Machine learning in trading: theory, models, practice and algo-trading - page 924
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
did you find out?
The answer is here:
Going to the whole depo on the real?
If "NO ! "Why?Go philosophize some more. When you're done philosophizing, be sure to share.
The answer is here:
On your picture you can at least make money with ribbets. And it's even worse - as soon as there was some big news and the market patterns changed - the profits went down.
That's why I said about the flip - he would have been in profit from that moment if he had listened.
On your picture you can at least make money with ribbets. And his is even worse - as soon as there was some big news and market patterns changed - the profits went down.
the pattern is always the same.
the beginning of the 95% ATS demo/real:
If this is present, the ATS-ku f*cked.Ahahahahah... fire away.
what other options?
On your picture you can at least make money with ribbets. And his is even worse - as soon as there was some big news and market patterns changed - the profits went down.
That's why I said about the flip - he would have been in profit from that moment if he had listened.
It means that the trend does not stand out and the trading is going noisy and flat, because it goes back and forth for a long time.
If it's NS, it also means that inputs are unfiltered before they are submitted for training. It seems obvious.
It turns out that coconut is so mean to us ) now it's going to cling to the test curve every day
I'm trying to help you. Look at the profit from trades - if the strategy is random, then each trade will bring a loss equal to the commission + spread.
If the average loss is larger than the commission + spread, then the account is trading in the opposite direction than it should. Look for errors, something is wrong. Try to increase the training period, take a couple of months.
Especially look at your trades on the pound - a bad random strategy cannot trade strictly in the red all the time, it is statistically impossible. The strategy is good, but you should have reversed the signals. Maybe you mixed up the classes 0 and 1 in the training data?I am trying to help. If I look at the profit of transactions - if the strategy is random, then each transaction will bring a loss equal to commission + spread.
If the average loss is larger than the commission+spread, then the cp is trading strictly in the opposite direction than it should. Look for errors, something is wrong. Try to increase the training period, take a couple of months.
All right, I know. For the euro, too, since the beginning of this week another model was trained.
I go step by step, now the TP is improving or descending, but the general trend is improving.
I don't think it makes sense to go into details, because there are better updates, like the screenshot above. Next week I'll add a new one.
How not to retrain the model, who has any suggestions?
I got it like this:
f1_score training: 0.96
f1_score test: 0.59
Did you retrain it?