Чтобы пользоваться чатом, пожалуйста, войдите в систему
Обратно к контактам

Fast Forex Millions back and forward testing by Remus

Jan 21, 2012 at 09:21
9,728 Просмотры
120 Replies
Участник с Apr 25, 2011   262 комментариев
Jan 23, 2012 at 07:54
I'm done with all the backtests I wanted to do, but won't post them here, since Birt uploaded his backtests now, and they match mine (I checked the charts, as well as looked at a couple of individual trades):

https://eareview.net/live-forward-test/fast-forex-millions

Bottom line: forwards tests are going to tell us if the performance shown by the backtests can be sustained.
He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
walker36
forex_trader_36599
Участник с May 13, 2011   1341 комментариев
Jan 23, 2012 at 10:24
there can be differencies between brokers for any EA that using different data feeds ( indicators ) + it has slippage setup this will also cause differencies + most probably it has spread filter that can also effect.
BenniB84
forex_trader_59882
Участник с Jan 16, 2012   31 комментариев
Jan 23, 2012 at 13:02
This is going to be a VERY interesting one to watch. Really tempted to jump on this one before is gets big and starts the blowing process... looking good overall?
Участник с Apr 25, 2011   262 комментариев
Jan 23, 2012 at 17:14

   walker36 posted:
   there can be differencies between brokers for any EA that using different data feeds ( indicators ) + it has slippage setup this will also cause differencies + most probably it has spread filter that can also effect.

Yes, the price feed will induce differences, and once people understand that, they will probably have the expectations set correctly (i.e. you cannot get all the same trades across brokers, but if the EA is based on a profitable strategy, the resulting expectancy will roughly be the same - discounting the spread, of course).

Also, what I'm trying to show here is that the demo account (at FXDD) is not always getting better fills than the real one:

https://www.myfxbook.com/members/NextLevelForex/fast-forex-millions-fxdd/228516
https://www.myfxbook.com/members/NextLevelForex/fast-forex-millions-fxdd-real/106578

The trouble with that is filtering the different trades, but once you do that, you can see that we already have trades that are better in the live account.

For me this (doing the backtests and now forward testing) was a very educational experience, through which I had some of my assumptions confirmed, and also learned new things.
He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
Участник с Apr 25, 2011   262 комментариев
Jan 23, 2012 at 17:22

   BenniB84 posted:
   This is going to be a VERY interesting one to watch. Really tempted to jump on this one before is gets big and starts the blowing process... looking good overall?

Benni,

It's looking OK from my perspective. Not great, but good enough to worth using it live in my account.

I don't like that the sales page advertises Pips Expectancy much higher than what I got in the backtests (I did EURUSD from 2007 through end of 2011, and I did not get anywhere close to what they claim to be). I also don't like that they (the vendor) did not put up a real money account, to show the commitment behind the EA.

However, I like that the EA passed the backtests for 4 of the 5 currencies (USDCHF is barely profitable from my testing), so I'm willing to risk 0.01 lots in my real account to learn more about the EA. The promise of this EA is that the wins are bigger than losses, and that's quite something, given that most of the EAs I looked so far at had that very much reversed (TP 15 pips, SL 120 - you have probably seen that already, more than enough :P)

The FXDD accounts linked above indeed show larger wins than losses, but it's too early to tell if that holds true for longer periods. What is good is that the EA takes trades often, so we'll have data to judge it pretty soon.
He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
Участник с Aug 20, 2011   588 комментариев
Jan 24, 2012 at 02:16
hi,

have you tried to test using different parameters?
usdchf appear to be the best , but not using the default parameters. but it was not tested using tick data and only for half year. (in fact I think a long term back test means nothing because the EA will be optimised for a past good period but not for the current 'flat' market)
so I put filter to 800, KofATR 4.6 and peratr to 26 for usdchf.

what did you get in a longer tick based back test?

for some reason... the GBPUSD did not provide any profitable configuration for the last 4 month in back test (again 90% quality) I'm not able to make it profitable
Участник с Apr 25, 2011   262 комментариев
Jan 24, 2012 at 06:40
I only used the default parameters (except for lot size, which I set to fixed 0.01).

USDCHF had the worst overall performance (talking about the whole 2007-2011 timeframe). Disclaimer: I did not do USDCAD backtesting at all, so I am only comparing the other 4 pairs here. If you look at only the recent months, yes, the USDCHF performance was impressive, so you should look at longer periods, otherwise you'll get only a glimpse of the picture.

GBPUSD had the second best performance over the whole period. And again, if you only look at the recent months, GBPUSD was flat.

I think 90% quality is good for this EA, but I wanted to use the tickdata since I believe that data is more accurate then the one in MT4 history center. So if your data is good, 90% testing should give almost the same picture was my 99% testing.

For more details, see the images in the attached an archive.

Приложения:

He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
walker36
forex_trader_36599
Участник с May 13, 2011   1341 комментариев
Jan 24, 2012 at 11:03
what is the difference between FXDD & Peppers??
Участник с Apr 25, 2011   262 комментариев
Jan 24, 2012 at 11:47
FXDD is what I consider an expensive broker, with very high spreads (2.2 on EURUSD, and above 4 one some of the other FFM pairs), while Pepperstone spreads are much lower.

For the moment Pepperstone account is behind FXDD, but I expect the situation to reverse after a longer series or trades.
He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
walker36
forex_trader_36599
Участник с May 13, 2011   1341 комментариев
Jan 24, 2012 at 11:53
but here must be reason for that. as you have said peeper spreads are lower. I had mentioned some issues about pepper at my thread earlier. I suspect now what I mentioned is correct. lets wait and see🙄🙄
Участник с Jul 24, 2011   91 комментариев
Jan 24, 2012 at 12:52
What are you talking about, Walker? Can you clarify?
Участник с Apr 25, 2011   262 комментариев
Jan 25, 2012 at 02:39 (отредактировано Jan 25, 2012 at 02:57)
Quick update: forward testing performance, in all 3 accounts, just sucks ;) The 2 demo accounts are running with the equivalent of a 2% risk, and the real account is running with the equivalent of a 1.5% risk.

On the positive side, the trade count is steadily increasing, so by the end the week we'll likely have 50+ trades in at least one account. I'm looking for 100+ trades before trying to make sense of the statistics, but 50+ is also acceptable (as the minimum).
He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
walker36
forex_trader_36599
Участник с May 13, 2011   1341 комментариев
Jan 25, 2012 at 03:28
this is opening Hedge trades. purpose is not hedge but they are hedged.
Участник с Apr 25, 2011   262 комментариев
Jan 25, 2012 at 03:47

   walker36 posted:
   this is opening Hedge trades. purpose is not hedge but they are hedged.

I intentionally activated DualMode, since it's the default (and that's how I did my backtests). That (the use of hedging) is one reason I picked FXDD (the other reason is that is has high spreads, so it's good to see how the EA handles that).
He who thinks he can, and he who thinks he cannot, are both right. ~ Henry Ford
Участник с Aug 20, 2011   588 комментариев
Jan 25, 2012 at 11:56
ouch... all demo and real account just get a big bump :)
well... opening a trade at the top of a trend and another one at the bottom opposite when the signal reverse on the 2 cases is not a good idea.
lol

the good point... apparently the backtests reflects the same trade like the real account.
so we will be millionaires!!!! well... only at 10% of risk or more... lol
Участник с Jul 24, 2011   91 комментариев
Jan 25, 2012 at 13:03

   NextLevelForex posted:
   Quick update: forward testing performance, in all 3 accounts, just sucks ;) The 2 demo accounts are running with the equivalent of a 2% risk, and the real account is running with the equivalent of a 1.5% risk.

There's this message from the FFM team:

'Hello,

I hope your having a wonderful day. Just in case
you didn't read the text file that was accompanied
by the zip file when you downloaded Fast Forex
Millions I would like to let you know what it said.

As this advice can help you manipulate the performance
of the robot for your broker.

FilterDepth setting in robot settings will range depending
on Dual Mode setting. (Info below)

When DualMode = true
Range: 0 - 1200

When DualMode = false
Ranger: 0 - 2400

Depending on your broker, and price feed and such, you
can control the sensitivity of the trade strategy to be safer
or less safer. If you find it's not trading well for you, you
can increase the filterdepth to the max number above between.
You can add 150 - 300 a week on the filters up to the max
range to improve results.

By doing this you are able to 'tune' the robot to your broker.

Best of luck!

FFM Team'



Did you adjust FilterDepth to your brokers?
Участник с Aug 20, 2011   588 комментариев
Jan 25, 2012 at 13:06
I do this, yes.
it was clear in the doc that there is some finetuning to do for each account.
Участник с Feb 09, 2011   320 комментариев
Jan 25, 2012 at 13:37
Wasn't this supposed to be broker independent ?


'he first million dollar robot that isn't broker sensitive.'

Another lie?
Better to lose an oportunity than a money
Участник с Aug 20, 2011   588 комментариев
Jan 25, 2012 at 13:43
depend on what's the broker sensitive means.
if its due to latencies, spread or some other common problems, then ok.
from a parameter point of view, I will never see any EA working correctly on different brokers using the default settings.

for me it was clear in the documentation that there is a need to be optimised for each account.
so for sure the 'broker sensitive' is not related to the parameters.
ForexScam
forex_trader_32776
Участник с Mar 28, 2011   1008 комментариев
Jan 25, 2012 at 16:06

 
You can add 150 - 300 a week on the filters up to the max
range to improve results.


Keep Fine tuning until your 30 days pass by. 🙄

if you add fractals to your M15 chart, your trade will enter right after the new fractal.
If its a Down fractal, you'll be Long,
If its an Up fractal you'll be Short,

The concept is ' Bounce off support and resistance and ride the longer momentum.

 If, nothing happens within 9 hours, or if Monemtum goes the opposite way, close the trade and/or adjust( tighten) the SL.

This bots Logic isn't sophisticated enough to run 24/7.

of course you have a Filter depth to decrease or increase activity, but how is this going to help?
You must be connected to Myfxbook in order to leave a comment
*Коммерческое использование и спам не допускаются и могут привести к аннулированию аккаунта.
Совет: Размещенные изображения или ссылки на Youtube автоматически вставляются в ваше сообщение!
Совет: введите знак @ для автоматического заполнения имени пользователя, участвующего в этом обсуждении.