Discussion of article "Money management in trading"

 

New article Money management in trading has been published:

We will look at several new ways of building money management systems and define their main features. Today, there are quite a few money management strategies to fit every taste. We will try to consider several ways to manage money based on different mathematical growth models.

Any method of money management can accelerate the growth of the trading balance, or, in other words, to increase the profitability of the trading strategy. Thus, the trading strategy is the basis, while money management is an addition. Let's see what requirements a trading strategy must meet in order to apply money management to it.

First, it is the mandatory use of stop losses and take profits. With their help, you can control trading risks: stop loss allows you to limit possible losses, and take profit makes it possible to assess the potential profit for each deal.


Second, it is a positive mathematical expectation. It enables traders to evaluate the expected profitability of a trading strategy in the long term, which allows them to make rational decisions and manage their capital more efficiently. However, mathematical expectation is important not only for the trading strategy as a whole, but also for a newly opened position.

Author: Aleksej Poljakov

 
Interesting author, scrolled diagonally saw the usefulness, a sign of good material, and the topic is useful. It can be expanded. The only thing began to read, the first formula immediately did not understand, because the definition of the theory of probabilities of winning is m / n (the number of outcomes belonging to this event / the number of all outcomes), I did not understand where you took 0.5 in the numerator and 1 in the denominator.
 
Evgeniy Ilin #:
Interesting author, scrolled diagonally saw the usefulness, a sign of good material, and the topic is useful. It can be expanded. The only thing I started to read, the first formula did not understand at once, because according to the definition of the theory of probabilities the probability of winning is m/n (the number of outcomes belonging to this event / the number of all outcomes), I did not understand where you took 0.5 in the numerator and 1 in the denominator.

We are estimating the probability of winning for a future trade.

if m / n is the probability (or frequency) of the already passed realisation, then for the future deal we get the following: n will increase by 1 in any case, and m - may increase, and may remain the same. Then we have two options: m / (n+1) or (m+1) / (n+1) We take the average of both options and get (m+0.5) / (n+1).

 
Aleksej Poljakov #:

we estimate the probability of winning a future deal

If m / n is the probability (or frequency) of the already passed realisation, then for the future deal we get the following: n will increase by 1 in any case, and m - may increase, and may remain the same. Then we have two options: m / (n+1) or (m+1) / (n+1) We take the average of both options and get (m+0.5) / (n+1).

Now I understand, but you can't do it this way, you understand ) it's your fantasy. For the future transaction the probability is the same as for the past realisation. This estimation of yours is absolute selfishness, no offence. Look in my last article, there is this topic. That's right... I I do not at all take pleasure in such posts, but I had to. I could have given you a more precise formula and you would have seen that it is identical to m / n, adjusted for "future". You would fool the reader, but not me )

 

We take a coin, toss it 5 times and get tails 3 times, whence p=3/5=0.6. Then we toss the same coin again 5 times and get tails 2 times, whence p=2/5=0.4.

A paradox? The probability is different for the same coin?

No, just another confusion from mixing the concepts of probability and frequency. Frequency can only be an estimate of probability, but only if certain conditions are fulfilled - those involved in the formulation of the law of large numbers, for example.

 

Good, I have some doubts about the formulas as I apply them in excel but I do not get results, could not you make a more detailed example with 5 operations and giving the data of these operations and the calculation of the variables for these 5 operations.

I don't know if I make myself clear, what I mean is if you can apply the formula without unknowns and with the prices of tp,sl, pip price, etc.

Thanks.

Thanks.

 
isanchez96 #:

Good, I have some doubts about the formulas since I apply them in excel but I don't get results, couldn't you make a more detailed example with 5 operations and giving the data of these operations and the calculation of the variables for these 5 operations.

I don't know if I make myself clear, what I mean is if you can apply the formula without unknowns and with the prices of tp,sl, pip price, etc.

Thank you.

Thanks.

Let me rewrite the script for you; it now creates a file where it writes the modelling result of each trade.

It is possible that you are not taking into account the impact of the pip value on the deposit currency. The result of each trade should be +/- lot*(TP / SL)*pip value.

Files:
 
Aleksej Poljakov #:

Let me rewrite the script for you; now create a file in which you write the modelling result of each operation.

It is possible that you are not taking into account the impact of the pip value on the deposit currency. The result of each trade should be +/- lot*(TP / SL)*pip value.

Thanks for the script, it has cleared up some doubts, although I still have some questions, since this strategy is used with fixed SL and TP, can't it be used with variable SL and TP?

 
isanchez96 #:

Thanks for the script, it has cleared up some doubts, although I still have some questions, since this strategy is used with fixed SL and TP, can't it be used with variable SL and TP?

Of course, you can use floating stop losses and still make a profit. Also, according to my observations, the optimal stop losses and profit taking for buy and sell positions should be different from each other. The only requirement is that the mathematical expectation is strictly positive. I used fixed values to simplify the modelling.