Machine learning in trading: theory, models, practice and algo-trading - page 350

 
Yuriy Asaulenko:

Pieces. There is no need to glue.

If you trade on the minutes for 1-2 months of history for training is enough for your eyes.


I have 2 months on the current contract... I'll try it one of these days.
 
Maxim Dmitrievsky:

In general, yes, the current contract has just 2 months... I'll try the other day
Just in case you haven't worked with futures. The futures are really traded and may be used for training, testing, etc. only for the last 3 months of its existence. That is, immediately before and after the expiration of the previous one.
 
Yuriy Asaulenko:

When calculating on the volume of the transaction, it does not matter what the deposit is, or even the specific volume of the transaction itself. The profit on transactions for the year is 8%.

As an example. I have 0.5% of a transaction per business day. The deal, say, 10,000 rubles. 200 days per year *0.5% = 100%/year profit from volume of transaction.

Now recalculate the leverage on the deposit, its load and the number of real contracts - we obtain the expected profit of a particular deposit. On Forts, the leverage is -4-5, the volume is N lots.

Certainly 800% looks good, but we get 8% p.a. from a lot. Imagine that there is no leverage - then there is no point in playing these games. It's strange.

Probably I don't have such a system in Forex, because not to count from a deal - all I get is some kopecks of profit)).


This is a loan. It is given for free in order to trade with 1 000$ even with 1 lot and to pay commissions. Well, it is free, we pay swaps from our $ 1000, but not multiplied by the leverage, here you see the arithmetic. And otherwise it would be strange if the broker would pay swaps for us with his own credit money, right? In the end everyone is happy - the broker and we with our 800% of profit. The broker trades a pool of buy/sell orders at a certain price from such orders of traders owning $1K micro-deposits. Once again, do not forget that besides us, the broker has big clients - legal entities and it is not difficult for him to compile the pool and sell it at the price, which is set by the array of traders - individuals and legal entities. That's how you can earn 800% from 1 000 dollars, that's how I understood it up to now and you want to make us doubt something in the profitability of algotrading))))
 
geratdc:

and you want us to have doubts about the profitability of algotrading)))
Is it so easy to make you doubt? ))
 
Yuriy Asaulenko:
Is it that easy to raise doubts in you? ))

So far I just have not seen reports of 800% from real accounts. I don't have any real trading signals. The question is, what do you think about the real trading robots in your account? Nobody will bother because of 8% and risk a deposit of $ 1k. I do not have to worry about it and to be honest it is also demonstrated remotely - in theory all this may turn out to be a fake or something like MQL, so it's a serious resource that you have to be 50/50. I think it may turn out that you're right)))) If you are right and that many traders really only dream about 8% p.a. after serial losses of their deposits)))
 
geratdc:

So far I just haven't seen any reports about 800% from real accounts. I mean if you don't take signals. I don't know what to do with them, I'll just try to avoid them. Nobody will bother because of 8% and risk a deposit of $ 1k. If you do not know how to tell the truth, you cannot tell the truth about the relation between the real and prospective customers. I think it may turn out that you're right)))) If you are right and that many traders really only dream about 8% p.a. after serial losses of their deposits)))

I am certainly right. The arithmetic cannot be wrong.) However, this is not about someone's fears and doubts, but about the profitability metrics of strategies applied. These are different topics.

Evaluating performance from a deposit is an evaluation of nothing. Let's say two strategies give 200% p.a. at a 20% drawdown, profit charts: twin brothers - which strategy is better?

You can show a dozen real reasons why one of the strategies will turn out to be great, and the other only belongs in the landfill.

I.e., to somehow compare these strategies we need a bunch more specific data, indicators and calculations.

 
Yuriy Asaulenko:

Say, two strategies give 200% p.a. at a 20% drawdown, profit graphs: twin brothers - which strategy is better?You can show a dozen real reasons why one of the strategies will be great, and the other only place in the trash.


What's the difference then? ..........................
 

Interesting article in the sense that I rarely see examples of machine learning for trading.

Here is the primitive thing: divide into two classes. According to the article, the classes are roughly equal. The zest is at the end of the article: you want a high probability of predicting a class, put 90%. That's it.

Machine Learning. Stock Market Data, Part 2: Linear Discriminant Analysis.
Machine Learning. Stock Market Data, Part 2: Linear Discriminant Analysis.
  • Data Scientist PakinJa
  • www.r-bloggers.com
It is important to mention that the present posts series began as a personal way of practicing R programming and machine learning. The bibliography and corresponding authors are cited at all times and this posts series is a way of honoring and giving them the credit they deserve for their work. The exercise was originally published in “An...
 
SanSanych Fomenko:

Interesting article in the sense that I rarely see examples of machine learning for trading.

Here is the primitive thing: divide into two classes. According to the article, the classes are roughly equal. The zest is at the end of the article: you want a high probability of predicting a class, put 90%. That's it.


package 'Smarket' is not available (for R version 3.4.0) :(
 
Maxim Dmitrievsky:

The more neurons and inputs the more stable but less profitable system

Actually, this is not right, imho.

As the system becomes more complex, both profitability and stability should increase at the same time. That is, with increasing complexity of the system its consumer properties should grow.

Let us consider an example of manual system development:

We take a bare trading idea and create a simple TS by optimizing the profit (we may disregard the losses).

2. Introduce restrictions that minimize the number of losing trades. Of course, a part of accidentally profitable trades will leave and in a part of profitable ones the profit will decrease, but drawdowns will also decrease and, as a result, the sum of profit and loss will increase.

Further complication leads only to the increase of profit, at least due to decrease of the number of losing trades.

If the amount of profit-loss does not increase as a result of complication, then there is something wrong. For example, we introduce inefficient conditions.