EA Switcher - page 9

To add comments, please log in or register
Ian Smith
1595
Ian Smith  

I think I can use fractals to detect market types based on switching S&R levels dynamically. I am already switching the S&R levels, I just need to define each market type based on how frequent the switching is. Then we have a bunch of little EA (as the thread clearly stated) for trading within each type. I can already zoom in and out and scale up/down TP & SL levels. Next I'm going to make an autoswitching mechanism, so each type of trade will automatically use the appropriate ratios for the type of trade being entered.

I already have a large chunk of usable code. I think I want to get the EA I have launched and earning profits, then I can work more on making a very large switching EA.

One of my concepts is to have a bunch of signals checked all of the time. We record a set of numerical values for each indicator. Let's say there are 20 indicators involved, and we record the results from each. Normally EA have filters that process that information, and then say "if Result > X then Buy" or something similar. I'll be recording market type, S&R levels, pivot points and possibly slopes in addition to that information. We couldn't test 20 indicators for 100 settings each, but we could test 20 indicators for 1-3 different settings, and then record the returned data from each.

By putting the information into a machine parsed format, I should be able to write a Neural Network that makes connections between reversals, trends, breakouts, etc for all of the signal pairs. Then we can go over the same data again, now with values for X and do the Bayesian Calculations.

There are a few things I'd like to do first. #1 devise a method for determing TP:SL levels within a scale automatically. I'm thinking Daily range, or H4 range for each of the 6 markets (JPY, UK+JPY, UK+US, US, AU). I can already work on multiple scales already, but the issue is resizing the scale appropriate to the range. I *might* have this covered, but I'm not sure. #2 have standardized formats for reading/writing data to/from signals. #3 standardize on a type of MA used #4 try to find out how many indicators we need to support in order to trade in each market.

#4 is really a community question.

#3 is probably the easiest to answer, based on advantages of EMA over SMA, etc. Do we really need any? Many trading systems mention 'compatible' settings for using EMA instead of SMA, etc.

#2 I might have answered in a week. I'm not sure I'm ready to give up my code since I've put months of research into these questions, and I think I have something very valuable. I'd like to see other contributors as well, because this will take months more of my time.

David SWINKA
611
David SWINKA  
daraknor:
I think I can use fractals to detect market types based on switching S&R levels dynamically. I am already switching the S&R levels, I just need to define each market type based on how frequent the switching is. Then we have a bunch of little EA (as the thread clearly stated) for trading within each type. I can already zoom in and out and scale up/down TP & SL levels. Next I'm going to make an autoswitching mechanism, so each type of trade will automatically use the appropriate ratios for the type of trade being entered.

I already have a large chunk of usable code. I think I want to get the EA I have launched and earning profits, then I can work more on making a very large switching EA.

One of my concepts is to have a bunch of signals checked all of the time. We record a set of numerical values for each indicator. Let's say there are 20 indicators involved, and we record the results from each. Normally EA have filters that process that information, and then say "if Result > X then Buy" or something similar. I'll be recording market type, S&R levels, pivot points and possibly slopes in addition to that information. We couldn't test 20 indicators for 100 settings each, but we could test 20 indicators for 1-3 different settings, and then record the returned data from each.

By putting the information into a machine parsed format, I should be able to write a Neural Network that makes connections between reversals, trends, breakouts, etc for all of the signal pairs. Then we can go over the same data again, now with values for X and do the Bayesian Calculations.

There are a few things I'd like to do first. #1 devise a method for determing TP:SL levels within a scale automatically. I'm thinking Daily range, or H4 range for each of the 6 markets (JPY, UK+JPY, UK+US, US, AU). I can already work on multiple scales already, but the issue is resizing the scale appropriate to the range. I *might* have this covered, but I'm not sure. #2 have standardized formats for reading/writing data to/from signals. #3 standardize on a type of MA used #4 try to find out how many indicators we need to support in order to trade in each market.

#4 is really a community question.

#3 is probably the easiest to answer, based on advantages of EMA over SMA, etc. Do we really need any? Many trading systems mention 'compatible' settings for using EMA instead of SMA, etc.

#2 I might have answered in a week. I'm not sure I'm ready to give up my code since I've put months of research into these questions, and I think I have something very valuable. I'd like to see other contributors as well, because this will take months more of my time.

Hi Daraknor.

Thanks for your comments and your contribution.

I'm still sure this idea, EASwitcher it's the "KillerEA"... but my knowledge concerning programming is less that nothing.

I will try to support you as far i can.

Ian Smith
1595
Ian Smith  

What indicators do you think we would need to include to make an EA for each market type? What indicators to detect the market type?

ASCTrend looks like a good trend detection system but I haven't used it and I'm not aware of an EA based on it.

It looks like we'll need a few major EA?

News (schedule based)

Trend (long term)

CounterTrend (shorter term)

Choppy or Whipsaw

Channel

Breakout - I think this can be based on "higher highs vs lower lows" method. If we detect the channel, this can predict the direction of the breakout with at least a slight statistical edge.

janka
64
janka  
daraknor:
What indicators do you think we would need to include to make an EA for each market type? What indicators to detect the market type?

ASCTrend looks like a good trend detection system but I haven't used it and I'm not aware of an EA based on it.

It looks like we'll need a few major EA?

News (schedule based)

Trend (long term)

CounterTrend (shorter term)

Choppy or Whipsaw

Channel

Breakout - I think this can be based on "higher highs vs lower lows" method. If we detect the channel, this can predict the direction of the breakout with at least a slight statistical edge.

Hi gents,

Im not sure if im right, but to complicate system can break neck. I think, here are very good EAs which can working in diferent time with cool results [day, night, news] For night time sould by good 10point3 V12 or goblin with diferent indicator [ASCTrend or WCCI - have good results by manual trading -other i think will by loosing time in], for news we have cooooool EA here, and day time im not shure, but very good idea is NDs 5min ASCTrend system, only program it but not to complicated and i think should by closing position by opening other one, or by out. And than to all of them give time filter and should by enough.

For longer trend are better AbsoluteSrange Indikator or Price Trender [ND use them in his system like base trend Indik. with cool res.]ASCTrend is for shorter term [ i think]

Only my opinion. U are older traders, so u should to know times when whitch EAs have best time to run, a specialy for night time im not sure.

J

Sergey Golubev
Moderator
114232
Sergey Golubev  

Hi janka,

We did not forget about this thread. Simple, i am doing the statements and posting for elite section 3 whole days incl weekend very week. And I am completely out now. Igorad is creating some tool for me to do it just for one day for elite section and after that I will be free for any development.

It was a lot of ideas. But the main subject is to estimate the market condition and to predict the market condition. We came very close to this first step of the development. Very very nearly.

Most easy step is to use EAs according to the market condition. I mean that it is most easy step is to code your idea. But it is most difficult to code some indicator which will tell us in the GMT morning: "today will be flat till 1 pm and after that breakout of the level ....". But we came very close to this development.

Just Igorad finishes some tool for me to generate the statements easy and we are ready to start.

janka
64
janka  
newdigital:
Hi janka,

We did not forget about this thread. Simple, i am doing the statements and posting for elite section 3 whole days incl weekend very week. And I am completely out now. Igorad is creating some tool for me to do it just for one day for elite section and after that I will be free for any development.

It was a lot of ideas. But the main subject is to estimate the market condition and to predict the market condition. We came very close to this first step of the development. Very very nearly.

Most easy step is to use EAs according to the market condition. I mean that it is most easy step is to code your idea. But it is most difficult to code some indicator which will tell us in the GMT morning: "today will be flat till 1 pm and after that breakout of the level ....". But we came very close to this development.

Just Igorad finishes some tool for me to generate the statements easy and we are ready to start.

ND, we have cool administrator!!!

If i can help with somethink, or so, let me know, ill by happy to do it for u and all here!

J

janka
64
janka  

Posible indikators

Hi gents,

Im watching few days at 2 indik, i found on tds threat and i think, CCI indik. i replace with _MTF_WPR_1_fixed [realy diferent and more helpfull] and other ang_DIpm3_v1 seems good for middterm trade too.

Just idea

J

Ian Smith
1595
Ian Smith  

I don't want to "merge 10point3" with a bunch of other indicators. My intent was to have a series of indicators, a series of settings, a series of exit strategies, and a series of lot management techniques.

So we could add Firebird entry signal with 10point3 lot management and phoenix exit strategy, running at the same time with a few other different ways to enter based on the type of market. I wish there was an easier way to configure arrays from the settings window. :/

janka
64
janka  

i think, is nearly done

Hi, gents!

Yesterday i found on Skydart treat new EA. When i open it, i just shockd was! There is everything what i was dreaming about, to be good for our think. So we have News Trader, Day Trader [ only change indikators in - CCI is ok, other one is fast so can be there, only replace MA... with ASCTrend and putt in Absolut Strange in and change setting for time trade to be able open/close hours + minutes -hour only there are]. And now to please someone to do the change, and look for some goot scalper and EA portfolio for trading will by done.

Is here someone so good man with litle of time and will do it for our comunity here? Is someone so nice? Please, please, please!!!

J

ibinsad
11
ibinsad  

any news?

Hello, is there any news about ea switcher im really interested.

How can i help ?

I think it can be the most profitable.

To add comments, please log in or register