Welcome to Forex-TSD!, one of the largest Forex forums worldwide, where you will be able to find the most complete and reliable Forex information imaginable.
From the list below, select the forum that you want to visit and register to post, as many times you want. It’s absolutely free. Click here for registering on Forex-TSD.
Exclusive Forum
The Exclusive Forum is the only paid section. Once you subscribe, you will get free access to real cutting-edge Trading Systems (automated and not), Indicators, Signals, Articles, etc., that will help and guide you, in ways that you could only imagine, with your Forex trading.
Elite Section
Get access to private discussions, specialized support, indicators and trading systems reported every week.
Advanced Elite Section
For professional traders, trading system developers and any other member who may need to use and/or convert, the most cutting-edge exclusive indicators and trading systems for MT4 and MT5.
i am using Interbank demo server to test DIN_KusKus_EA 2.4. The backtest is not giving any good result. What server do u guys use? Can the EA be done in such way, like the buy or sell will only be executed when Kuskus Starlight changes from negative to positive or vice versa? And there will be only one trade when the signal occur?tq. I found out that Kuskus is very accurate in 1 hour tf, and if you only buy or sell when Kuskus starlight just changes from negative to positive or another way round, basically u will have more than 80% winning trades. I only look at Kuskus starlight and Stepma_stoch.
To test on IBFX you need to change the trading times to match your server times. The tests I run are all with FXDD server times which are 2 hours different than IBFX.
So you need to subtract 2 hours from the tradining hours settings.
Especially for MiniMe
2006.08.18 - now
Results for Tenchi_v1:
risk 0,1
Drawdown 17.76%
Profit 56k
Profit factor 2.38
risk 0,08
Drawdown 14.35%
Profit 37,5k
PF 2.49
85% winning, 8 wins in a row, 1 loss in a row (average)
I am not sure how Tenchi_v1 works but now when fister==0 and MA 24 & 50 crossed, HMA=0, then I got two future trades, sell and buy stop. interesting
-------
the first order sell stop was placed and the second order buy limit is still there and is not cancelled !!!
------
Now its cancelled after the first position start to make some pips
EA version 1.6 is pure Isakas mode , ver 1.8 and 2.4 have more than Isakas rules... all the rules are met or no trade will occur and please read Mrpip comments.
Tests are done on database collected over time, which you can download from alpari site, there is a procedure to run backtests to achive 90% you can find it somewhere on forex-tsd.com, so running interbank demo server or any other server does not matter for backtest, as they are done offline, although the data from one broker to another may differ but the end reults will not be very big difference
One of the biggest issues with backtesting is that markets change. You must remember that during 2005 and most of 2006 the feds kept raising the interest rate. In 2005 I had an ea that traded on the daily cross of lsma 7 and 16. No take profit and a very large safety stop loss. Starting with $5000 on Jan 2 ended the year with over $140,000. In 2006 it failed big time and continues to fail.
Another EA I wrote ran on GBPUSD 15 minute chart and had 0 losses from March 2006 til October 2006. It only traded at most 2 times a week because of the filter I used to enter trades. Since then it has about 50% wins. When I run the backtest the trades match almost to the pip.
A more recent EA started with $3000 on Aug 20, 2006 and reached a high of $45000 on Feb 6, 2007. Maximum drawdown at that point was only 10%. The balance is currently $31,000 with only losses since Feb 6 and now drawdown is 32%.
The reason I run the backtest on lower timeframes is for greater accuracy of exits. Trades are entered on closed candles but when stoploss or takeprofit is hit the backtester is more accurate running on lower timeframes. The signals from the indicators are still from the time frame set for SignalTimeFrame input. When I display the chart and switch to the 30 minute the signals look correct for entry.
I have run backtest on alpari and fxcm data and often get different results.
This is even more true with IBFX where there are often spikes that are not found with FXDD or Alpari. Stop hunting??
Sorry for the long post but when an ea fails for 2005 that does not mean it will fail for 2007. And because an ea is great in 2005 does not mean it will continue to be profitable.
i am using Interbank demo server to test DIN_KusKus_EA 2.4. The backtest is not giving any good result. What server do u guys use? Can the EA be done in such way, like the buy or sell will only be executed when Kuskus Starlight changes from negative to positive or vice versa? And there will be only one trade when the signal occur?tq. I found out that Kuskus is very accurate in 1 hour tf, and if you only buy or sell when Kuskus starlight just changes from negative to positive or another way round, basically u will have more than 80% winning trades. I only look at Kuskus starlight and Stepma_stoch.
Kuskus Starlight was replaced with Fisher_M11. If you only want to use that indicator you need to set all other switches to 0. Keep the switches for useFish set to 1 on both entry and exit. Then for only trading on the change from negative to positive turn off time filter by setting useTradingHours to false. Backtest with Fisher only does not do to well. I can only assume you are using different settings or some other rules like a lower take profit that are not found in the current ea.
Nice to see you are still around.
If you still need some pivot indicators to get ideas from them please notify me, I have a handful copies of them.
(I think it'd better to keep this thread tidy so I would not like to post them here.)
Quote:
Originally Posted by Shinigami
okay lets see...
The "thing" I used to get that result (I can't call that an EA because it works only in the backtest) makes great results which I can't beat. Only in backtest though.
I'm posting here the new EA I wrote. Yes, its a new EA, it doesn't use any indicators. I'm posting .ex4 so you can play around with settings, optimal ones for 2006 are defaults (optimals I found, using SL of 200 makes things even better but thats almost a suicide)
If you have any problems with it, let me know.
Dominic-chan you rock! Thanks for noticing! And sorry for being that sarcastic but thats exactly how experts work
If every time we got a new tick we tested if we crossed EMA and all other conditions, it would take years to backtest such an EA.
Instead we can use a veeeery simple thing:
if(prev_time==Time[0])
return(0);
prev_time=Time[0];
this way we would only work (open new positions, close old positions ect ect) once per bar, on the start of the new bar (seems to be the best solution I've found)
Nice to see you are still around.
If you still need some pivot indicators to get ideas from them please notify me, I have a handful copies of them.
(I think it'd better to keep this thread tidy so I would not like to post them here.)
chrisstoff
Don't worry you can post whatever you want, since its related to development of the EA .. if they are too many just put them on a zip file.
I found that there are some sites that we can use to upload then link the upload here, so we don't use much space on forex-tsd.com, doing that we avoid adding extra cost to the owner of forex-tsd.com http://www.zupload.com/ http://www.imageshack.us/
okay lets see...
Instead we can use a veeeery simple thing:
if(prev_time==Time[0])
return(0);
prev_time=Time[0];
this way we would only work (open new positions, close old positions ect ect) once per bar, on the start of the new bar (seems to be the best solution I've found)
This method is already built in to most of my EAs by checking exit signal from prior closed bar. This is the same method that is done to open new positions.
The signals from the indicators are checked for the most recent closed bar.
Checking time of bar does have one advantage in that it would speed up testing by only checking indicators at the start of each new bar instead of every simulated tick.
However when I run backtest on 1 minute timeframe with signals on 30 minute time frame there is not much savings.
This method is already built in to most of my EAs by checking exit signal from prior closed bar. This is the same method that is done to open new positions.
The signals from the indicators are checked for the most recent closed bar.
Checking time of bar does have one advantage in that it would speed up testing by only checking indicators at the start of each new bar instead of every simulated tick.
However when I run backtest on 1 minute timeframe with signals on 30 minute time frame there is not much savings.
Robert
Run check for signals every 30 minutes and open/close positions on closed bars only. You can trail on M1 or M5 or every tick. As for me, I only open positions twice a day and trail every tick. And it already takes a while to run a test on a year's history. If I find a way to make experts run faster, I'll let you know!
Linear Mode
