Forex Software — Good Strategies - how long use and then?

Re: Good Strategies - how long use and then?

2019-02-14T16:04:24Z

ok, i understand

Re: Good Strategies - how long use and then?

2019-02-14T16:00:55Z

rantampla
In the multi-market use positively correlated pairs, negative correlated pairs and uncorrelated pairs. To find them use internet, they are free services, the number is at your discretion . Before checking the strategies you choose immediately the pairs to insert in the multi-market, do not decide after finding the strategies!

Regards

Re: Good Strategies - how long use and then?

2019-02-14T15:05:27Z

Yes, that's how we do it.
Last night I already started to set up 2 instances of EA Studio.
1) 10 years data taken
2) Optimizer off
3) All other Robust Test
4) Multi Market equipped with the top 7 pairs
EA Studio has been running for over 12 hours and has only found 1 strategy that has passed the tests, but does not look all that neat.
I will now run EA Studio permanently until 100 is full, then I look which I can delete, then again to 100, etc ...
the 100 then fit on the last 25% data. Then maybe I can all throw away or stay 1-2. Am very excited

P.S.: I am gladly for our discussion here

Re: Good Strategies - how long use and then?

2019-02-14T14:59:52Z

rantampla
Ok, concentrate on how to use the machine, not on generating many strategies because more strategies generate more the probability that it is just luck increases. Use my advice: 50%, 25% and 25%; use at least 10 years of history maximum 20; if you find a strategy a month or every 2/3 months you're on the right track, and if you find a strategy a week you're wrong; be patient because it is not you who decides which strategy is winning or losing but the market; if your strategy so developed passes my suggestion try it also with 1,2 or 3 pip of spread; never optimize, you're not ready, you need other things to make money; observe the returns of top traders or speculative funds on forex, you'll be realistic and aims at those performances.
Good luck!

Regards

Re: Good Strategies - how long use and then?

2019-02-14T14:43:07Z

ok, I almost thought so
I continue .... would be great if the forum continues to live. I think it helps everyone to exchange
Thank you

Re: Good Strategies - how long use and then?

2019-02-14T14:40:21Z

rantampla
It's a good question and I'll give you a horrible answer because you will not find it here on the forum and it's not the same for everyone: a good strategy is the one that beats randomness, here the odds come into play. Unfortunately, identifying a good strategy, in addition to a solid construction, requires that you be able to prove to be better than other traders (not speculative funds and top traders, I refer to retail). If you can not do it you'll never know if you're lucky (you'll lose money over time) or good. These probabilities change over time and are linked to many factors so if, for example, I provide you my calculations you could lose a lot of money because we operate differently.

Regards

Re: Good Strategies - how long use and then?

2019-02-14T14:22:21Z

Yes, I already understood that. Of course, I am also willing to learn. But what is a good strategy? I have had strategies that went very well for several weeks. Some even 2-3 months. Then no more.
What do you think about what a good strategy is?

Re: Good Strategies - how long use and then?

2019-02-14T14:01:34Z

rantampla
1) You have not grasped the point: manual trading and algotrading with machine learning are completely different things, you're not a good algotrader because you have been trading since 2004 by hand. If you have studied oriental languages in your life, do you think you can become a doctor because you have studied so much? no, you can not, you have to start studying again. I've already written that before optimizing a strategy you have to learn how to build a good one and you can not do it now. If you optimize a losing strategy the result will be an optimized losing strategy. Almost all the people who try to become traders are in your condition and I am helping you to overcome this obstacle. You can not make money by beating the best traders or the best hedge funds, you can earn it by beating those who can not trade okay? To overcome this obstacle you do not need to optimize a losing strategy but understand what to do in order not to lose money, requires research / study / many tests (about 1,2 or 3years for me)
2) when you will be able to do what is written above then you will realize that the management of the portfolio or of a strategy is even more complex, much will depend on the reference market, much the type of strategies generated.

Regards

Re: Good Strategies - how long use and then?

2019-02-14T12:40:54Z

Thank you for your response. However, I do not quite agree. I am not a bloody beginner, I deal live since 2004 with Daytrading. In manual trading I have synonymous with normal results. However, it is simply not possible for me to sit in front of the PC all day because I still have 2 companies to run.
With the software here I have actually imagined to simplify much and solve the time problems. I misjudged it, that's right. However, I am still convinced that it can continue to work. Therefore, I am very grateful for your hints, they have given me the way, which was done, something directed. It will show how this will help.
Still, I'm still unclear and as I read in other forums, from the competitor, what do I do when the strategy starts to lose.
a) I import it back and optimize it
b) I throw her away and take a new one

Both variants are conceivable.
to a) in other forums there were attempts of daily re-optimizing, unfortunately, this has not been tested long enough, because this is too expensive.
I would rather go in the direction of weekly or monthly re-optimize .... and I'm missing something. Namely how do I properly optimize when the time is right. Does it make sense to run the optimizer with + -5 points or more? Should one rather use the validator instead of optimizing the strategy individually?
These are the actual questions to which I can not find an answer ....

Re: Good Strategies - how long use and then?

2019-02-14T09:18:56Z

Rantampla
Fast answer: no, you should not. You do not have the knowledge/skills to do it like almost all users here. footon would not agree but I showed you with a simple suggestion of being able to deny you. You probably read all the posts of this forum but it is not the solution to become a trader, to become one you have to study more serious things. From personal experience I say that there are 3 important components: you, machine learning software and trading platform. You are not a trader but simply a person who experiences machine learning. Take one step at a time, take your time (1, 2 or 3 years), do not confuse things and you'll improve a lot. Instead of asking questions here on the trading forum, look for the answer elsewhere because it is difficult for a novice to find useful information here, you should understand what is wrong and what is right to study but you are not able to. Follow big traders like Dr. Ernest P. Chan (QTS CM) and others, look at their performance and be honest with yourself. You do not have to do what they do but understand why they manage millions and you lose, you have to understand why some have failed and so on. If you do not understand this you can use 1, 10, 100 machine learning software but the result is that you will lose more and more.

Regards

Re: Good Strategies - how long use and then?

2019-02-14T00:07:59Z

Sorry another Question:

When i check the last 25 % with my own OOS Data. Can i check this in the Validating tool or is that not a good idea, better only load in strategie?

Re: Good Strategies - how long use and then?

2019-02-13T23:40:03Z

Yes you are right. The whole thing really brings new momentum into my thoughts. It's a terrible experience to test for weeks on demo, test with MT4 Tracker and then push on Live to see how it gets worse and worse, much easier and faster. I'm even convinced that if you find good strategies now, you can skip this week-long demo Gesteste and can certainly take 1-3 months directly on live some, but must constantly run the reactor to keep finding new strategies. Furthermore, I think, well checked weekly with MT4 tracker, you can have everything well in view. I hope Sleytus comes back and extends Sidekick on EA Studio, I would love to use it here.
I thank you very much again!

Re: Good Strategies - how long use and then?

2019-02-13T23:24:56Z

Rantampla
It's a pleasure. Yes, the goal is to save oneself from statistics and data mining. Remember to use 50% with IS, 25% with OOS and 25% "remove" it from the software with the time limit settings, leave it only on MT4 and then test the whole period (100%) in MT4. If you have not found a good strategy after a few days, you're on the right direction. If you find a single strategy in a month you're already good/lucky. Once you have found the strategy with this method you will have to start examining it carefully. Trading is a minefield, if you want to survive and make money you have to study and study. I was referring to this at the beginning when I told you that you are the problem. Whatever site or blog you open, be careful. Trading with fake money is useless, if you build a strategy with a solid foundation you can immediately put it live. Try to build 1 or 2 and then put them on demo, if they work forget the demo and use only real money with 0.01 lots until you feel safe.

Regards

Re: Good Strategies - how long use and then?

2019-02-13T22:52:21Z

Wow, the first test was over. That's crazy. All the strategies that I have previously run in OOS have been directly attached to IS, now with self-set OOS data, as suggested above, suddenly totally bad. It's like making a 1-year live trading with it. Can not imagine it differently. That's amazing, what a great tip. Now I'm curious if in the next few days a few useful strategies will be left. Many thanks again for the tip!

Re: Good Strategies - how long use and then?

2019-02-13T18:28:42Z

footon
If MT4 were reliable I could avoid the test on it, but we agree that it is not. Is the platform on which I trade? yes, personally this is enough for me. If I have to choose I do it, no one of us has predictive capabilities and FSB can not guarantee the correct functioning of MT4. If you want to know I test eas on different brokers and, to be worthy, the algorithm must behave well everywhere. This path is a crushing stones