Yuyostox – my amazingly smart
A.I. stock picking algorithm

Hi, I'm J.H. Lim, a Canadian computer scientist working on applying artificial intelligence to stock market analysis (see my bio under "Learn More").
Are you a Registered Investment Advisor (RIA), portfolio manager or professional investor? Do your clients demand better returns on their equities portfolios? Could your stock trading performance use a boost?
If so, I have some information that you may find interesting. This site is about my YUYOSTOX System, an AI stock picker algorithm I created that has so far produced very high returns on U.S. equities portfolios (30% to 130% p.a. in recent measurements).
Short Term Profits: Every day, my algorithm scans over 4,500 stocks in almost 150 industry sectors, yet it generates only 10 to 20 buy signals a year. The reason is that it looks for an exceptionally high level of confidence that a stock will rise 5% within a few days. Just as with day trading, compounding these small profits results in a significant return on investment over the year.
Detailed, unedited results of applying my algorithm to actual U.S. stock data for 12 recent, one-year periods can be viewed here.
Signing up for membership gives you access to my algorithm's buy signals as they occur. The page is updated on each trading day (10 to 20 buy signals per year), and if you choose, the results can be sent to you by e-mail before the market opens each day.
NEW: Not ready to buy? Then the one-month Trial Package (FREE!) is for you. It's the same as the full plan, but limited to one month, and you can only use it once. Whenever you're ready, upgrade to the paid plan: US$360 for a full year of algorithm results.
Recent actual test results using NYSE/NASDAQ market data (12 one-year portfolios):

  • Return of 30%-130%/year
  • 10-20 buy signals per year
  • Positions closed within 30 days
  • Avg profit/position: 5-15%
  • Avg loss/position: 5-15%
  • Ratio of profit/loss positions: 3:1
Image Image
Use the power of AI in your investment strategy! Sign up now to receive daily updates of my algorithm's output, including all buy signals for one year.

Who is it for?

Stock portfolio managers, regd investment advisers (RIAs) and professional investors who:

• manage a portfolio that includes substantial holdings in US equities
• hope to achieve annual returns much higher than the S & P 500.
• are able to take 10-20 long positions, liquidating each within one month (20-40 buy and sell trades/year)
• can wait one year for the returns

A.I. Stock Picker

Yuyostox is my proprietary AI algorithm for stock price/volume technical analysis, developed over 30 years by myself. It is based purely on technical analysis and excludes fundamentals. Technically speaking, this is an innovative, non-neural, abstract feature recognition, multidimensional AI algorithm for equities analysis of North American markets, not published in any journal, nor available anywhere else.

Actual Results

During the bull market of recent years, running the algorithm on actual stock price and volume data from Jan 2016-May 2018 showed that the system produced 30% to 130% returns per annum (before taxes and brokerage) for 12 one-year periods. These results were considerably superior to the S & P 500 index, and were obtained by the algorithm without parameterization, or human adjustment of any kind.

How it works

Each year, 10 to 20 buy signals are generated, with no human judgment involved, assuring objectivity. Automatic profit targets and stop loss limits determine when to sell, but if these are not reached, the system indicates that stocks are to be liquidated within a month. After each sell, the funds are compounded into the next buy signalled by the system, and this is repeated over one year.


First of all, I am not paid by any company or promoter to recommend any particular stock. Neither do I buy a stock and then manipulate the algorithm to recommend that stock. Yes, I may buy stocks the algorithm recommends, but my total outlay per purchase is generally less than CAD$10K, so obviously I am by no means influencing the price in any way.

I tested my algorithm on actual NYSE/NASDAQ data for over two years, from Jan 2016 to May 2018. No curve-fitting or parameterization was involved. The algorithm runs without human input or judgment of any kind, avoiding the limitations of human subjectivity.

Results of testing 12 simulated US$10K portfolios on actual stock data for 12 one-year periods starting a month apart from July 2016 to Jun 2017, showed returns of 30% to 130% per annum. I chose this period based on data available to me (a window of 6 mths before and 12 mths after the start date is required for each test).

A list of all buy signals generated by the algorithm over the same period showed that 75% of the buys resulted in an average gain of 1.62%, including 30-day liquidation sell transactions, over an average span of 13.29 days (annualized return of 38.88%).

Although about one out of four signals results in a loss, in fact, as can be seen in my results based on actual market data, it is the 75% success rate that really matters when it comes to the annual return. Simply rolling the proceeds, whether gain or loss, into the next buy resulted in 30% to 130% compounded annual returns in the 12 different one-year periods I tested. Each period was started one month later than the previous period.

The algorithm is based on accumulating several small gains of around 5% each time, compounding to yield an impressive annualized return. Positions are not held for longer than one month, ensuring constant turnover.

However, note that past performance is no guarantee of future performance, so going forward, , the returns based on my algorithm, if any, cannot be guaranteed to be at the tested levels, or any level for that matter!

As developers and owners of the Yuyostox algorithm, neither I nor my wholly-owned Canadian company Yuyosoft Innovations Inc., is a registered investment adviser, and we offer neither general nor specific investment advice.

I am not a financial industry person, but rather a researcher publishing my results, namely a log of my algorithm's output, on each trading day to paying customers, for their reference. I do not recommend that they follow me, they are free to do as they wish with the information I publish. I do not assess their risk profile, nor does my algorithm take into account any information about my customers or their holdings, nor do I provide financial advice of any kind, nor do I trade on their behalf, nor do I tailor my algorithm to suit any customer's requirements.

Customers are free to trade based on the algorithm output, or to consider it in their own stock picking process, or to ignore some or all of it, as they wish.

View actual results

Thank you for your interest in my algorithm. If you wish, you can view detailed results of the algorithm applied to actual stock market data, showing profits of 30% to 130%.

View Actual Results