Here are some answers to questions you might have about my Yuyostox algorithm. Click on a question to reveal its answer.

  • Q. How is this algorithm able to produce such exceptional returns?
  • My algorithm uses a blend of various artificial intelligence techniques that I have refined over 30 years. The result is a system that intelligently identifies specific patterns indicating short-term growth (at least 5% within one month).

    Applying this to recent, actual U.S. stock data, I was able to obtain a strong profit over a one-year period, with per-annum returns of 30% to 130% based on recent, actual U.S. stock data. The important thing here is that the algorithm has NO parameters to tune its performance, and it requires zero human input other than the starting date.

    All buys are based entirely on the signals output by the algorithm and sells take place when a profit target or a stop loss limit is reached, failing which the stock is liquidated one month later.

    As humans, we all have limitations, and when it comes to stock picking, we will sometimes allow subjectivity to prevail, based on an emotional response to some analysis we have read, our own study of fundamentals or technicals, how we think the market is moving or an unfounded belief that our stock market instincts can be trusted.

    My system is completely automated–it works with absolutely no human judgment involved, and does not change no matter what the market is doing–no manual adjustments of any kind.

  • Q. What exactly does the algorithm do?
  • On the morning of each trading day, prior to market opening, my algorithm studies stock prices and trading volume for six months up to that day, across more than 4500 NYSE and NASDAQ listed stocks, spanning about 150 industries.

    First, a shortlist is produced, eliminating penny stocks, stocks with low trading volume, etc. Then, each shortlisted stock is compared against its industry peers, based on evaluating and quantizing a number of what some might consider intangible factors, such as strength, sustainability, predictability, accumulated momentum, etc.

    In many cases a short-term growth signal to buy a specific stock at a certain price does not emerge–the signal only appears for any given stock if the system has high confidence in predicting a 5% rise in price within one month.

    Whenever a signal appears, which is quite infrequent, my algorithm reports the stock name and the signal price. The named stock should be purchased within the same trading day if it can be purchased at around the signal price (at most 1% above the price), or lower. Otherwise, the signal is ignored.

    Excluding ignored signals, there are generally about 10 to 20 valid signals over a one-year period. Please see the results page on this site for the results of applying my algorithm to two years of actual U.S. stock data.

  • Q. How frequently do signals appear?
  • Past records show a frequency of between 10 and 20 signals per year. However, this is completely unpredictable, as the algorithm will only output a signal if it believes that there is a very high probability of a short-term gain of 5% within less than a month.

    This will sometimes result in a long gap between signals. For example, April 26, 2018 to June 27, 2018 was a period of two months without any signal. But on the 7th and 8th of November, 2017, the signals were only one trading day apart.

    On the other hand, there were 10 signals in January 2019 alone.

    So do not be alarmed if there is no signal for some time. Instead, please be prepared to act when it does finally appear!

  • Q. What do I get for the subscription price of US$360/year?
  • Subscribers have access to the Members Only area of this site. This is updated each trading day, before the NYSE and NASDAQ markets open. Subscribers may choose to have the same information sent to them by e-mail.

    Most of the time, there will be no signal, but when there is a signal, the stock name and target price is shown to members. If you wish to act on a signal, you should buy the named stock only if you can get it for a price at the signal price (or at most 1% above), or preferably below the signal price. The purchase should be made within the same trading day, but if the price rises and the conditions mentioned above are not met, the signal should be discarded.

    That’s it! I do not publish any kind of market overview, specific industry or stock analyses, etc. Just the output of the algorithm, each trading day. Is that worth $360/year to you? It is a decision only you can make, based on your perceived value, if any, of my algorithm’s potential to help you make significant amounts of money!

  • Q. How does the free one-month trial work?
  • This is a low-risk way to try before you buy the Full package. You get all the same privileges as the Full Package subscribers, except that you can only use this offer once, and your package expires after one month, not one year.

    The main purpose of the trial is for you to satisfy yourself that this is not some shady operation, and that I do in fact update my web site daily with the latest output of my algorithm, and that I send out e-mail notifications if that if what you wish. If you like my system, please upgrade to the Full Package–this can be done at any time.

  • Q. How much risk is there of losing my original investment, or perhaps a lack of liquidity?
  • Please carefully read the disclaimer section on the Home page of this site. It states that nobody can predict the future with certainty, and it is impossible to guarantee that you will never lose money if you use this algorithm.

    However, what can be stated with certainty is that tests of my algorithm against actual historical stock data for twelve one-year periods, showed only profits and no losses, with annual compounded profits of 30% to 130%. Furthermore, a review of all signals generated showed that about 75% of signals resulted in profit taken, while 25% of signals resulted in a stop loss.

    This is how I use my algorithm (you should, of course, have your own trading strategy): Whenever my algorithm generates a buy signal, it indicates a signal price. I try to buy at around the signal price (at most 1% up), or preferably when the price falls below the signal price.

    I generally set a profit target of 5% up from the signal price, and a stop loss limit of 10% down from the signal price. So there may be some unprofitable transactions over the course of a year, but based on our full set of results, they are more than made up for by the profits taken.

    One way in which the loss of your investment may conceivably happen is the rare case, not encountered in any of my tests but imaginable, in which a company is delisted and your holdings become worthless. In the first place, these companies are very unlikely to be identified as buy targets by my algorithm, but I am mentioning it as an outside possibility.

    For example, if massive embezzlement or some financial irregularity is discovered at a company suddenly, and the exchange puts a stop to trading in its shares, you will be unable to liquidate your holdings. For this reason, my algorithm avoids penny stocks, which tend to be riskier, and if a stock price is below $5.00, it is only selected if there is a very obvious upside in the very near term.

    Similarly, a recent average daily volume of less than 500,000 shares disqualifies a stock from consideration. Again, unless there are very exceptional circumstances, this is a safeguard against not being able to sell your holdings due to illiquidity.

  • Q. Why are you sharing this information with subscribers, instead of just using it for your own trading?
  • I get this question a lot! Usually, it is more along the lines of “If your system is so great, why ain’t you rich?”. Well, the truth is, I have been using the algorithm to trade using my own funds, and I have been obtaining significant profits. However, the amount of money I can invest in the stock market as an individual is limited.

    I would love to set up a hedge fund, and manage other people’s money, taking a piece of the profits, but I am a computer science researcher, not a licensed investment adviser or portfolio manager, so I cannot do so under the law.

    Hence, the only way I can legally monetize my invention is to be a publisher, publishing the results of the algorithm on a daily basis. I do not know anything about my subscribers’ trading portfolios, objectives or risk profiles, and I neither answer investment questions nor do I provide financial advice or planning of any kind. I simply publish the results of my algorithm on a daily basis, to subscribers wishing to receive that information.

  • Q. To use the system, how much monitoring and trading will I have to do?
  • Not much. On many days, there will be no signal. Over a year, there may be 10 to 20 signals, in total (but please see the question about frequency of signals, above). How you use these signals is entirely up to you, but I will describe what I do. To trade using the algorithm, I simply wait for a buy signal from the system, and then purchase that stock within that same trading day if its price is at 1% above the buy signal or lower.

    I myself monitor my holdings daily, but people with busy schedules could then set limit orders with their broker, at 5% above the signal price (not the actual buy price), and a 10% stop loss. If a month goes by, and neither the profit target nor the stop loss is reached, I simply liquidate the stock at market price (more often than not, I realize a small profit rather than a loss).

    Every time I cash out on a position, I roll the funds into the next buy signal to obtain the benefits of compounding. After one year, I take the accumulated profit, and start over. That’s it, easy peasy!

  • Q. How do I become a member?
  • Simply click on the “Sign Up Now” button at the foot of this page. The signup process involves e-mail verification, so be sure to enter your e-mail address correctly. Everyone starts out with the trial package, which is free, but once you log in using the password sent to you by e-mail upon signing up, you will be able to purchase a paid upgrade to a Full package.

    When you log in from the “Member Login” item of the menu at the top of the page, you will be shown all the latest signals, but remember that the only one that matters for trading purposes is the signal, if any, for the current day.

    Renewals: any time you purchase a Full membership, your membership is extended by a year. At any time, you can see the expiration date of your membership by simply logging in as a paid up member.

  • Q. How do I cancel my subscription?
  • Rest easy–I do not charge your credit card year after year. You have to manually renew if you wish to do so. Because there are no automatic payments, and no refunds, there is no real need to cancel. You can simply switch off the e-mail service (on the Member’s Page), and never log in again. Your membership will expire.
  • Q. Machine learning and neural networks are trendy in artificial intelligence research. Why are these techniques not used here?
  • One simple reason. The algorithm is now frozen and will not be developed further, as it performs well, and predictably. Allowing it to learn beyond this point assumes the risk that its performance may degrade. Moreover, any learning component would likely attempt to fit data to the curve, in the sense of evolving to reflect the latest market data.

    Because my design was from first principles, rather than a data-driven modelling process, my thesis is that this algorithm captures core characteristics of NYSE/NASDAQ markets, and unless there is a fundamental change in structure (e.g. overall number of counters sees a significant change, or regulations of some sort that impose constraints on the free trading of listed stocks, or basic price and volume information become unavailable for some reason), there is no need to fix something that is not broken.

    However, at the beginning of each year, I feed the previous year’s recommendations and results to the algorithm, and it adjusts accordingly, so it does undergo some learning, in that respect.