Forex-random numbers-- Sooner or later "death" will come. The law of nature is this!!! - page 4
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
It is reasonable to assume that if price has crossed the first grid line, it will reach the second. The 38.2%b 50% etc. lines will then act as support/resistance lines.
I'm sure you know this as well as anyone here.
The problem is precisely in the construction of the grid. So far it's been very tricky with this, even in the manual version...
kch, that's a lot of chatter for nothing. If you are seriously interested, take it up and write it. A brief scheme, as I see it, is something like this:
- take a working TF (say, H4),
- You build a phase on it (a standard one or any other),
- build an array of swings of that zone,
- Out of these swings you select only those that affect the future (described in R. Miner's book "Dynamic Trading"). Miner's book "Dynamic Trading", in the section about time projections),
- build a Fibo grid on each of these swings (usually there are no more than 10 swings, although it is theoretically possible there will be more),
- You assign a weight to each level out of all the received (it is very subjective, but it has to be done, because different levels obviously have different value).
Now repeat the same procedure for several more TFs (say, H1 and D1). As a result, we obtain a dense grid of levels (a complete grid can contain several hundreds of levels) with different weights. And then you cluster all this magnificence and determine the most probable support/resistance levels.
There are many nuances in this scheme, which I am not going to explain. But the system will not work without their reasonable solution. The implementation of the system in the code with the current capabilities of MQL4, which does not allow to create new data types, is a feat (I haven't managed to do it yet). But judging by various scraps of rumour, the feat may well pay off.
And talking about Fibs with a grid built on one or two or three swings in the current TF alone is simply pointless. It's an exceptionally thin and not very stable structure that can get very rearranged with not much change in history.
It is reasonable to assume that if price has crossed the first grid line, it will reach the second. The 38.2%b 50% etc. lines will then act as support/resistance lines.
I am sure you know this as well as anyone else here.
The problem is in the construction of the grid. So far it is very complicated, even in the manual version...
The question was: which line (support or resistance) the price will reach faster.
In other words, the problem is not about constructing lines, but about something else entirely.
kch, this is chatter about nothing.
The chatter about the importance of Fibo levels is also boring...
And putting weights on a few hundred levels from history - in my opinion, it's like fitting something to history, to put it mildly, is not quite correct (past experience).
No, it's not a fit. The significance of the level (in my scheme) is determined by three factors:
- the length of the swing that generates the level,
- the number of the Fibo level itself within the grid attached to the swing,
- remoteness of the swing itself in time from the present moment.
No optimization, God forbid. And it is not realistic with such calculations.
And talking about Fibs with a grid built on one or two or three swings only on the current TF is just pointless. It's an extremely thin and not very stable structure, which can be greatly rearranged with not too much change in history.
In my opinion, three is just enough.
No, it's not a fit. The significance of the level (in my scheme) is determined by three factors:
- the length of the swing that generates the level,
- the number of the Fibo level itself within the grid attached to the swing,
- remoteness of the swing itself in time from the present moment.
No optimization, God forbid. And it's not realistic with these calculations.
I'm sorry if I misspoke...
The problem is that we do not know now about the crowd's memory (i.e. which of the levels is still significant to it and which is not anymore).
We can't see the levels of orders accumulation in MT, we have only guesses about possible places of their accumulation and no more...
Well, if that view is already established, then why ask questions?
that's an answer.
Thank you.
Back to the very first post in this thread. I totally disagree I've been trading forex for 7 years and this is my only and main income, I manage investors accounts.
I use my own trading system, starting point is Neural Network -> further input and output according to technical analysis. I don't use Fundamental analysis at all. I have been developing this system since year 2000 using Easy from TradeStation and now I have started to port it to MT, but it is too slow in calculation of sines and cosines of neuron movement, so I am finally porting it to C++ DLL and building it now.