Strategy Tester accurate?

New comment
 
Is the strategy tester portion of MetaTrader a reliable estimation of what would actually have happened if I had run my Expert Advisor for the specific time period? I am seeing some great results and I just wanted to make sure I am not doing something stupid (like the tester data is not real data, or something). Does anyone have an idea about this?
 
Did you read Larry Williams - http://www.amazon.com/gp/reader/0471297224/ref=sib_dp_pt/102-2544486-9324950#reader-link ? His book can give you many answers.
 
Is there any way to get an answer to my question without buying this book? I just want to know generally speaking if the Strategy Tester is supposed to be a good way of testing an EA.
 
Here is a simple one word answer for you. NO.

MT4 backtesting is far from accurate for many reasons.
One example is the way it "peaks into the future" at your closing prices.
EAs that work based on bar closes will always trade in the direction of the close, showing nice profits.
In real time, we do not know what the closing price will be, so results are much different.

Backtesting (mostly visual) helps with building EAs, so you can at least get an "idea" if it is working right.

People may argue my comments, but what I have said is the simple truth.

Demo trading is as close as you can get to real results with MT4.
 
Hello there,

Backtesting is not reliable, in my personal opinion, I have 2 pcs on 1 there is metatrader downloaded data and on the other pc, i have downloaded alpari data, my ea is based on purely mathamatical formulations and nothing to do with the built in functions of metatrader. but on metatrader terminal the chart is going sky high and on alpari terminal it is going rock bottom. the ea is same.

What I personally suggest is that rely on forward testing only, if you have any logic in mine, apply it by making several variants of your ea and installing metatrader several times on different directories and checking all the ea's on all metatraders at a single point of time.

Here I also recommend you all to develop EA on mathamatical formulations only, if you average the results on bars behind that will be the best EA. If you run after the bar high low time values etc individually for different different brokers, you are likely to go in crisis very high.

I do not know how did you understand this valuable information, but it is the fact.

Regards
New comment
Reason: