HOW TO TRADE FOR NOTHING ? - page 8
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
Prival
Here's what I did for interest( this is whatPrival probably wanted to see, not a forecast), as I said above...
I do not have a signal without noise for our example, I made my own (by informativity it is very good for TS) - it is also a teacher for the grid (3-layer grid)
green line - signal with noise
red - teacher (what I want to separate from the signal with noise)
yellow - result (grid output)
I attach the scatter diagram -.... it is of course possible to make a better cut - but in real conditions it would be of little use (informative areas (kinks) will lag or appear earlier) and you would need to train on every bar... that is why it is better to simply filter in my opinion....
https://www.mql4.com/go?http://s16.radikal.ru/i190/1007/a5/00a689e43a38.jpg
https://www.mql4.com/go?http://i069.radikal.ru/1007/96/773a22155904.jpg
Sores you C-4(https://ru.wikipedia.org/wiki/C-4_(explosive_matter)) yourself remember when you started.
we are not afraid of plastid )) h ttps://ru.wikipedia.org/wiki/Авиационная_вакуумная_бомба_повышенной_мощности
I'm being kind, almost homely.
On the subject, it's all well and good but totally useless. But who told you that in real markets there is only one signal? Change the problem, enter a couple hundred signals at once and display their cumulative result on the chart. Then, so that the model is not too primitive, dilute it with chaos of random numbers. But this picture is too primitive. Make sure that these signals are not equivalent and also that their strength is constantly changing. Then change the formulas of the signals themselves, so that the same signal would generate a different pattern at different moments of time (however it should produce some fuzzy regularity). Then feed it all to the neural network, and wait for a few months for it to respond. If it does not go crazy, then that is a reason to wonder whether you are not crazy yourself, because while traders make money on what they have, you are solving the problem of the square of the circle, which has nothing to do with the market.
much worse with noise - and the scatter plot confirms it, and so does the forecast itself (55 points ahead)... if we smooth it out and then feed it into the network, of course it would be better....
but here again we have cyclicality in the data ... if it is removed, the grid and any predictive model can not properly forecast.... (about forecasts for the future)
your approach is different - removing noise ... it's great to remove noise - and what goes up and down is not important - we need the line itself, or rather the peaks and troughs (the most informative part for decision making) ...
Maybe we should just try it on real quotes ? and then compare it with MA for instance... which one is better.... try it for your interest....
....
Tried it. Made it, looked at it. Put it on the shelf and it is gathering dust.
You may also remove the permanent component, no problem.
Vizard is so, so, dabbled whether the NS stands up to the man. I wanted to make it clean on this example. Just here a bunch of "true" believers that NS can solve the problem better than a man able to think...
Matkad files attached, the code is 5-6 lines, you just need to know what is the spectrum, PF, who wants knowledge and wants to understand, will understand...
I've got more tasty stuff for neural networkers. Input. Picture below, each line is a currency, by definition lies only in the interval from -1 to +1, no normalization is needed. The second one is rate analysis (probability of going up or down, also lies in the interval + - 1) I wrote to Neutron but he's not saying anything, is busy probably, I would like to study this tool...
Excuse me...but where did you get all this crap? ))))))) Is he (the trader) a visionary? to know where most people think the low is? etc......
everyone here on this forum and on other forums put together has no influence on the market... even if they stand on the same side ..... (this is about playing against the crowd )))))))))
)))) hilarious..... to the point !
that's how we trade ))) everyone is the smartest ... but everyone is in the red )))))))))
...
On the subject, it's all well and good but totally useless. But who told you that in real markets there is only one signal? ...
If the market were like the one on the first page, it would have been destroyed a long time ago and Forex would not exist anymore, it is much more complicated there...I don't think it's useless at all, before you can predict the market, this incredibly complicated substance, you have to learn to predict something simple, even a sine wave with noise, many people don't even know how to do that, but they want to predict the market... from simple to complex, that's the way it's always been. If it's different, you'll get a branch like Fourier doesn't work(https://www.mql5.com/ru/forum/125844) or help with Fourier https://www.mql4.com/ru/search/?keyword=%D0%A4%D1%83%D1%80%D1%8C%D0%B5.
Why Ritchie is right. There is a rule of thumb not to trade on extremes...
the equation was wrong - you see - you understood the crowd as an extremum ... and how Ricci understood it - I asked him ... he probably thought I was joking and sent me away to watch phonetics ... and I was asking seriously ... no jokes ...
But then again about extrema... you have to mark it first - understand that it is one... it's easy to do on the story... but in real time it's another matter... how do you mark them for example?
Vizard was just trying to see if the NS could stand up to the man. I wanted to make it clean with this example. Just a bunch of "true" believers here that the NS can solve a problem better than a man who can think...
i don't like grids, maybe because i got nowhere good with them...+train time is killing me )))
the bottom one from the screenshot - not so impressed with it.... but this may be a small piece of history... Just a quick look - I think I'll pick it up with some simple tools (ma, etc.)... and normalization is not really needed if the signal is taken through 0...
or not through 0 in this example ?...but for colour change...if someone has it, drop it off in Tsv files - i may play with it later - i'll pick it up....
I'm being kind, almost homely.
On the subject, it's all well and good but totally useless. But who told you that in real markets there is only one signal? Change the problem, enter a couple of hundred signals at once and display their combined result on the chart. Then, so that the model is not too primitive, dilute it with chaos of random numbers. But this picture is too primitive. Make these signals not equal and also constantly changing their strength. Then change the formulas of the signals themselves, so that the same signal would generate a different pattern at different moments of time (nevertheless it should give out some fuzzy regularity). Then feed it all to the neural network, and wait for a few months for it to respond. If it doesn't go crazy, then it's a reason to think whether you are crazy yourself, because while traders make money on what they have, you are solving the problem of the square of the circle, which has nothing to do with the market.