현재 랭킹 3위의 The_Wild_13이라는 EA가 눈에 띄는군요. 대회마감일이 얼마 남지 않은 상황에서 3위를 기록중이며 별다른 이변이 없다면 3위이상의 성적으로 마무리 할것으로 보입니다.
다음은 The_Wild_13의 밸런스 그래프입니다.
보시는것처럼 트렌드는 우상향을 그려주는군요. 그래프만 봐서는 자금관리형 또는 마틴게일형 로봇인듯 합니다.
아래는 해당 포럼에 올라온 내용을 적어봅니다. 추후 번역해서 한번 읽어봐야 겠네요.
포스트 #1
Hello,
i’d like to try to answer the questions some of you asked:
I started getting into MQL4 about a year ago. Still i’m far from beeing a professional programmer and have more ideas about trading systems in my head than beeing able to programm them. But since i started i learned a lot and the EA now taking part in the ATC 2011 is the result of this learning process. Back at the time the ATC 2010 ended i finished my first EA that – although it followed quite a somehow complex strategy – traded more like it followed chaos. These results took from me my illusions about receiving the holy grail soon. But then Avoitenko got to the 4th place in ATC 2010 and was so generous to publish his code. I was fascinated by the possibility to have a look at the code of an EA that made it to the 4th place in the ATC. So i read Avoitenkos code and at the moment i did understand it from A to Z i was surprised how simple yet effective it was. It inspired me to take up my experiments in MQL4 again.
So roughly following the ideas of Avoitenko, i started to develop a simple breakout system based on the ZigZag. That’s where this EA taking part now comes from. However what makes it more succesfull then just that is the fact that – to my believe – a promising strategy has to be well programmed and is the result of a – what sounds unusual in the field of programming – creative process. It took me a lot of try & error and coding mistakes that revealed as improvements to finally create this EA.
Since me and my brother (with whom i developed this EA together) haven’t had much time to adapt the EA from MQL4 to MQL5 it is coded and trades a little rough. Thats the reason for the 0.01 Lot Orders and the many deleted and recreated PendingOrders. Since there are always PendingOrders of the Buy and Sell type at the market we had to find a simple way to always bring the full amount of the allowed LotSize to the direction the market is trending while saving the location of the “almost deleted” orders with the 0.01 Lot Orders.
Me and my brother would be very happy to at least make it to the top-ten of the ATC. Although MQL4 offers a lot of possibilties i however believe that a truly profitable and reliable system has to go beyond MQL4 and the use of any indicator. I think to go there one has to get aquainted with Machine Learning Techniques and their use in MatLab and R.
@ APeng: keep experimenting – and consider that a promising EA in MQL4 might be rather simple than complex. I think a good EA depends less on signals of different timeframes or very many indicators or hundreds of parameters BUT on a well written code and a strategy that evolves in live trading, optimisation and making mistakes
So far …
포스트 #2
Hello,
the EA is using the ZigZag that comes with the Metatrader 5.
@ APeng: The EA is not using an indicator (MA or MACD etc.) in order to consider trends but does that by treating the pending orders in a certain way. I think – since the EA follows a breakout system – using some indicators in order to tell it about a higher trend would cause more harm than improve its trading very much. Because i found that the problem with estimating trends on higher timeframes is that the higher the timeframe the more reliable is the indicated trend of course but also the slower the EA reacts on a change of the trend. Too much noise is bad for the trading but almost no noise at all is maybe even worse. I wouldn’t say that calculating trends on higher timeframes makes no sense but then it is hard to find the right timeframe (1W or 1D or maybe even 1M ???) along with the right indicator (MA, JJMA, AMA, MACD, …) and finally the right parameters for that indicator. So there you have a lot of unknown parameters that might make the EA too sensible for once and also might have very little effect compared to the amount of ressources they require.
But: As i worked almost only with the ZigZag until now i haven’t got much experience with other systems. So all i can tell you is that its worth getting in the direction of a breakout system based on the ZigZag. But then i also think almost any direction will be rewarded if followed consistently. As you said yourself, its a “neverending topic”.
포스트 #3
More and more waiting for the first loss trade – it’s getting a little weird.
Telling you the ZigZag Parameters would mean nothing since it is the programm that makes them what they are.
짧은 영어로 힌트가 될만한 포스트들만을 선별해서 옮겨보았습니다.
2010년 참가이후 올해도 참가하려고 했었으나 과중한 업무로 ㅠㅠ 참가 못해 못내 아쉬웠는데 이런저런 데이터를 모아서 내년엔 꼭 참가해봐야 겠습니다.
