Artificial Intelligence - page 2
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
Thank you Sir...please test thouroughly before posting your work...please (because I will:)). This is fun isn't it?
but wouldn't you agree that writing in percent rather than dollar amount would be better & self-updating...trade set to trade set...? please include the 1 decimal percent input field, as requested in the opening post.
There is no hurry folks...accuracy and well thought out presentation and testing is far more important than speed.
I hope that somebody can get something from this "plain speak" Journal. Don't worry I will FORCE these guys to speak in "Plain Speak English"
Please folks slow down now...lets give this a day and do not go off in several different directions as I must test one change at a time...
ES
i would say yes to all the above.
here is my little fib tweak.
im going to try to combine the ai ea and the ea lepericus posted.
its just a cut and paste job.
then it can all be backtested.Thank you for posting this elitecamper.
I have reached this far in the thread and this represents one changefrom the opening link posted.
I will test this as I got sidetracked here and there was 2%in the float as I post... I took it!...HeHe I used a "close all" tool that I have and quickly exited and got flat in the entire portfolio. Then I turned off the advisor..in preparation for this new test...now some of you are asking to yourselves..Electric? why 2% ? This is my answer: Because we are working together with SL's and TP's and I am trading with 1 lot which is a total of 4 lots per 3k. 2% seems reachable before the AI system kicks in...which I consider protection instead of being naked. I like getting out early with diversification just to see what the AI does with its directions on the new entries at the next opening of the one hour candlestick. We are combining and calibrating, thus my need for one change at a time folks.
Thank you.
Do not post your PERCENT MM change until tomorrowplease! as I must test and it will get lost in this thread. Lets limit our posts now ok? This can become a very large thread...or maybe we can be quiet and test (please do not rate this thread).
Forget backtesting multiple pair..it does not work together...This is AI with MM. Plus I will Forward test.
ES
P.S. With the four independent pairthat I trade representing eight currencies...there are losses and gains in the mix..so when I exit and re-enter...the AI can go in another direction...sometimes it stays the same....and the individual losses that was averaged previously turned around and becomes the gain in the next exit...each pair fortifies each other and who cares which one is carrying the weight! Remember we are not naked and we have SL's and TP's...also...This technique that I am using is advanced trading and should be realized by some of you as brilliance, and believe it...I found this all by accident and years of trading experience and not MY brilliance. (now I am going to eat my words if all this goes bad).
I have made 6.39% in two dayswith my Demo trading, with the link in the opening post here, and I used the default inputs on the four pair We can now compare this one changeto this to see how it does, but I need to watch and exit at 2%, as I have been doing, to compare apples to apples.
Late edit: the EA portfolio of four pair with this latest change, is already in plus territoryin under 10 minutes...some pairs are losing and some winning. the reentry that the AI evaluates is always new after a 2% profitable reset...and again...if the hold becomes overnight we are not naked.
i would say yes to all the above.
here is my little fib tweak.
im going to try to combine the ai ea and the ea lepericus posted.
its just a cut and paste job.
then it can all be backtested.oh i see good idea thx
to not overexpose on one currency movement. In today's environment the general consensus is that there is an "unwinding" of the carry trade...I am not agreeing or disagreeing with this.
Nobody can predict or forecast with reasonable accuracy INCLUDING AI. but....we need the ebb and the flowfrom the basket with enough force to get out and restart with 2% each time...if not the individual SL's and TP's will protect this. There really is nothing wrong with this approach and I challenge someone to show me otherwise....We have diversification...stop losses...money managment...good entries...defined plan (hint: if anybody wants to flame...try attacking the entry method..I am game...lets rumble, perhaps try it by PM though as we do not want to fill up this thread with the last word
)
...But when you choose pairs that overlap/corrlate in the same or opposite directions you will get double exposure with this EA and if you want to trade with 1 lot per 3k (total 4 lots..this is heavy, but again we have SL's and early exits)..this is the best balance and approach to use.
We need good entries that move separately for the MM to work, as this is basket trading.
ES
whats the benefit of trading in combos?
Does anyone have a link or a good explanation on how this AI actually trades? I mean what is the gist of its logic?
4 accelerators, set apart in intervals of bars, on the same time frame.
then a "weight" (a variable) is assigned to each accelerator. it determines which accelerator of the 4 has more weight, or more importance. also the weights determine the relationships between the accelerators are they tight or loose? . the equation looks like this..
(a1 * x1) + (a2 * x2) + (a3 * x3) + (a4 * x4)
except the AI EA was only optimized one time. and then distributed like that.
that doesnt really qualify as artificial intelligence.
because it doesnt evolve itself. it has all the tools to do it tho.
the way u guys are using the EA ur only utilizing like 10 percent of its capabilities. granted u guys make good use of this 10 percent.
alot of traders on these forums put down back testing and optimization. and cry about curve fitting.
but they are using it on linear systems and thats why they get bad results.
This EA being discussed here is a non-linear system, thats why it's so special.
its the next level of system building. and its application is universal in this case is being used for forex and in my opinion there is alot of potential here.
i think its important to understand what this system here really is and really appreciate its creative design, it is not just another linear tool that traders customly use. and if u cant wrap ur mind around it thats fine but dont bash what you dont understand.
oh this is basicly just a trend ea its all it is.
oh no bashing yet. just that i have recieved alot of bashing in the past.
when i have tried to share ideas that involve non-linear systems.
they tell me that im a noob that they already been there and done that, that they want to save me the time, that there is no holy grail.
im not even looking for holy grail! im looking for a system!
Folks,
Why do you think I am here & started this thread?...I have been trading for many years and I seem to be able to recognize some approaches that seem promising...if not I should just go bury my head in the sand...I am not bragging as I am first to admit that it takes me longer than most to learn new stuff and that I may not be the sharpest nail in the shed...but when I "get it", I really do. I am reaching out.
I haven't seen any bashing yet...is that good or bad?
Heuristic approaches and the training required that the other 90% of the capacity that this poster suggests is possible with this approach is fascinating. I want to learn it..I want it...I need it...I feel it...By the way when I learn this approach...I will charge for it and I will not be talking to the retail crowd then as there is no money there for too much work. The trouble is that most offers are not good and I write my own ticket and those with deep pockets do not want to put out cause' I am not good enough yet.
ES
4 accelerators, set apart in intervals of bars, on the same time frame.
then a "weight" (a variable) is assigned to each accelerator. it determines which accelerator of the 4 has more weight, or more importance. also the weights determine the relationships between the accelerators are they tight or loose? . the equation looks like this..
(a1 * x1) + (a2 * x2) + (a3 * x3) + (a4 * x4)
except the AI EA was only optimized one time. and then distributed like that.
that doesnt really qualify as artificial intelligence.
because it doesnt evolve itself. it has all the tools to do it tho.
the way u guys are using the EA ur only utilizing like 10 percent of its capabilities. granted u guys make good use of this 10 percent.
alot of traders on these forums put down back testing and optimization. and cry about curve fitting.
but they are using it on linear systems and thats why they get bad results.
This EA being discussed here is a non-linear system, thats why it's so special.
its the next level of system building. and its application is universal in this case is being used for forex and in my opinion there is alot of potential here.
i think its important to understand what this system here really is and really appreciate its creative design, it is not just another linear tool that traders customly use. and if u cant wrap ur mind around it thats fine but dont bash what you dont understand.
oh this is basicly just a trend ea its all it is.you can go toCurrency correlation | Mataf.net
and learn alot about pair correlations.
I have looked at 4 pairs representing 8 different currencies before but its hard to match pairs that have both low correlation and similar daily ranges.
For example a set like:
AUD/CAD
GBP/JPY
NZD/USD
EUR/CHF
represents 8 different currencies and have low correlation but the gbp/yen pair is "weighted" more heavily because of its insane volatility. If the yen crosses decide to run away like they often do, then its going to skew your trading basket because the other pairs dont have the movement to offset it.
I guess what I am saying is that ES's approach to this is smart and can be very powerful once you get the right pairs picked out.