Machine learning in trading: theory, models, practice and algo-trading - page 164
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
Well, no, but I really wanted to tell you... Seriously, it would immediately remove a lot of questions, and the result of your systems would be better. What can you do, pride is a great thing :-)
So tell us. Let's hear it.
All right, I'll tell you, but it's only a secret.......
Before you automate any process, you need to find out what this field is and how to work with it. You came to the market with a baggage of knowledge in the field of programming and all sorts of other sciences. But as far as I understand, no one is interested in studying the market, and this is your big mistake. So listen here. I will not beat around the bush, describe the dithyrambs, etc. I'll get right to the point.
Have you noticed that it's not uncommon for the situation when yesterday the TS worked, and today it does not want to work, or works with errors. It's all about the context of the trading day, and this context is formed of three values.
The result of trading on the exchange is the volume traded during the day, the Open Interest of the same day and the difference between the opening and closing of this day. And this context, as well as the news background, expectations, etc.. form a kind of predisposition. Okay, I'm getting some kind of water again.
Every day we know the volume of trading of the previous day, the open interest and the difference between open and close. We are interested in only three variables, but we are not interested in the actual values, but in their changes, so we should train our TS according to these changes.
In short. Today: Volume is down, OM is up, Dclose is up. We save the data for training the network on those days where there was such a situation and train the network to work on those days. That is all.
Three variables - 9 possible market states. Train your 9 TS for each state and you will be happy!!!!!!!!!!!
Are there statistics used to determine the probabilities of occurrence of the events highlighted in blue in the diagram?
How were the "expectations" determined?
Are there statistics used to determine the probabilities of occurrence of the events highlighted in blue in the diagram?
How were "expectations" determined?
This is just the context of the trading day, i.e. recommendations. I have an indicator that builds these differences to be able to save the history for the day I want. Otherwise, these are just recommendations. Moreover, there are only 4 situations in the picture, while in fact there are 9 of them. So...
Well, how did you find these "recommendations" - did you collect any statistics?
For example, did you analyze 824 trading sessions in which the price rose, the volume rose, the OM rose, and based on the analysis of the subsequent price dynamics you determined that in 89% of cases the price "continued to rise" by at least x%?
Well, how did you find these "recommendations" - did you collect any statistics?
For example, analyzed 824 trading sessions in which the price rose, the volume rose, the OI rose, and based on analysis of the subsequent price dynamics determined that in 89% of cases, the price "continued to rise" by at least x%?
No, these recommendations are fundamental. You just need to understand the meaning of OM and Volume. Look it up on the Internet and you'll understand everything at once. Read, it's already been explained a long time ago on many sites, why it is so...
So, there are no statistics and the "regularity" is just a blank canvas?
So, there are no statistics and the "regularity" is just a blank check?
Especially for lazybones like you Dmitry
http://tempofox.com/interpretaciya-otkrytogo-interesa-i-pokazatelej-obema/