New build and backtesting issues...

Has anyone had any backtesting issues with the new build? I was testing an EA that showed a normal slow and steady profit with about 380 trades (90% modeling quality) and with the new build the same EA is showing about 40,000 trades showing an astronomical profit graph. Any thoughts???

Be happy, astronomical profit graphs are supposed to be good .

I've not used the new build's strategy tester yet.

I haven't experienced any issues like that. Have a look at the changes of the new version. And you could post the EA and some information that we could test it and then we might find the reason why its behaving so strange. Without any information that impossible :/

That is waaaay to funny saaedb !

On the build, just realized when loggin in after a black out here on the Oregon coast yesterday there was a new build. Have yet to use the ST, but i did download a new 1m Alpari data and create 5 & 15m charts and am using the data to manually test one day at a time my model as i'm still not happy with the inconsistancies of letting the ST do it's magic (if it has any!).

: ) Thom

Yes, those astronomical profit graphs looks great. This one is nutz. It goes from 10,000.00 to something like 45,353,867.00...I'll re-run it on my other machine that hasn't been updated yet and see what happens...Should of thought of that first, then maybe download the historical data again from Alpari. If nobody else is having any issues than it might be just me...

Palmer, it doesn't seem to be "just you". Actually, I've entered FF now to see whether anybody has any explanation for the following. I compared the old and the new build LOW 1HR GBPUSD prices for 2004-2006. The absolute discrepancies are up to 62 p. for 2006 (mean = 1.8, SD = 3.2), up to 127 p. for 2005 (mean = 2.0, SD = 4.2), and up to 214 p. for 2004 (mean = 12.1, SD = 15.5)!!! Needless to say, a strategy I've developed during the last month that backtested very well with the old build data is worth nothing now. I'm using a live MT4 with FXDD. Any ideas, what happened? Which data should I trust?

Guys,

If you can, hold back on upgrades when they first become available.
If you have a build that's been keeping you happy, say "no" to the upgrades for a few days at least. Any issues seem to come to light the Monday after the weekend upgrade most of the time . If no-one mentions anything over the following week there's a good chance you can upgrade no hassles.

These guys are sending new builds out quite often, there's no way they would be able to test for every circumstance in the time-frame they are doing this.

Just a suggestion.

Bundyraider, thanks for your advice. However, I seem not to understand a basic thing. Shouldn't have been there one and only one price quoted by a given broker on a given hour of a given day in the past? My statistics show that there were at least two such prices. How is it possible?

I had a build that kept me happy. Do you mean that if I continue using it my broker will continue quoting the same sort of prices that fits my strategy? And if I switch to the new build, I'll be quoted different prices? Can we choose the reality we trade in? It sounds strange to me...

FWIW:

The previous build and the EA equity graph that had a nice gentle upslope of profit now pretty much reads the exact opposite, a nice gentle downslope to zero.

Bundyraider:

I'll give it a few days and see if the new build results revert to the previous builds bactest results. If they don't then something is wacked somewhere. I'm using FXDD as well......