R-Portfolio - a diversification method

 

Happy Holidays and Happy New Year to all!

I have more or less finished my previous development. I am putting it out there for the local residents to judge.

----------------------------------------------------------------------------------------------------------
What is the fundamental difference between R-Portfolio and Modern Portfolio Theory methods?

Let's look at the basic tenets of Modern Portfolio Theory:

1. For the evaluation of individual portfolios of securities and the evaluation of individual securities the same methods are applied, which are based, as a rule: the ratio of the potential profitability of financial instruments to the standard deviation - risk.
2. The risk of non-risky securities, as well as portfolios formed from these very securities, according to the Modern Portfolio Theory, can under no circumstances be negative.
3. A portfolio of securities reduces risk, at the cost of lower returns. I.e. to increase the potential yield, it is necessary to risk more.


To be brief, the Modern Portfolio Theory, offers methods with the help of which the investor, having selected a list of securities and having calculated their statistical characteristics can:

1. Specify the amount of potential risk and obtain a portfolio. By multiplying the shares of the portfolio by the returns of the securities and adding up the results, you can obtain the potential return of the portfolio
2. Set the value of the potential return and obtain the portfolio. Calculate the risk for the resulting portfolio.



Now consider the same points with respect to R-Portfolio:

1. R-Portfolio portfolios refer to a separate set of portfolios consisting of securities and are synthetic financial instruments that benefit from a strictly upward stepped trend without dips (drawdowns) compared to other financial instruments. If the time series of quotes of separate securities in the portfolio is a random walk on Bernoulli's scheme, i.e. consists of a number of Japanese candles of different colours, with various highs and lows, coming from these very bodies, the R-Portfolio is a synthetic instrument which consists only of candles with always white bodies - growing candles. Random Bernoulli rambling of this very portfolio applies only to the yield highs, as they are unpredictable and not constrained by anything. The candle body for an R-Portfolio cannot be less than some certain value.
2. The potential risk for any R-Portfolio, is always negative, even when there are no risk-free securities in these very portfolios.
3. There is no concept of potential risk for R-Portfolio because this very risk is always negative, i.e. for any period of time the portfolio can only give potential profit. Instead of potential risk, the notion of potential minimum portfolio return over a period of time is introduced. The maximum potential return is unpredictable and not limited by anything from above. The geometric space of R-Portfolio portfolios is one-dimensional and is characterized only by the values of potential minimum returns of individual portfolios for a given period. The relation between the potential minimum return of the R-Portfolio and the potential risk is inversely proportional. i.e. the higher the potential minimum return of this very portfolio, the lower the potential risk. For this reason an investor has no choice, as R-Portfolio calculation method will result in portfolio with the highest potential minimum return, and consequently, the lowest risk.

To be brief, the R-Portfolio method allows to get a portfolio with the highest potential lower limit of the yield, by converting the list of securities, by converting their quotes. In order to obtain a potentially worst-case portfolio, one or more of the assets in the R-Portfolio generated by the full list must be removed from the list of securities for which the previous portfolio was formed.

----------------------------

This is just a description of the R-Portfolio in the spirit of, "differences from conventional powders". More detailed information, i.e. algorithm, software implementation, as well as open-source software can be found at: http://r-portfolio.ru

At the moment there is a Java implementation compatible with MT4. Subsequently a full-fledged MT5 Expert Advisor written entirely in MQL5 is planned,
---------------------------

Since the above website is currently not available for feedback, please: Discussions, complaints or requests can be posted in this thread
 
Tantrik >>:

Зависимость между потенциальной минимальной доходностью R-Portfolio и потенциальным риском обратнопропорциональная. Т.е. чем выше потенциальная минимальная доходность этого самого портфеля, тем ниже потенциальный риск.


Чем выше доходность - тем выше риск. Чем ниже доходность портфеля тем более там надёжных бумаг с медленным ростом(имхо)


I used to think so too. In the Tobin-Markowitz portfolio there is indeed a direct proportionality between potential return and potential risk. But in this case it's the other way round. That is, for example, the minimum potential return for one portfolio over a period of time is $x. The minimum potential return for one portfolio over a period of time equals $y.

With x > y > 0

Since the maximum return is from minimum to infinity (Sortino's coefficient is infinite), and the risk zone starts from zero return and below, a portfolio with m.p.a. $x is less risky than a portfolio with m.p.a. $y, as it has the bar further away from the risk zone.

 
Tantrik >>:



Набор слов - abracatabra (имхо)

It's elementary, Watson. The risk of falling into the abyss is higher for the one who is closer to the edge of the abyss.

 
You can make it even simpler.

Suppose there is a bank that is contractually obliged to pay each depositor a mandatory n-th amount (% of the deposit) + an optional premium every month. That is, if the bank is doing too well, it is obliged to pay the n-th amount. If the bank is doing well, it pays out a premium from the additional income.

Where is the contractual risk for the depositor here? It is clear that there is no risk for the contractual investor. But there is a potential risk that the bank may default or go bankrupt.

How does this same bank operate? It buys assets that have a potential positive yield. But these same assets have unstable returns, which means that in some short periods of time they may show losses. The bank forms a portfolio of these securities in such a way that each period of time there is a return of at least a fixed amount - the minimum potential return. The maximum return is not taken into account - it is impossible to calculate, i.e. it is non-stationary. And the required n-th amount of payouts to the depositors is calculated with this very min. If the portfolio yield exceeds the min. interest rate, the depositors receive a bonus from the additional yield. Everybody is happy, everybody laughs.

The R-Portfolio methodology is an algorithm for forming such potentially risk-free portfolios from unstable potentially profitable assets.

The essence is that non-stationarity of quotes of separate assets, in the portfolio, is transformed into a stationary minimum potential yield, by increasing non-stationarity of the maximum potential yield. Portfolio returns are, to a greater or lesser extent, depending on the shares, the average returns of the assets in the portfolio.

That is, it is not a perpetual motion machine at all. The elimination of non-stationarity - risk in one place - comes at the expense of increasing it in another.
 
So the algorithm combines stock price curves so that the synthetic has an even up-trend?
 
neoclassic >>:
То есть алгоритм комбинирует кривульки цен акций таким образом, чтобы синтетика имела ровный ап-тренд?

No, the up-trend is stepped, no dips. There is only a fixed minimum height of each step. This makes the maximum height of the steps even more unpredictable than for the individual assets in the portfolio.


A linear uptrend is usually observed with Martin. But TC with Martin has the equity line below the balance. R-Portfolio has equity above the balance.

In addition, Martin's balance linearity is achieved by increasing lots (bets) at unfavourable times.
 
I would like to join the previous speaker.
The work is extremely interesting, but has any research been done into the application of this methodology to foreign exchange markets?
 
HH: as you know, most of the most popular brokerage companies (which work through MTs) have extremely stingy securities listings, and it seems doubtful that one can create an optimal portfolio out of such a small number of assets.
Therefore it would be very interesting to know how (and if at all) this method could be applied to currencies or CFDs.
 
Tantrik >>:
Полность с Вами. Но Маэсто мы на форе. И вот вопрос(вы его и ждете?) это можно применить к форексу? К его нестабильности? К отсутвию надежных безрисковых активов?

Don't speak for everyone. There are traders on this forum who trade not only currencies, but also stocks, futures contracts and spot metals.

If you are not allowed to trade anything else besides currencies, please find other threads for flaming.

 
lexandros >>:
Присоединюсь к предыдущему оратору.
Работа чрезвычайно интересная, но делались ли какие либо исследования в направлении применения данной методики к валютным рынкам?

No. For currencies, nothing but a history fit was possible. The reason is banal. Currencies can be counted on the fingers, although there are a lot of currency pairs obtained by combining them.

For the implementation of this method it is necessary to have at least several tens of financial instruments quoted in one currency, and preferably several hundreds because a portfolio needs to be formed out of a variety of investment assets most of which will not fit in it and will be rejected because of low trustworthiness. This is why positive results have only been achieved in the stock markets, as there are at least 30 financial instruments on the markets.

 
Reshetov >>:

Нет. По валютам ничего кроме подгонки под историю получить не удалось. Причина банальна. Валюты можно пересчитать по пальцам, хотя валютных пар из них путем сочетаний получается большое множество.

Для реализации данного метода необходимо как минимум несколько десятков финансовых инструментов котируемых в одной валюте, а желательно несколько сотен, ведь портфель необходимо формировать из множества инвестиционных активов, большая часть которых в него не попадет - будет отсеяна по причине низкой благонадежности. Поэтому положительных результатов удалось добиться только на фондовых рынках, т.к. на них представлено не менее 30-ти финансовых инструментов.

a direction, by the way, for the funds is promising.
Only I did not understand why it is called the Reshetov method:)))