Machine learning in trading: theory, models, practice and algo-trading - page 243
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
look for properties of the market that do not change
If they were, they would have been found by now.
the only thing that doesn't change is that 99.9% lose money
If they were, they would have been found by now.
The only thing that doesn't change is that 99.9% of them leak the money.
So you haven't tried it yet. Maybe the 0.1% have already been found.
An example? Please. Try to enter in the opposite direction after 2 candles in a row in the same direction, after 3, 4, 5, .... 20и. What is the difference? Your conclusions. Limit the SL and TP depending on 80% of the size of the candles in a certain period. Just try it.
Try also to calculate what the candlestick return statistics are. Surprised, but at all currency pairs, this value is constant and does not change, even on the metals, it is the same (in stocks and other exchange-traded instruments has not checked).
Well, clearly you are doing something wrong. Try to approach the problem from the other side, try to think differently than 99.9%....
Your own fault - you anthropomorphized him with your communication with him.
But it's not a problem of MO's stupidity, it's our problem, because you need to pre-process the data before you put it into MO, you need to be able to make sure that MO can really understand what you want it to understand
And you are always singing about picking out noise from non noise, but don't you understand the fact that you are working with noise from the beginning and you want to pick out the non noise in the noise ))))
?
Logically it doesn't add up.
If everything is noise, you throw it all away and don't put it anywhere.
If you find something in the noise, you put it in the model. Let's look further, since we won't necessarily get a low percentage of errors.
You're writing about the same thing, and for some reason the conclusions don't follow from the premise.
Don't forget to add what Wizard said then throw it all away and forget it like a bad dream))) On the maps, I would first do a few
"snapshots" with the window at different sites. With training and application on new data. With and without a teacher. In the case without the teacher, then I'd stick
it on the input as well. I would see where things pile up (clusters) and hit density matrices. Only then would either throw it away or check it.
And only buy/sell, wouldn't complicate things. Before all this I would try to "tune up" (initialization, radius, etc) with simple examples.
Well, the quintessence.
How is all this different from "three soldiers", or "head and shoulders"? No way! Let's take it and then see how it piles up. What's the point? Where is the criterion for judging this "piling up"? There are a million of such Expert Advisors on this site: they draw indices, look through the combinations of candlesticks and start coding. They have received the grail in the tester, and then they are glad that they have lost it with the help of the grail. Of course, it should be corrected a bit more, and then it is definitely a "grail". And you can start to drain again. And so for years. And the pile of this grail manure is getting bigger and bigger....
It's all bullshit - the MO. It's like trying to know the weather for tomorrow from historical weather data. It's like applying facial recognition to a person who constantly has cosmetic surgeries and uses new makeup every time. The person is the same, but you can't recognize him or her.
In general, why try to catch and constantly adjust to changes in the market? Wouldn't it be better to look for something that doesn't change?
It's not bullshit, it's the only thing we have. TA is also MO, just a very inefficient, profane form of it.
PS: you can predict the weather accurately enough (70-80%) for the next few days, the face of a person in makeup is also possible to learn as well as a person does.
Not bullshit, it's the only thing we have. TA is also MO, just a very inefficient, profaned form of it.
PS: the weather can be predicted quite accurately (70-80%) for the next few days, the face of a person in makeup can also be recognized no worse than a person does.
heh heh. how misguided you are.
https://nplus1.ru/news/2016/11/03/glasses
Heh heh. how misguided you are.
https://nplus1.ru/news/2016/11/03/glasses
Humans can be fooled by glasses, too.
You just need to put glasses on a neural network, and it will see a person with glasses better.
Seriously, we need to increase the number of parameters considered by the network, i.e. to determine a person not only by his face, but also by his gait, clothes, skin color, voice, accent, smell, etc...
Not bullshit, it's the only thing we have. TA is also MO, just a very inefficient, profaned form of it.
PS: the weather can be predicted quite accurately (70-80%) for the next few days, the face of a person in makeup can also be recognized no worse than a person does.