F.A.Q.

The program has not been tested on Apple’s operating systems. When demand becomes evident, we will elaborately test and adapt the software to also run on Apple’s operating systems.

Our software runs smoothly on Windows 10. It has been elaborately tested on simple and advanced laptops and desktops. For larger portfolios of more than a hundred stocks, a user will notice the difference in processing time between running the estimation techniques for optimum portfolios on a simple laptop and on the most advanced desktops.

For instance, automatically composing, ranking and weighting portfolios of 1500 long and 1500 matched short positions takes about 30 seconds for a weekly validation over 11 years on an advanced desktop with an Intel i9 extreme edition processor, a pair of 8GB active memory set at 3200MHz, and two high-speed SSD’s configured in RAID0. This processing time increases by a factor of eight to ten when a simple laptop is used.

In statistics, one has introduced the concept of expectation values. For instance, the present life expectancy of 60-year-old males in The Netherlands that do not smoke is about 82 years. That does not guarantee anything about how long any particular 60-year-old male still has to live. However, we can say with reasonable confidence that, past “performance” from “womb to tomb” was on average some 82 years under the conditions that the male was 60 years old, lived in The Netherlands, and did not smoke.

In a similar way, we can quantify with reasonable confidence the expectation values of risks and rewards of optimum portfolios based on past performance from “womb to current” under a given set of some 35 conditions and personal circumstances for stock screening, ranking and weighting. Shouldn’t an investor or trader be legitimately interested in his risks and rewards he can expect from his investment or trading scheme in the future?

We do not know of any program other than ours available to the retail investor that answers this question.

In May 2017, the Automatic Programming Interface (API) of Finance.Yahoo.com, the public domain site for US historical exchange data, changed unannounced. Therefore, we had to adapt our data-collection software, which cannot just be done and tested overnight.

As timing is crucial for rebalancing portfolios, one usually cannot afford to be cut off from the current historical data. To minimize potential errors and download uncertainties, we are happy that CSI was willing to work out a customized solution with us aimed to providing Clean historical data for our analyses in secured timely download processes.

CSI may offer more timely price and volume corrections, IPOs (Initial Public Offerings), Splits, and Dividends, and its service may be more reliable, providing free customer support by phone and email. It offers this service directly to our users for a modest yearly subscription fee of $165.

There are armies of experts that regularly give us guidance about stock price, company, market and economic developments. If you would consistently know these guidance statements ahead of publication, you could consistently bet on the outcome. That would be a lucrative business, but it would not be legal as that is interpreted as trading with foreknowledge.

The fact remains that the future guidance by experts on fundamentally unpredictable processes may induce relatively short-term market movements that could be exploited in short-term interactive trading strategies.

Our algorithms completely ignore the continuous stream of breaking news and only take the end-of-day historical exchange data of Wall Street as a measure of historical performance. Optimal portfolios by definition can only be selected when risk and reward objectives are satisfied over the long run, irrespective of this continuous stream of breaking news.

DigiFundManager enables a user to screen and rank stocks based on relatively short-term value fluctuations. The program enables a user to group screened and ranked stocks into portfolios and time and weight those portfolio positions based on his long-term investment objectives. It then enables a user to validate the conformance of his investment plan to his investment objectives over 10 - 55 years. To our knowledge, there is no other software program available to the retail investor that enables this validation.

Modern Portfolio Theory reduces portfolio risk by selecting and balancing assets based on statistical techniques that quantify the amount of diversification by calculating expected returns, standard deviations of individual securities to assess their risk, and by calculating the actual coefficients of correlation between assets, or by using a good proxy such as the single-index model.

Like practically any portfolio theory, MPT is based on the linearization of the solution space, usually posing additional assumptions on this space. These theories select different portfolios when time progresses.

Such portfolios will fluctuate in value over time. In these value fluctuations, the drawdowns and annualized rewards of the portfolios determine the risks and expected returns rather than the standard deviations of individual securities. The question of how portfolios selected by MPT behave through financial crises and recessions in terms of risks and rewards is not answered by MPT.

That validation of sequential investment decisions in terms of portfolio risks and rewards through economic cycles of booms and busts is quantified by DigiFundManager.

According to Wikipedia and Bloomberg, one of the best fund managers is the mathematician Jim Simons, the founder of Renaissance Technologies. His quantitative investment strategies achieve risk/reward ratios approaching zero over 29 years with annualized rewards averaging 41%/year of multi-billion-dollar funds.

It appears to us that the keys to success of such algorithms are (1) the short-term, even intraday correlation times down to the order of a second that determine the ranking of the assets and (2) the estimation technique to calculate the optimal asset allocations for each rank in line with the investment objectives of the investor.

Stock ranking is used to rank the stocks in a given collection at any time in terms of decreasing likelihood to increase in price. Renaissance Technologies mostly self-invests in its funds. Once you have the digital tools like DigiFundManager to self-invest your money, you are able to getting to trade down to each week or multiple of weeks the optimal stock portfolios of Wall Street with expectation values of risk/reward ratios validated over 20+ years that you yourself may be able to get well-balanced of the order of one or even smaller.

According to Bloomberg, https://www.bloomberg.com/news/articles/2016-11-21/how-renaissance-s-medallion-fund-became-finance-s-blackest-box , this is what Renaissance Technologies used to develop their highly effective Medallion Fund strategies during the past 30 years.

When you are a pioneer in a field, you usually don’t want to take any chances when it comes to possible critical paths in your applied research.

Thirty years ago, computing power was one of them, and brain power to develop effective statistical ranking techniques and efficient search algorithms for estimation techniques another one. Once you know that these correlations are out there in the statistical behavior of the financial markets, they are easier to find.

According to Moore’s law, the density of computer power on a chip increased by 2^15 = 32,768x over these 30 years. The convergence of estimation techniques has been in focus of active research, especially in the field of signal processing.

The availability of historical financial data on Wall Street stocks has developed into a major business arena. According to the same Bloomberg article, quantitative strategies are now outperforming classical fund managers in absolute dollars.

DigiFundManager brings the configuration of quantitative strategies to the retail investor for the first time. It enables the retail investor to design his own unique quantitative game plan that fits his own maximum risk endurance and satisfies his own reward expectation.

These two models appear to represent the two camps of quantitative trading. Trend-following originates from what technical analysts call “the trend is your friend (until it bends)”. According to financial experts, mean-reversion implies that the statistical return variations are not a random walk, but that periods of lower returns are systematically followed by compensating periods of higher returns.

Economists use the tautology “regression to the mean”, which appears to imply that being lucky in one instance has relatively high probability of being unlucky the next. These two models must be rewritten in terms of a ranking system where stocks are ranked according to their probabilities to increase in price. You are then able to select the long positions with the highest probability to increase in price and the short positions with the highest probability to decrease in price. The stock with the highest probability to increase in price has rank one, etc. Trend-following and mean-reversion are statistical processes that analyze the statistical behavior over time.

DigiFundManager assumes ergodicity for the statistical behavior of the financial markets. Ergodic behavior assumes that the expected behavior over time equals the expected behavior of an ensemble of stocks traded in the same market at each time. The ensemble of stocks is defined by your Watchlist. Hence, rather than performing the time-consuming processes of analyzing the time-series on trend-following and mean-reversion, DigiFundManager simply ranks the stock in a given Watchlist for a given set of correlation times at each specified time of trading.

The ranking model that appears to work the most effective is that the steepest fallers and climbers in a Watchlist have the highest probability to turn around at each specified time. DigiFundManager allows you to base your ranking on two correlation times of your choice. The shortest correlation time is one week and all possible correlation times are a multiple of weeks. Usually, shorter correlation times give higher expected returns and maximum risks. It is not surprising that correlation times of (multiples of) 13 weeks can be found over 30+ years in a Watchlist of the 505 current S&P500 members to produce optimal portfolios of 50 long and 50 short positions with well-balanced risk/reward ratios.

DigiFundManager assumes that all trades are done in perfect competition. That implies that stocks are screened for traded daily-dollar volumes that are large compared to the trade sizes planned by the investor.

Trading such stocks should then only negligibly move the market. According to Bloomberg, almost from the beginning, Jim Simons understood that the fund’s overall size can affect its performance: too much money destroys returns. Experts who designed algorithms for quantitative strategies have expressed that researching how your own trades move the market is not an easy challenge.

For a retail investor, planned investment amounts will usually leave ample room for large numbers of available stocks with sufficient daily-dollar volumes over the past 30+ years. So you can validate the automated sequential decision-taking process of weekly selecting optimal portfolios of the size of your choice in terms of the risks and expected rewards they produce over those 30+ years.

In our opinion, when you keep the solution space linearized, it is not difficult to extend the estimation techniques we apply with a model for a linear feedback mechanism associated with trade sizes. However, in the retail investment market, that extension has little relevance.

At times of their choice, fund managers keep reselecting their portfolios by screening and ranking their assets, possibly from a Watchlist. They then weight their financial assets with the objective to produce the risks and rewards targeted by their clients.

The only data needed to quantify the risks and rewards of this sequential decision-taking process of rebalancing already selected portfolios is the continuous stream of historical exchange data. As many brokers offer trading Market On Close (MOC) orders, trading on end-of-day (eod) prices offers a sure case to enable validation of the risks and rewards of sequentially selected portfolios over time using simple eod data.

The timing and weighting processes of fund management are optimization processes in terms of risks and rewards. Hence, these processes only need the simple historical eod data. Screening and ranking of stocks can be done using the simple eod data and/or using any breaking news on financial, economic, and political analyses and outlooks that may influence the financial markets.

Portfolio rebalancing on breaking news requires intra-day trading, which cannot simply be validated by a back test. DigiFundManager only uses simple eod data for its quantitative screening and ranking processes and assumes that investors and traders who choose to apply the program use MOC orders.

Its own weekly screened portfolios of top-ranked 250 stocks outperform the 24%/year rewards of the weekly screened portfolios of 238 Zacks’ rank #1 stocks by roughly 10%/year over the 18.4 years prior to April 2018 when we performed this comparison test.

Our stocks were screened for minimum daily-dollar volumes of US $0.1million. Zacks uses its own ranking system based on forward-looking estimate revisions of some 3000 analysts at over 150 different brokerage firms.

At any given point in time, they are monitoring well over 200,000 earnings estimates and other related data looking for any change. Our weekly screened portfolios of the top-ranked 25 stocks generate 88%/year. It has yet to be seen whether intraday-trading is more effective in terms of risks and rewards than eod-trading.

Contrary to DigiFundManager, quantitative trading software like SmartQuant, QuantShare and DLPAL searches for successful intraday charting patterns over time.

It is well known that successful patterns have a finite shelf life and suffer from alpha decay. They can come and go, and come back and go again, but you never know when, only in hindsight.

It is believed that the reason rather than cause of this performance decay is the efficiency of the financial markets. One way to measure alpha-decay in the stock market is to measure the t-stat of your investment objective over time. In simple performance terms, this implies that the free cash generated by a sequence of optimal portfolios with constant investments is decaying over time.

DigiFundManager is designed to practically instantly reveal such performance decay validated over time spans of 20 – 55 years for each individually configured quantitative investment strategy. The program shows that successful investment strategies of large investments are usually short-lived.

However, there is room for smaller investments that accommodate retail investors for already 20+ years. The key to individual success is each individual configuration of our screening, ranking and risk modules in combination with an individual watch list. Broker fees play an important role, but the possibility to go short in any of the entries in your watchlist is even more important. It appears to us from the London-Whale incident that the results of apple orchards shouldn’t be hedged by shorting the results of orange orchards.

Short positions intend to take care of the damage control during drawdowns and recessions. Each quarter for already 30+ years, the slowly changing watchlist of the 505 current S&P500 members keeps enabling to configure optimal portfolios of 50 long and 50 short positions with well-balanced risk/reward ratios producing about 12%/year net of broker fees and taxes. The shelf life of this quantitative strategy is already 30+ years as our last demo reveals.

DigiFundManager allows you to use any watchlist and screen, rank, time and weight optimal portfolios of the size of your choice from those lists. The screening, ranking, timing, and weighting is done automatically by the program after you configure these processes based on your personal choices and circumstances, including your broker, tax and data provider.

All of those resulting game plans are discussed as tutorials in the Manual. Our choices for configuring those game plans usually give risk/reward ratios of roughly 1 – 1.5, except for the one based on investment funds. That latter game plan turns out to be highly risky through the last financial crisis.

It has the highest risk and lowest reward. Of the remaining game plans one needs to rebalance the portfolios every week except for the game plan that quarterly picks the optimal portfolios from the current S&P500 members. Except for the last one, you can see so-called alpha decay. Over the past eight years, these three game plans show decreasing performance.

As these three game plans require weekly rebalancing, broker fees and dividend tax further significantly exert a downward pressure on the performance. The game plan with highly-covered stocks gives the best performance. The game plan based on current S&P500 members gives the most steady performance without any alpha decay over as many as 30+ years.

It is known that only about 5% of the brokers’ accounts for stocks in the end makes money at the cost of roughly 95% that makes a loss. A recent analysis can be found on VantagePointTrading. Those percentages appear to hold for the equity markets and for the crypto-currency markets (Distribution of wealth).

Hence, as an investor, your probability to make a long-term profit on Wall Street is of the order of 5%. This holds for retail as well as for professional investors. This small probability stands remarkably in contrast with the probability that companies create a long-term investor value above their IPO’s. These latter probabilities can simply be calculated by DigiFundManager with progressing time by realizing that all IPO’s are normalized to “1”.

By calculating the ratio of screened and unscreened stocks where screening is on adjusted share prices larger than “1”, the weekly probability can be charted as in the following chart:

Probabilities of success for investors’ stock accounts (green curve) and for adjusted share prices of company stocks relative to their IPO’s (red curve) starting on April 2007, the onset of the credit crisis. A minimum daily-dollar volume of US $0.1 is assumed for the stocks.

It is remarkable that the likelihood of success for the companies in terms of their investors value increasing above their IPO’s is more than an order of magnitude larger than the likelihood of success for stock brokerage accounts. The chart shows that since the onset of the credit crisis in April 2007, currently some 11 years ago, the likelihood of success for companies for delivering long-term investor value increasing above their IPO’s is slowly decreasing from 88% to 75%.