I do not use one, no. I don’t have the capital required to invest in one. The hedge funds I know who are using one spend several million dollars on it and I am not at that level of trading yet.
True but hedge funds dont use 100% of their potential.They put “very” small positions compared to their total balance because they are afraid to lose their customers money.But you as individual trader only trade for yourself, so your can risk as much as you are comfortable with.
And sending your money to other people to invest in is not really learning forex trading, because one they are gone, your income is gone too.Every people must learn to make money one their own, maybe put only a % of your total assets in HF but not all.A little bit of portfolio diversification.
Maybe you dont have the technology and computer power to run a very sophisticated one that may be the case, but still a many people with average PC’s use them so maybe even with a little less performance you can still get better results in trading, have you ever thought about that?
The Forex Market is news driven and news and market psychology are inherently unpredictable phenomenon.
Nope.The news make up only ~1 % of the market structure.The market is interest rate driven and other economical factors.
But we dont look at fundamentals they are really unpredictable but you can react to them quicker since not always they the market adapts right after release to it.
Technical tools can “predict” the market better i believe, but you dont have external information here.
I don’t think Technical tools are reliable at predicting the market. If Technical analysis could accurately predict the market everyone would be making a fortune trading Forex.
The big moves I have found are after major Economic News or seem totally random. There are periods where a currency moves in a narrow price range and all of a sudden it ticks 15-20 pips.
I’ve been able to make a moderate profit margin by gauging market psychology based on Economic news and price action. I don’t use any Technical instrument other than EMAs and candle sticks.
I totally agree. This is why the market moves in a completely digital way. News comes out and there is a perfect move to find equilibribum between buyers and sellers and then the market doesnt move at all until there is another input of news.
You can write an NN in excel in about 5 minutes. I can do one for you for a deal price of $500,000 if you like?
Ok there are more ways to make a fortune in this game, but none of them are that what majority thinks.The forex market is not a democracy, here the minority is always right.That is why the majority loses money,because they dont have a damn clue about how to trade profitably.
Totally agree ? Why ? Digital way, what is that ?
Show me your research data an methodology about candlestick distributions based on a random sequence and i believe you.Just by saying something that sounds cool doesnt mean that you are right.
The random walk argument that price is totally upredictable is misinterpreted, which means the chance of a white candle being followed by a black one is partially random only, but it can be approximated if not predicted to some degree.For example if you look it that way by black-red candle being coded as 0 or 1 then the sequence would look like this:
0,1,1,1,0,0,1,0,1,0,0,1,0,…
which may look random to you, but take a look at closing or opening prices
1.2345,1.2346,1.2347,1.2340,1.2349,1.2348,1.2346,1.2345
Does that look random to you ? I dont think so.There is alteast 1 mathematical function in the universe which can describe that sequence.The fact that you dont know it, is your problem, but that has nothing to do about the randomness of the data sequence.
So its predictable in a certain degree, not fully,and also the news period is really unpredictable, it is a noise in the point of view of the “market universal function” but if you know or atleast have an aproximation about what the news could bring then its no longer a news.But you mix the fundamental factors with the technical ones.
I was only talking about technical factors
Ha ha, you are very sensitive. I was joking btw, i was trying to point out news is only half the story.
To put it into plain english, classical micro economics goes on about about the price finding a level that suits supply and demand. If the market moved only on news the price would just move in perfect steps. It doesnt but why?
I dont know…hmm… maybe because news doesnt matter? :54:
It is only a junk data in the perspective of the price flow.The market in most of its time is in equilibrium , and the news are a deviation from that equilibrium, possibly a human intervention because it is know ahead that it will happen but the outcome is a mystery so it cant be random if we know that at 15:50 there will be a GDP release, but we dont know the outcome so its hard to explain, but definitely not random, only the outcome of it.
But the equilibrium movement of the market is not affected mostly by fundamental data or it is soo hidden that its impossible to decypher it.The most prominent theory is that of interest rate fluctuations and SUPPLY DEMAND levels, but that still doesnt explain much.
But definitely not human psychology, not the forex, the stock market may be because in stocks the down movement is negative so its obvious that that is the sign of fear, but in forex the down movement doesnt show anything it is negative for 1 currency and positive for the other , so in the end its neutral.
So stocks are very different from forex market.
And why the market moves the way it does, tell you the truth, nobody knows, if they’d do they would be the new Goldman Sachs.Probably a good combination of many factors,which can be predicted but not with the current computing power.
Not even the Tianhe 2 the world’s fastest supercomputer can crack the market’s movement i can tell you 
I think that i know, but it part of my secret sauce.
in regards to super computers
I got told a very interesting thing a long time ago about from my Aikido sensei. He said “If you take the time to learn how the body works, then breaking it is very easy”
wax on, wax off, Michael san.
[QUOTE=“pipwhip;578228”] I think that i know, but it part of my secret sauce. in regards to super computers I got told a very interesting thing a long time ago about from my Aikido sensei. He said “If you take the time to learn how the body works, then breaking it is very easy” wax on, wax off, Michael san.[/QUOTE]
Neural networks or not, the market will not tolerate an anomaly against it.
It will quickly adopt that anomaly as its own.
Its how the market functions.
So u cannot have a situation where one force takes over the other completely, each and every trade.
Like water in a bowl, equilibrium will be sought out immediately.,
A neural network or some supercomputer having the market for breakfast lunch dinner and supper will produce a situation like in the third series of Matrix, where Mr Smith goes loco.
Why? Because for every buyer, there must be a seller, vice versa.
Thats a bit exageration.I believe the part with market equilibrium but i dont think a NN would disrupt the market that much.I mean from the start of the millenia there was a big increase of automated trading, it is shown that around 40% of total market executions are done by bots not humans, and it didnt disrupted the market that much.If you’d trade with a NN all day that anomaly would be reflected only in your trades, in nothing else.And if you put a trade here and there, guess what, many people do it too, so its not like you would put trades only at spots where nobody trades because there is no such spot.On every tick there are trades put on so you technically wouldn’t even scratch the market.
You seem to have all the answers.
One has to wonder why you asked a question.
Yes i had my own theory before asking the question, but i was interested in your opinion guys
I enjoyed your post a lot. It got me thinking about something from a while back in regards to how the exploitation of inefficiencies leads to the inefficiencies not existing any more.
What it got me thinking about is, are there a stable source of inefficiencies? This is a philosophical question I suppose but I was wondering your take on it. I personally would say structural inefficiencies like the January effect are not sustainable. There are some technological ones that HFT has truly killed from over exploitation, I would say these too are not sustainable. There are however people who have consistently made money in the same asset classes and must be deriving their edge from a sustainable source. I can think the only constant is the human condition and what people call predictable irrationality but was wondering what your take was on this question?
Did you really? I mean really, deep down, coz it really sounds like you don’t.
I find this Proximus guy really funny. He seems to know how everything works, refuses to accept any opinions that do not necessarily align with his beliefs and yet, still asks people for their opinion (only to attempt to shut it down of course).
To answer the original question, and this is purely based from my experiences and studies:
I think the original concept of neural networks sound straightforward, but the devil is in the details. Neural networks aren’t hard to build; they’re structure is not difficult to design, one could build one pretty easily after a couple weeks of study (with prior programming experience).
Like pipwhip said, one could make one through Excel or any programming language. The main problem when it comes to the neural network is the training process. Data, data quality, training algorithm are some of the main factors involved.
What we’re seeing a lot in financial neural networks, is the use of the backpropagation algorithm for training neural networks (closing prices as data sample). People kind of just take assume that because everyone is using it, they should too, without fully understanding the ins and outs.
The guys who know what they’re doing will analyze the problem they’re trying to solve with the NN and from there, pick the most appropriate training algorithm to use, under the given situation, to determine the best weightings for the NN.
There’s more problems that arise after the successful completion of a NN with training as well, but I’d rather not get into too much detail as it can be quite lengthy.
Personally, I’ve played around with several NNs (high frequency and low) and ensemble methods and my conclusion is that they’re just a bunch of hard work for very little end gain.
The funniest thing about this thread is thinking that a 51% win rate with a 1:1 risk reward guarantees a winning system.
But that seems to be overlooked because someone thinks the math works out.