a go at converting the speed test EA from the first page of this thread. Here it is ...
Now the idea was to compare speeds in the strategy tester doing like-for-like trades. Of course I forget the most fundamental change in MQL5. You can't hedge. In fact you can only ever have one position open per symbol by the looks of it. That means multi-strategy, separated by MAGIC_NUMBERS, seems to be impossible.
My test EA failed miserably as it just places a trade every hour to see how long it takes. Doing a like-for-like comparison this way is not possible :-(