molanis.com/forum - Tools for MetaTrader : The place to ask about the best expert advisor builder, expert advisor downloads, and expert advisor programming

Good news! New versions of MT4 now allow multi timeframe backtesting. You only need to make sure you have all data (quotes) for all timeframes available in mt4. If you do not have it you may or may not get an error. For next month will create a short explanation about it.
Mind you, mt4 does not backtest multi currency eas. You can get signals from different timeframes and pairs during backtesting but you can only trade one pair (the one selected on the backtesting settings)

Hello Molanis,

I think I want to buy the latest version of SB for MT4 Build 600+ but I need to be able to do accurate back testing. However, historically the problem with back testing in MetaTrader has always been tick data accuracy. In the past, I've always had to run Birt's Patch for back testing in an older build of MT4 and then not allow MT4 to run its Liveupdate.exe.

QUESTIONS:

1) Will EAs compiled under your newest SB run on both MT4 build 600+ and MT4 build 409 (is there backwards compatibility)? If not, then I would have to run all my back tests under MT4 build 600+, which means losing access to all the historical tick data (FXT files) that I created using Birt's Patch for build 409.

2) Do you know much about any improvements in the backtester environment of MT4 build 625 (you can use multi-timeframe EAs, but how's the Modeling (%) quality)?

MetaQuotes does not seem to reply to questions very well on their site, so I'm trying to find out from people who have a reason to use MetaTrader for their own third-party produce development purposes such as Molanis.

Thanks for the help.

If you buy a Pro version you get the code so you can compile it with your old mt4 version (open it with metaeditor old version and click on compile). By default the new version compiles for the latest mt4 version and not for the old version so having the code is needed for old version backtesting.

Modeling quality is similar. You need to get your own data to do 90% or more.

Our products generate normal eas - backtesting is done with mt4 so any improvement on backtesting is done on mt4 and not on the ea. Thus we do not control the quality of backtesting.

Ok, thanks! Pro version it is!



I figured as much. I did not know whether or not you were aware of any internal changes that Metaquotes may have included into its newer versions (600 or higher) that make backtesting better for the active trader depending on tickdata.

Out of the box, for the active trader, MT4 testing lacks good modeling quality unless you force feed the engine tickdata. The problem is that most of the retail brokers out there don't drive their servers using tickdata, nor do most of them store tickdata on their back-end databases. So, when you download from their back-end through your MT4 client, the lowest timeframe you get is M1 and that won't yield sufficient modeling for an active trader where tick-precision on the entry and exit is crucial (M1 data works fine for the swing traders or longer term investors).

Anyway, like I said, I just wanted to know if you had any new inside info on their new Backtester Engine. It sounds like they are using the same exact engine for backtesting, however.