Huge differences in backtesting - Leverage - General

Sergey Golubev 2007.03.16 17:26 #1

Different number of bars in your statements. Seems gaps in hostory for some broker.

Besides if EA is coded on open bar so open bar price and high/low of every bar is different for different broker.

For M5 timeframe it is very different (not 1 and 2 points - even more).

Interesting and Humour 6 sma cross Electra EA ** Released

Kevin 2007.03.16 17:34 #2

I downloaded Metatrader history for both tests. shouldent it be the same data?

Sergey Golubev 2007.03.16 17:55 #3

in your statement:

IBFX, bars in test 90479

Velocity, bar in test 89699.

780 bars in difference.

If everything is the same so there are some holes/gaps in the history. Not because of you. Because of the brokers.

Besides i think your EA was coded for open bar or for high/low of the bar and in this case it should be very different for different broker.

Newdigital's EAs EA Problem when testing All forward testing on

Kevin 2007.03.16 21:45 #4

I

The long term results are the same. . . too bad it caps out at 100 lots. should be an exponential growth curve instead of a linear.

Files:

2000_-_2007_ibfx.jpg 118 kb

[Deleted] 2007.03.17 20:04 #5

Yes i agree that different brokers have different data feeds. Id be more interested to see the difference in the results between backtesting, demo forward testing, and live forward testing.

It would be really intresting to see the difference of each is...

-Mikhail

Please don't stop Demo,metaquote.net Petit Pip Eater Why are there differences

[Deleted] 2007.03.18 04:03 #6

I can almost guarantee that there is some problem with that data. I got that kind of great results, and I was using the data I downloaded from Metatrader by using the download data button from the history option in MT4. It is really representative data and I think it is faulty. However, the results may not be that great, but maybe good enough to make you a millionaire few times over

good luck.

Where is the history MT4: How to rebiuld MetaTrader 5 Strategy Tester:

Kevin 2007.03.20 06:05 #7

Anyone here know why the live server and demo servers have different results.

I have 2 live accounts and 2 demo accounts running currently, all with the exact same EA. All with different results.

are the brokers delaying the demo servers or something ? anyone have an explination ? I am not complaining because the live servers give better results than the demo, but i dont like suprises. And would like to know the differences so that I can make the proper adjustments when i am testing new strategies.

Thanks,

KK

Experts: DVD Level (100-50 Difference between Demo and Sti

Sergey Golubev 2007.03.20 07:36 #8

Basicly it should be the same: real and demo for most of the brokers.

There are small differencies for real and demo server concerning execution of the orders (real server may do it more "slowly"). And the other is the spikes for demo server more than for real. And data "gaps" are in different places may be ...

But, basicly the different results you are getting because of different bars testing. It is not your mistake. It is the broker's gap in data.

Besides, may sure that you did not download the data from metaquotes because metaquotes data is coming "from nowhere" but you need the data exactly for your broker. And sometimes as i see the MetaTrader may download the data automatically for new versions of Metatrader and it is difficult to know: the data "from nowhere" or downloading from broker's server. That is why i prefer backtesting on the biuld 198 for example. Because if i use biuld 200 or higher so the data may be mixed.

You know what I MQL5 metaquotes data Is it true that

pipsweak 2007.03.20 13:13 #9

Kurka Fund:
Anyone here know why the live server and demo servers have different results.

I have 2 live accounts and 2 demo accounts running currently, all with the exact same EA. All with different results.

are the brokers delaying the demo servers or something ? anyone have an explination ? I am not complaining because the live servers give better results than the demo, but i dont like suprises. And would like to know the differences so that I can make the proper adjustments when i am testing new strategies.

Thanks,

KK

Does the EA generate less trades on the live account or more trades, or does it enter/exit and different prices than demo? There might be a way to remedy that problem. I do believe demo servers are slightly delayed.

OOP vs procedural programming Gathering a team to Trading by experts is

Kevin 2007.03.20 14:04 #10

I am getting the mixed results while foward testing.

It opens more trades in the live acounts and usually gets an extra pip. Its a scalper system, this might have something to do with it. Either way the results are consistant with the backtest that i posted.

What are your suggestions on a remedy. . . I am working on another system and would like to get as accurate of a model as possible.

Thanks,

Kk

Experts: Grid Template EA Fire EA: Can you Portfolio: PriceChannelExpert and others