![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
1. limit losses
2. limit losses
3. limit losses.
It's not trendy, but you don't have to pray while drinking beer, which by the way is not healthy.
The only 100% way to fully eliminate losses - do not gamble at all. It's not fashionable, but it's guaranteed.
I recommend it if you like :) True, no profit, but the plum=0, for sure.
:)
Read it carefully! It says limit losses if you don't play at all, there will be nothing to limit. Apparently Neutron was right... Ostap was out of his depth...
You're right.
Feel better?
:)
paralocus, no offence, please. Here, play with this. I dug up some of my stuff from before last year.
Here's a little something for you to play with. Made in the form of a turkey. It uses, as a pseudo-science statistic.
is a simple correlation. More exactly the product of two values (1) correlation of the current pattern to the historical one
(2) future (relative to historical) growth of MA in some period. As an input receptor
here is a stochastic.
Parameters:
PreCorPeriod: indicator pattern length (in this version Stochastic)
PostCorPeriod: forecast period length (the period of time during which the MA will either increase or decrease)
CorWindow: length of time sample on which statistics is gathered.
MA_Method, MA_Period, IndPeriod - do not need any comments.
Ind_dt: frequency of reading of a value from the indicator. If it is 1 - on every bar. If it is 2 - every second, etc.
Don't criticize the style - I was written in a hurry "not for showing off, but for science". :)
I dug out the year before last in my stash. I'm throwing in one sample for you to try out.
What's up, I took a close look at the clock... It's quite an outperforming turkey. I almost regretted
I almost regretted putting it in the forum. :) Come on, I'll draw some more. :))) Tomorrow night, though, I'll make a grail
on a couple dozen of these turkeys, if I'm inspired.
Почему бы для этого не использовать:
- асс. правила или
- Кохоннена?
Они и дают ту самую вероятность и подержку.
Why?!
After all, this work in the case of binary input data may as well be solved by statanalysis pattern. Judge for yourself - no troubles with training and searching for the optimal architecture. As paralocus rightly pointed out, "...the Expert Advisor would be childishly simple but maturely effective"!
asst.rules is the statistic you're talking about
paralocus, no offence, please.
I'd rather be offended than shot. I'll look at your turkey when I have time.
Using the base:
1. Taking a pattern from the market.
2. search the base for the most similar pattern(s) // synonym in Neutron terms - the closest node(s).
3. we play either (1) by the most similar-nearest (2) by a Fuzzy-ified set of the most similar ones
(var: (2.1) by linear combination (2.2) by linear combination S(x), where S() is sigmoid, etc.)
Please, explain to me if I'm not an expert, why do I need a base when I can gather statistics directly for the current market pattern? One of us is obviously stupid. I admit that I am. What's the catch?
We don't look for "similar", we identify the pattern unambiguously!
When we find one, we look at the stats for that particular pattern, and play there. MetaDriver , there is no fuzzy logic here, everything is definite with determining the node and the database is needed for statistically accurate determination of the probable direction of position opening (this is from another point of view).
Portfolios are played (as a rule in a supposed trend) to reduce the risk of trend reversal. However, the risks are only half as much if we play in two opposite trends. I think it's obvious. The probability of two reversals is clearly half the probability of one. And regardless of the absolute value of the probabilities.
No? ;)
Well, you're making a lot of sense. That's right:
Portfolios are played to reduce risk. (That's it. Nothing more to add.)
Efficiency increases if you use instruments with a negative correlation coefficient in the portfolio.
Ass rules are the statistics you're talking about
Ok!
The pincypsee is quite a leading indicator.
You, that's it... stop with the playboy language. Not kids. You can make a special color for fun and waving there before the fools by "anticipating" turkeys and graal systems, fortunately there are enough fools on this forum - there will be no putback and your popularity will go up considerably (not for long, though)!
And about the stats I was wondering...
From those half year ticks you gave we get a little over 400 PT counts when split with H=9 pips. Not enough for statistics, for a split with H=10 it's already 300 counts. By the way the results of dissection too hint at using of smaller H as possible because at H over 12 dispersion of projections of PT arms lengths on Y axis practically nullifies statistical advantage by sizes of risks. What to do here? Reducing H is the only thing that comes to mind, but then the size of the payoff turns out to be in pips
It is for this reason that I now use the opening prices of the minute bars for my research. Not great, of course, but enough for understanding. And saving resources at the stage of studying the subject plays not the least role.
Fyodor, look how intensively the girls dance:
In the picture, the red color represents frequency of meeting pattern for the 2-entry classifier on SP with H=10 points. Blue is the real BP (EURUSD minutes). The length of the training sample is chosen rather large for the reason of decrease of statistical variance of the pattern classification. We can see that for SV the frequency distribution is almost the same, to within a statistical dispersion (whiskers at each experimental point). For the real series, however, there is a statistically significant (see size of blue whiskers) deviation of the distribution from the equilibrium one. You can see that for some reason, patterns like 00 and 11 are noticeably smaller than 01 and 10.
This is an interesting fact.
Ha! - This means that the Market does not like to move in one direction. After all, pattern 00 corresponds to two downward moves and 11 corresponds to two upward moves. I can already see this whole scheme: At the first node (pattern 00) we need Buy, at the last node (pattern 11) we need Sell, at the second node (01) we need Sell, at the third node (10) we need Buy.