Alpha9 EA *** Released to TSD Elite Members *** - page 86
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
Alpha with RSI settings for small deposit size performed very good. But as you see - I am making the statements weekly. But Alpha EA with this settings - it is not trading often: it can be 1 or 2 weeks without any single trade.
That is why I stopped this Alpha for small deposit size.
I will start it with new version of Alpha and it is really good idea to keep/create the settings/version of this EA for small deposit size.The new version being v20? v21?
I do not remember the numbers sorry ...
New version will be with Ichimoku indicator for market condition checking/forecasting.
So, on a small account, it's alright to use v19 with the RSI settings until something new is available, right?
Thank you.
I am using 1.20 and 1.21.
1.21 version is more good so you can use it.
Or wait when new version will be uploaded to this thread (it was started to program but not finish yet).
...
Some answers :
__________________________________
As of ArrrayCopyRates errors : they can be somewhat minimized if you manually "force" download not only 1 minute data but data for time frames from 5 minute to monthly for the symbol you wish to test (I normally test 6 to 12 months that way) I too hope that they are going to solve that issue (ArrayCopyRates is a completely legitimate metatrader function and having it work like this is not too understandable. With a lack of metatrader API we are forced to use what we have and it limits us in more than one way)
Frankly I do not understand to much the metatrader "upgrade policy" that posts an upgrade and does not tell us a single word of what has been changed or fixed or anything similar
As of version 5 we all know the drill : baby illness first and only after some time there will be a more or less stable version, so ... We are certainly going to rewrite code to version 5 but please be prepared for those "baby illnesses" too
__________________________________
As of hopes in this platform : this platform is a sound one and it deserves additional work and time invested in it. We already have some "future" versions, but we are trying to solve some issues that we know that can happen. I will elaborate that a bit more (hope no one minds) :
For start see the attached equity curve (just the curve itself) : that is an EURUSD test with default settings of what Alpha might become (I will remind that it is not a "no stop loss" test since alpha always has a stop loss - hidden or not). The problem is that it behaves like that on one broker and on another it behaves differently. What we are trying to solve right now is that "behavior" (even at a cost of results not so good for the broker where these results were achieved)
Also, as newdigital already mentioned, the deciding logic of Alpha is what we are working on. I would not like to make announcements since it than sounds like it is already done. We prefer to forward test things and than, if the results are OK, we are going to post EAs (that way there will be no "fake hopes")
till new versions
regards
mladen
...
Makes me feel that you have serious hopes on this one as a platform for multi-TF trading.
However, when i run 1.21 in tester mode on a laptop with little memory i get the "arrayCopyRates error - internal error" for EACH call. Even if i backtest only one day's data, and shut off trading on any other timeframe than M5 and use only one currency pair. So there is indeed a substantial bug somewhere. Is there any hope that it will be fixed, or shall we just wait for another rewrite under MT5?Some comments (and observations) of how to use Alpha :
Alpha is by default, set to use monthly, weekly, daily..., m5 heiken ashi values to determine the trend. But also, there is an interesting effect when one is trying to find "alignment" (when they show same trend) for multiple time frames : it seems that those tend to align as a kind of "forecast" for higher time frame.
What does it mean :The interesting part is that those "alignments" really do tend to go in the direction that the next higher time frame is going to have and opposite alignments are rather rare (this "alignments thing" I have been testing on EURUSD and it almost never failed, it rather caused faster entries and avoided some late false direction entries) I am not saying that it is a "holly grail of prediction" but the effect is interesting and, coming to think about, logicalYes, as logical as that the next hurricane in the Carribean will start with someone sneezing in Africa.
Trends of course do start on the smallest TF, and are confirmed on higher and higher TFs. The trick is to sort out the false trend deviations on the lower TF. Alpha is doing that, in a rather peculiar way, looking at a trend deviation in a lower TF, which is then followed on the next bar of a deviation in the next higher TF. However, if the lower TF returns to the previous trend, the devation is discarded as a false one.
There are other ways of sorting out if a trend is changing, though. I have rebuilt the trading locic of Alpha to search for trend confirmation on higher TF, much along the thoughts of mladen above. But - unlike Alpha9 - trading will always only take place on M5, because that's where the trend change actually starts!
As a result i get a much more active MTF trading platform, but keeping all the safety measures built into Alpha9.
Currently, i'm optimizing parameters on EURUSD only.
Also, I've added a trailing SL based on the OS Gaussian Support Resistance, as well as a TP based on breakout from Bollinger bands with 3 st.dev. The SL and the TP are the two weakest points, they are a bit too sensitive for parameter changes. As a result, optimizing is unlikely to provide good guidance for forward trading.
Is there anyone who has any ideas how to improve SL and TP?
Of global variables : Alpha 1.21 adds prefix "testing-" to global variables names when testing (simple solutions are almost always the best
) and that way Alpha can work with global variables even when tested. I recommend you apply that kind of global variable treatment, since it is as close to "real thing" as it gets (you can not get actual spreads for variable spread brokers that were in the moment of testing, but then, nothing is perfect)
mladen,
I'd like to clarify further about the treatment of global variables when testing. I'm already testing/optimizing and all seems to work fine, will soon start real-time forward testing, so I'm worrying that backtesting efforts are not relevant due to the argument you raise about MT's inner logic.
In the code, you replace the following five variables while testing:
Alpha9EqFlag
AccountEqty
EquityTrigger
MaxDrawdown
DrawdownTime
As I understand it, other global variables, such as TakeProfit at different TF's, numerical values for SMAfilters, etc - don't need to be handled separately? Would appreciate some further explanation as to what qualifies a variable for separate treatment during testing!
Thanks for your help...
kiasom,
If you are using 1.21 as basis you don't have to worry about global variables (in earlier versions global variables were completely shut down when testing)
In Alpha (and any other EA testing) the one that is making a substantial difference is the spread : as you know there is no historical data for spread that were valid in some historical moment, so, in those cases, metatrader uses what is a "normal" spread. By doing so, automatically the maximal draw down can be altered (if stop losses are not hit) So that is the only global variable used by Alpha for which we can say that it is not accurate. But since it is used for informational purposes only...
_________________________________
The rest of variables do not need any additional handling.
Just be careful not to curve fit (optimization is OK, curve fitting is not)
regards
mladen
mladen,
I'd like to clarify further about the treatment of global variables when testing. I'm already testing/optimizing and all seems to work fine, will soon start real-time forward testing, so I'm worrying that backtesting efforts are not relevant due to the argument you raise about MT's inner logic.
In the code, you replace the following five variables while testing:
Alpha9EqFlag
AccountEqty
EquityTrigger
MaxDrawdown
DrawdownTime
As I understand it, other global variables, such as TakeProfit at different TF's, numerical values for SMAfilters, etc - don't need to be handled separately? Would appreciate some further explanation as to what qualifies a variable for separate treatment during testing!
Thanks for your help...Indicators and settings on the EURUSD
Hi Guys
This is the first time I am looking into the Alpha9 and are wondering if I need to use the two indicators posted in the first post with the 1.21 version? The Heiken_Ashi_Smoothed.mq4 and VHF_v1.mq4.
Mladen, regarding your testing on the EURUSD, are there any specific settings that you could share (in addition to setting monthly and weekly trend to false)?
Thanks
#import "libAlpha.dll"
Hi,
When looking into the code I saw the #import "libAlpha.dll".
Is this something I have to manually put into the "libraries" folder?
Thanks