EA request for profitable gbp/jpy system - page 8
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
backtest reliability
I do not know anything about programming. I just know that in order to have reliable backtest, an expert should be programmed in such a way that he is only checking for new entries/exits at the close of each bar. Then the backtest with option "open price only" will be fully reliable.
I do not know anything about programming. I just know that in order to have reliable backtest, an expert should be programmed in such a way that he is only checking for new entries/exits at the close of each bar. Then the backtest with option "open price only" will be fully reliable.
When you're letting the server take care of your stoplosses, that can be true, but in order to make sure we don't reenter the same trade without the conditions changing first, we needed to contol the stoploss in the EA. If we only look at bar open/closes, our stoploss won't be closed when hit. Rather it will be closed when the next bar opens AFTER it's been hit. With a good running bar, that could mean a lot of additional loss. With the 1h timeframe on the GBPJPY, that could easily pass an additional 100 pips.
FYI: Entries are based on the results at the previous bar. While the EA will check for new entries at every tick, it won't actually enter a new trade except at the open of a new bar (with the exception of when it's loaded onto the chart for the first time, or the EA/MT4 is restarted and it isn't in the middle of a trade.
-MRE-
When you're letting the server take care of your stoplosses, that can be true, but in order to make sure we don't reenter the same trade without the conditions changing first, we needed to contol the stoploss in the EA. If we only look at bar open/closes, our stoploss won't be closed when hit. Rather it will be closed when the next bar opens AFTER it's been hit. With a good running bar, that could mean a lot of additional loss. With the 1h timeframe on the GBPJPY, that could easily pass an additional 100 pips.
FYI: Entries are based on the results at the previous bar. While the EA will check for new entries at every tick, it won't actually enter a new trade except at the open of a new bar (with the exception of when it's loaded onto the chart for the first time, or the EA/MT4 is restarted and it isn't in the middle of a trade.
-MRE-Thanks for explanation
After taking a look at it, I believe this system is only profitable with the aid of FXCM's bullshit charts. In real world trading, it won't be nearly as profitable or profitable at all. QQE does not match at all with MetaTrader's and FT's platform. You are at a severe disadvantage using FXCM's data.
Thanks for explanation
No problem. After reading what I typed, I hope it didn't sound like I was getting defensive about the issue at all. I want this kind of discussion and suggestions to emerge, so please don't feel I was jumping on you about this or anything. Besides, you're completely right in that it should have a real MagicNumber system. That'll end up being incorporated into v0.4, but I'd like to get a good amount of testing under this current incarnation of the EA.
-MRE-
After taking a look at it, I believe this system is only profitable with the aid of FXCM's bullshit charts. In real world trading, it won't be nearly as profitable or profitable at all.
That is a very real possibility. I have found that a lot of systems that have been used manually are reported to do much better than they do if the rules are followed mechanically over every instance available. I think either there's often a bit of discretion used with manual systems (even when the user doesn't realize it), or when first evaluating a system found, the user often uses earlier entrance and exit points than otherwise noted. For example, in this strategy we wait until the close of the bar before we make any decision on a bar, thus putting us behind price action by up to an hour. When traded manually, the user can determine whether the two lines crossing is from a "strong" or "weak" movement. If it's strong, they can choose to exit much earlier than the mechanical system allows. Additionally, sometimes people's minds may dismiss certain losses and concentrate on certain wins when first developing a system, thus convincing themselves that the system is better than it really is.
I am *NOT* saying that Lepiricus has fallen victim to any thing like this. I'm just saying I've seen this type of behavior before in trading systems. Lurk around forums enough, then test the EAs that emerge for a system, and you'll likely find the same thing a lot.
It seems testing out different SL/TP levels using 15/30min charts instead of 30/60min could prove lucrative as well. Maybe this should be evaluated on a longer-term setup as well. Who knows. I'm simply providing the ability for people to take this strategy where they want. Give people a framework for finding things out for themselves about this system. Besides, I've been getting quite rusty at programming by not doing it for quite a while, and I figured smaller programs like EAs would be a great way to shake some of the cobwebs off.
Has anybody discovered any new errors that exist with this version (v0.3.2-beta) of the EA thus far?
-MRE-
I've been letting 0.3.2 run untouched for a few days, and think I found an error, but I'll need someone else to confirm. My last two trades were sells. According to my trade history, the first was closed for a profit, but did not hit its take profit. The second one was opened immediately after the first and was closed when it hit it's take profit. The 30 and 60 min charts show no reason for that first trade to have closed. The best I can think of is the trailing stop, but that would mean the exit is at least 100 pips away from the low of the since the start of the trade, but it was only about 85 pips away from the low.
Has anyone else seen something like this? How about any other anomalous behavior?
-MRE-
I do not know who to credit for the QQE indicator, but I know it *
* was not created by myself, Matthew Ebersviller. I take no credit *
* (or responsibility) for the EA.
QQE was created by Roman Ignatov with a copyright of 2006 yet it is exactly the same as TrendStrength made by Igorad with a copyright of 2007.
Just for your info cos you had that in the EA.I put TrendStrength here cos it has about 6 variable to use possibly IgorAd made it better than QQE version.
Mode = 0; // 0-RSI method; 1-Stoch method; 2-ADX method
Length = 34; // Period of evaluation
Smooth = 5; // Period of smoothing
Price = 0; // Price mode : 0-Close,1-Open,2-High,3-Low,4-Median,5-Typical,6-Weighted
ModeMA = 3; // Mode of Moving Average
Deviation = 10;
kkkkkkkkkjhgfdkjhgfds,
I have alway wondered about that. Can we use igorads copywrited material here in tsd? or does the EA need permission?...or does the EA become the property of tsd if no permission is obtained?
Perhaps El Cido might know...
ES
Is a patent...trademark or a copyright the same? and how does one obtain rights to OHLC?
I just asked Igorad: no problem with copywrite.
TrendStrength_v1 was created by Igorad long time ago. It was somewhere in July 2006 according to the post in elite section dated by 08/07/2006.
It is public version now. It is modified AbsoluteStrength with trend view (AbsoluteStrength is Igorad's Indicator as well).
TrendStrength_v1.1 is fixed and updated TrendStrength_v1 to works good on any timeframes (it was small bug in v1 so it was fixed in 1.1 version).
Those v1 and 1.1 are public versions now.
There are 2.2 version in elite section only: version 2.2 is identical with VT version. But 2.2 is in elite section only sorry.
As to QQE so I have no idea. I only know that Roman Ignatov is not Igorad.
We may use 1.1 version as public version with no any problem.