Machine learning in trading: theory, models, practice and algo-trading - page 1451
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
Just so you know, I haven't stopped trading for a single day. I don't tolerate downtime, I just don't communicate here. I just haven't communicated here. Maximka scares away all the people with his English. It's not right, so I decided to join the team back...
I see, so I was wrong.
If you're referring to English, I don't know what you nerds are doing here, but scientists are still studying the Brownian Fractional Motion to model volatility. There are no more accurate methods of describing market movements in the world yet. That is, starting with Black and Scholes and on to newer research.
https://tpq.io/p/rough_volatility_with_python.html
https://www.quantstart.com/articles/derivatives-pricing-ii-volatility-is-rough#ref-gatheral
So far all I see from you is a discussion of predictions of candle colors, zigzags and other kindergarten nonsense.The color of the bar in the articles of Vladimir Perervenko is very well defined with an accuracy of about 0.8. And it is easy to repeat. I did it again including bare quotes, the results were 2-3% worse than with digital filters.
It is strange that your results are so bad. But much depends on the symbol and the TF.
I do not guess bars but I try to keep the result not lower than 80 on validation.
I want to correct. I never predicted the color of the bar. Only classification with target ZZ and different variants of predictors. And the best results are obtained using large ensembles with different aggregation methods.
Another important experience is that the more complex/refined the preprocessing, the simpler the model solves the problem.
Good luck
Didn't anything work out for you with ZZ?
The faces seem to be the same, but the questions and answers... the same... It's like Groundhog Day or something)))
ZZ is 80% again...
An unsophisticated person would suspect some kind of conspiracy or problems with his or her tower.
Looks like this is the 3rd or even 4th "wave" about the same thing, I can not say exactly, as more than 2 / 3 posts have not read.
WHAT'S GOING ON GUYS?
The faces seem to be the same, but the questions and answers... the same... It's like Groundhog Day or something)))
ZZ is 80% again...
An unsophisticated person would suspect some kind of conspiracy, or a problem with their own towers.
Looks like this is the 3rd or even 4th "wave" about the same thing, I can not say exactly because more than 2 / 3 posts have not read.
WHAT'S GOING ON GUYS?
The fact that ZZ is useless for MO is a myth.
Well, the outcome of the 1 hour TF just interested me as an application of my predictors to other targets/strategies.
By the way, first selection of predictors revealed Accuracy 0.53 and mathematical expectation of 6 points, which shows the meaning of further tuning in the direction of predictor selection.
The fact that ZZ is useless for MO is a myth.
As for the "owner is the boss"))) From what I have seen said on this subject more than that, there is probably a question of "karma", everyone has their own way.
Well, the outcome of the 1 hour TF just interested me, as an application of my predictors to other targets/strategies.
By the way, the first selection of predictors has revealed Accuracy 0.53 and mathematical expectation of 6 points, which says about the meaning of further tuning in the direction of predictor selection.
Now think about it, you have Accuracy 0.53 and mathematical expectation of 6 points, it's a year Sharpe co. ideally 1-1.5 and ZZ Accuracy 0.8-0.9, but ASR is not better, when HFT companies have Accuracy 0.6 and ASR>10 when processing terabytes of data per day, think about it...
But, as the saying goes, "As you wish")))) I've seen more than what is said on this subject, there is probably a question of "karma", everyone has his own way.
Well now think about it, you have Accuracy 0.53 and expectation of matrix 6 points, this annual Sharpe co. ideally 1-1.5, and ZZ Accuracy 0.8-0.9, but ASR is not better, when the HFT companies, processing terabytes of data per day Accuracy 0.6 and ASR>10, think about it ...
Enlighten me, what is ASR in this context?
I have a strategy on ZZ, which Accuracy is not very important - the most important is Precision, because the decision is made only on the target, while the bar strategy involves a decision in the direction of entry, respectively, the strategy on ZZ Accuracy is around 0.65.
What's the point of me thinking about myths about HFT?Enlighten me, what is ASR in this context?
I have a strategy on ZZ, which Accuracy is not very important - the most important is Precision, because the decision is made only on the target, while a bar strategy involves a decision on the direction of entry, respectively, the strategy on ZZ Accuracy is around 0.65.
What's the point of me thinking about myths about HFT?ASR - annualized sharpe ratio
I agree that Accuracy is not a very good metric in this case. But you've already quotes 0.53 and I've used it as a guideline to give figures accordingly.
Myths are not important, I've cited real averages, HFT because they have the biggest ASR, the intraternity (5-10 deals a day) prediction quality figures are lower by 3-5 times (forecast hours ahead).
0.65 - it's fantastic with such data, it's almost smooth exponent of equity, you simply predict a linear mix of future and past data and everything above 50% is a cheat, it has already been said many times.
The ZZ target makes sense only for"grail sellers", probably that's why there is such a fierce opposition to this audience, otherwise an honest approach would result in the fact that the MO does not make sense in the market, while the indicators are out of the question, it's just random. Only if you devour a lot of paid and free data, a team of highly skilled professionals can thinly generate a slight advantage over the market, and "directly" on the data from DC or Dukas, one gentleman of luck, on the left libraries, to find 65% of this glitches.
ASR - annualized sharpe ratio
I agree that Accuracy is not a very good metric, but you're the one who mentioned 0.53, that's what I'm basing my calculation on.
Myths are not important, I've cited real averages, HFT because they have the highest ASR, the intraternity (5-10 transactions a day) prediction quality figures are lower by 3-5 times (forecast hours ahead).
0.65 - it's fantastic with such data, it's almost smooth exponent of equity, you just predict a linear mix of future and past data and everything above 50% is a cheat, it's already been said many times.
The ZZ-target makes sense only for the sellers of "grails", that's probably why there is such a fierce opposition to this audience, otherwise a fair approach would mean that the MO is meaningless in the market, while indicators are out of the question, it's just random. Only if you devour a lot of paid and free data, a team of highly skilled professionals can thinly generate a slight advantage over the market, and "directly" on the data from DC or Dukas, one gentleman of luck, on the left libraries, to find 65% of this glitches.
Have you read too much fiction or something?
ASR - annualized sharpe ratio
I agree that Accuracy is not a very good metric, but you're the one who mentioned 0.53, that's what I'm basing my calculation on.
Myths are not important, I've cited real averages, HFT because they have the highest ASR, the intraternity (5-10 trades a day) prediction quality figures are lower by 3-5 times (forecast hours ahead).
0.65 - this is fantastic on such data, it is almost a smooth exponent of equity, you simply predict a linear mix of future and past data and everything above 50% is a cheat, it has already been said many times.
Let's distinguish two different strategies - one I have by bar and the other by signal ZZ and trawl, in general ZZ.
The first strategy is bar strategy, I have obtainedAccuracy 0,53 so far, but it is the normal metric for this strategy because the entry is generated on every bar and the average value of bars in the entire sample is strangely enough 0,5. It turns out that I have a 3% deviation from the average, which is not much, but it already gives some information for thought. At the same time I used the same predictors as for the second strategy.
The second strategy with the result 0.65 is not a fantasy, the decision there is made only when one appears, but it is possible to detect one in 23% of cases, and correctly detect in 65% of these 23%. It should be taken into account that the risks are approximately 1 to 1.3 at the beginning of the position opening but a portion of them is taken out by trawling - in general the balance will be wavy enough (here it makes no difference whether it is equity or balance due to imputed stops).
If you're referring to English, I don't know what you nerds are doing here, but scientists are still studying fractional Brownian motion to model volatility. There are no more accurate methods of describing market movements in the world yet. That is, starting with Black and Scholes and on to newer research.
https://tpq.io/p/rough_volatility_with_python.html
https://www.quantstart.com/articles/derivatives-pricing-ii-volatility-is-rough#ref-gatheral
For now, all I see is a discussion of predicting candlestick colors, zigzags and other kindergarten nonsense.I think this is a reprint of some 6-7 years ago I read that, but it is about valutility trading, I wondered a couple of times how to imitate options trading through simple orders - I couldn't find anything