DislikedOanda is what I use.
MT has always been demo, only to test, apply and learn indicators.
Currently I have all setups in "Al Trade MT". I switched to this one only because it has DAX Futures, which I've been trading with CFDs in my local broker, for 1 year;
Last before it was "FxPro", which is still installed in my laptop.
But speaking of brokers and switching to a MT one, I will need/apreciate some advice from you all, concerning which one...
Having mini-lots... having low spreads... of course, being serious... are some of the conditions I find essentials.Ignored
EMA 25-50 Trading System (Detailed)
- #2,164
- Aug 11, 2008 7:07pm Aug 11, 2008 7:07pm
Corre, you mean the big increase in black crosses (and consequent low winning) in EURUSD? I only looked at 2007.
That is unavoidable... two things you'll learn over time are:
Markets don't move the same over time. There's always something different. What's obvious to me, watching 2007, is less trending and more ranging, and also very low ATR (12 to 16 most of the time);
One good Indicator/EA today can turn into a crapy one tomorrow... exactly due to those changes in the market.
We have several editable inputs which can serve the purpose of adapting over time... but there's always a delay (the time that passes until you noticed current setup isn't right anymore).
What did I told you by PM earlier today? There's certainly things to look at and improve or correct
Look at some of the code Trader posted here days ago, about his TPs;
His TPs are coded so that they're different under different ATR conditions... It can be a possible approach to that, too.
That is unavoidable... two things you'll learn over time are:
Markets don't move the same over time. There's always something different. What's obvious to me, watching 2007, is less trending and more ranging, and also very low ATR (12 to 16 most of the time);
One good Indicator/EA today can turn into a crapy one tomorrow... exactly due to those changes in the market.
We have several editable inputs which can serve the purpose of adapting over time... but there's always a delay (the time that passes until you noticed current setup isn't right anymore).
What did I told you by PM earlier today? There's certainly things to look at and improve or correct
Look at some of the code Trader posted here days ago, about his TPs;
His TPs are coded so that they're different under different ATR conditions... It can be a possible approach to that, too.
- #2,165
- Aug 11, 2008 9:46pm Aug 11, 2008 9:46pm
OK...
To make it possible to test in fixed past periods, different from " x until today", I've modified the Indicator once again.
It is now possible to choose 2 dates to count in between them, and so search the best configuration for different market behaviors...
I also did a cleanup to unnecessary and deactivated code.
Now you people test it... 'cause I'm going to
To make it possible to test in fixed past periods, different from " x until today", I've modified the Indicator once again.
It is now possible to choose 2 dates to count in between them, and so search the best configuration for different market behaviors...
I also did a cleanup to unnecessary and deactivated code.
Now you people test it... 'cause I'm going to
- #2,166
- Aug 11, 2008 11:02pm Aug 11, 2008 11:02pm
Just for the record, I haven't found anything wrong with the counter, and I also get those results in 2007.
If you put a test period that only has 1 or 2 trades, you will observe it's counting right.
In fact, 2007 is a bad year, compairing to the present...
I think there can be 2 approachs to improve the results:
1) Either you try to optimize the configuration as the behavior/results of a pair starts to change (with the delay realizing that means, you'll be loosing big time);
2) You test different configurations in several past time periods, and try to find common configs that give you good results overall (sticking to those).
I'm for the option 2...
I prefer winning less than potential maximum profit in a certain period, if it means winning more than the minimum profit (or simply not loosing) in a bad period.
For example, and sticking with EURUSD, I did some tests compairing beginning of year until now, of both 2008 and 2007 (same periods of different years):
Note: First I had similar result as Corre (arround +2000 pips) for all 2007, with the configs that were giving +9700 for these last 8 months;
After playing a little with it, I came up with a config that gives even more for these 8 months, and more than doubles for all 2007:
these 8 months = +10172;
all 2007 = +4909.
The changes made to defaults in the config was:
- Stoch_Slowing = 5;
- Stoch_MA_Method = 2;
- Use MACD = False;
- Multiplier = 2.3.
By the way Corre, I forgot to mention that when I've been testing "from beginning of year", it was with the default date of 2008.01.15... not the first day - maybe it contributes to our differences;
Also now, testing past year, I leave the first and last 2 weeks out of calculation.
Now I'm really going to 
If you put a test period that only has 1 or 2 trades, you will observe it's counting right.
In fact, 2007 is a bad year, compairing to the present...
I think there can be 2 approachs to improve the results:
1) Either you try to optimize the configuration as the behavior/results of a pair starts to change (with the delay realizing that means, you'll be loosing big time);
2) You test different configurations in several past time periods, and try to find common configs that give you good results overall (sticking to those).
I'm for the option 2...
I prefer winning less than potential maximum profit in a certain period, if it means winning more than the minimum profit (or simply not loosing) in a bad period.
For example, and sticking with EURUSD, I did some tests compairing beginning of year until now, of both 2008 and 2007 (same periods of different years):
Note: First I had similar result as Corre (arround +2000 pips) for all 2007, with the configs that were giving +9700 for these last 8 months;
After playing a little with it, I came up with a config that gives even more for these 8 months, and more than doubles for all 2007:
these 8 months = +10172;
all 2007 = +4909.
The changes made to defaults in the config was:
- Stoch_Slowing = 5;
- Stoch_MA_Method = 2;
- Use MACD = False;
- Multiplier = 2.3.
By the way Corre, I forgot to mention that when I've been testing "from beginning of year", it was with the default date of 2008.01.15... not the first day - maybe it contributes to our differences;
Also now, testing past year, I leave the first and last 2 weeks out of calculation.
Now I'm really going to
- #2,169
- Aug 12, 2008 8:03am Aug 12, 2008 8:03am
I am more than happy to assist with testing, however I'll need to know the specifics of how you are doing it. If it were an EA, would be real nice to just pop that into the strategy tester, setup some optimization parameters and have it crank out the best settings.
Are you just manually scrolling back, and entering the trades into a spreadsheet? How are you basing the price, opening price of bar after the cross? How are you basing the exit price, bar open price after a recross?
Anyhow, if you want to share how exactly you are testing, I can help with the project if you'd like.
I was suppose to answer this yesterday, but I got absorved by calculating
Well, until now the tests have been very light and simple, because the Indicator always had some functionning bug.
About the exit price (both SL & TPs), the system calculates the exact price (based on Multiplier and ATR inputs), and if the candle reaches it (max or min of candle), it sums that value to the counter.
It's not based on crossings neither openning of bars.
- #2,176
- Aug 12, 2008 9:00am Aug 12, 2008 9:00am
Another example (moved to other pair):
GBPUSD:
2008 (these first 8 months): +6574 pips;
2007 (same first 8 months): +7133 pips.
See?
In the first one I've tried, I've found settings that give similar results in both 2007 and 2008, even better last year.
Take out some thousands if you want, for differences in servers
GBPUSD:
2008 (these first 8 months): +6574 pips;
2007 (same first 8 months): +7133 pips.
See?
In the first one I've tried, I've found settings that give similar results in both 2007 and 2008, even better last year.
Take out some thousands if you want, for differences in servers