How it works
Keep the finger on the pulse of the market and cater to your style of Investing with Upturn
In just 3
Receive AI/ML Based recommendations
Receive "Buy" and "Sell" recommendations daily after market close for stocks in your portfolios
Create Portfolios to plan and manage your investing
Plan and create a stock market portfolio on your own with Upturn either through a stock screener or directly adding stocks into a custom portfolio for the stock personas of your choice. Choose stocks based on Upturn Star Rating and Historic performance
Tap into the stock market to trade and invest smartly
Plan and create a stock market portfolio on your own with Upturn
Receive "Buy" and "Sell" recommendations daily after market close for stocks in your portfolios catering to your style of investing by using personas (some investor personas could plan to hold a stock for a year or the persona 'long', while some want to buy and sell a stock within a 'quarter' or a few weeks to a 'month')
Use Upturn-Star rating specific to the persona ('long', 'quarterly' or 'monthly') to get an indication of how the stock performed historically for that specific persona.
Upturn Personas to cater to your style of Investing
Upturn Corporation’s (the “Company”) proprietary algorithms provides alerts as to when a stock might potentially be on an “upturn” based on a technical analysis (including machine learning and quantitative statistical analysis) of historical data. Specifically, the Company identifies “upturns” or “BUY” recommendations in a specific stock by using the end of day stock data (price and volume) and apply quantitative analysis (i.e., machine learning and proprietary statistical algorithms) to identify and recommend these “upturns” or “BUY.” Alerts are provided to users daily after markets close, typically within 3 to 5 hours. Based on the post-market close analysis, the Company may recommend a “BUY” or “upturn” for a stock or alternatively, a stock which had a “BUY” or “upturn” may change to a “SELL” recommendation. Currently, approximately 300 large market capitalization stocks are covered by the Company’s for analysis to provide recommendation. The Company runs proprietary algorithms on each of these 300 stocks for each different persona after end of day market close to come up with updated recommendations for each stock.
These alerts are to be taken purely as “potential” upward indicators and the Company provides no guarantees on the stock movements, success of trades or profit. A user should always perform his/her own due diligence before making any stock purchase or selling decisions.
Every user has his own method or style of trading. For example, a user may want to trade frequently (i.e., buy a stock and sell it in short intervals of time) or may prefer to buy a stock and hold it for a longer period of time. The Company currently offers users 3 different investment strategies which are referred to as “personas” and represent 3 types of recommendations for every stock the Company analyzes based on a user’s trading style. A user can select from the following 3 personas:
A Long Persona
A Quarterly Persona
A Monthly Persona
In general, a persona is based on how long the average "hold" period for a stock. A “hold” period is defined as the time duration for a recommendation going from “BUY” to a “SELL” i.e., it is a measure of how long a user is going to be invested in that stock if following the Company’s recommendation. The average hold period for each persona is based on average of all hold times for the recommendations for a specific persona when performed for a large pool of stocks (for instance, 100 largest cap stocks) and this is calculated as an average in historic simulation starting from January 1, 2016. In this context, a "hold" period would generally be considered as the duration of time a specific stock changes from a "BUY" to a "SELL" recommendation. For example, the hold period for a “Long” persona is on average typically a year or longer. A “Quarterly” persona hold period averages typically 3-4 months and a “Monthly” persona is on average a month or more (i.e., 4-6 weeks). While recommendations of a persona for a stock is indicative of the hold time for a recommendation, hold periods may be longer or shorter, depending on such factors, including, but not limited to changes in trends due to various factors in the market conditions, company performance, quarterly results and others. This means that, for example, a "hold" period for a "Long" persona could be less than a year if the “upturn” or “BUY” trend analysis for a stock begins to change and the Company could recommend a “SELL” within a few days or weeks even though it is a “long” persona.
The Company currently uses anywhere from 5 to 40 different proprietary internal metrics to come up with a “BUY” or a “SELL” recommendation in connection with each persona.
The following example is for illustration purposes only and shows through simulation how “BUY” and “SELL” recommendations may vary for different personas for the same stock. The simulation is from January 1, 2016 through August 24, 2018 applied to each of the 3 personas based on Apple, Inc. (Ticker: AAPL) (Note: the method of charting might change over time with implementation). Past performance as shown in the simulation is not indicative of future results. Hypothetical performance (such as back tested results) does not reflect the performance of actual users and are not intended to serve as a guarantee of future results.
Long Persona for AAPL (Apple Inc.)
A “Long” persona, after a “BUY” recommendation for AAPL on March 17th, 2016, the recommendation didn’t change to a “SELL” and continued to be a “BUY” (or “upturn”) as of early Jan 2019.
Quarterly Persona for AAPL (Apple Inc.)
A “Quarterly” persona, after a “BUY” recommendation for AAPL average invested-days was 62 days (the 62 days was a calculation of the total ‘hold’ time during the simulation which involved 7 ‘BUY’ and ‘SELL’ cycles divided by then number of the cycles).
Some recommendations where over 3 months as seen between Nov 7th 2016 with a ‘BUY’ recommendation to ‘SELL’ in June 2017, which was actually over 7 months of hold time and a sell was not suggested as the trends did hold and didn’t trigger a ‘SELL’ recommendation.
Some recommendations were lesser but it is in the 3-4 months range as an average with few recommendations “hold” time being less than three months, for instance the first ‘BUY’ on March 3rd 2016 and a sell on May 2nd 2016.
Monthly Persona for AAPL (Apple Inc)
“Monthly” persona, after a “BUY” recommendation for AAPL on August 1st 2018, ‘SELL’ recommendation was in two months on October 10th 2016. However, average invested-days was 27 days (the 27 days was a calculation of the total ‘hold’ time during the simulation which involved 4 ‘BUY’ and ‘SELL’ cycles and one ‘BUY’ without a ‘SELL’ yet divided by then number of the cycles). Some recommendations where over 1 month, some lesser but it is in the 1 month’s range as an average.
The Company currently offers 3 different investment strategies which can be selected by a user for a specific stock. These investment strategies are referred to as “personas” and represent 3 types of recommendations for every stock [the Company analyzes] based on a user’s trading style.
The Company provides a star rating for every stock it has provided a recommendation for. These star rating represents the simulated historic cumulative returns of the stock based on the Company’s "BUY" and "SELL" recommendation specific to the persona ("Long", "Quarterly", or "Monthly") and are not fundamental ratings of the stock or company.
The simulations began on January 1, 2016 (date chosen to show at least a few years of historic consideration for the simulation) through last close price of the stock the previous day traded. A higher star rating [is not a current recommendation of the stock] and does not guarantee that the stock will perform well in the future, nor does a lower rating indicate poor performance in the future. Further, since the recommendations are based on technical analysis of historic patterns, historic performance is not guarantee of how a stock could perform in the future.
Descriptions of terms and metrics
The following section describes performance metrics and terms used
Note: Please see the Disclosures for Performance Metrics and Simulations for assumptions and considerations used in the simulation (hypothetical) returns and ratios.
Stocks and Stocks Chart (Simulated)
Following are terms and its definitions and assumptions used when showing the Company’s historic recommendations.
Note: The method of charting might change over time with implementation
This is simulated compounded returns expressed in percentage Compounded means, in the simulation, the assumption is, the user gained or lost x% returns on its initial principal invested, that entire principal plus (or minus) the returns was reinvested in the next set of recommendations. Returns can be negative or positive.
The compounded profits noted in the above chart takes into consideration the simulation of having reinvested the principal and returns for all the recommendations of the Company for that particular persona of the stock since the simulation start date of January 1, 2016 and is measured in percentage. The profits can be positive in case of a profit or could be negative in case of a loss. (The disclosures below provides details about the assumptions while calculating returns in a simulated transaction)
Avg Profits (or Average Profits)
This takes the average of all the returns (profits and loss) expressed in a percentage for all the “BUY” recommendation through the period of simulation start date of January 1, 2016 till previous day close. Some “BUY” recommendations might have a corresponding “SELL” recommendation which is called the “Realized returns” or might not have a “SELL” recommendation yet which is called “Unrealized returns”, i.e., this is calculated by the simulation which includes transactions which have completed a full “BUY” and “SELL” recommendation (also called realized profits or loss) and transactions which has had a “BUY” recommendation without a corresponding “SELL” yet (also called the unrealized profit or loss) and averages it over the total number of such transaction cycles. This is expressed in percentage. In short, this is the compounded profits by number of “BUYs”.
UnRealized (or Unrealized profits)
This is the profit (or loss) expressed as a percentage which has had a “BUY” recommendation without a corresponding “SELL” yet. The stock with this is still a “BUY” recommendation and the “SELL” recommendation could come later point in time. From a simulation point of view, the stock is still bought and has not been sold and hence the returns are not yet realized.
Number of BUYs
This is the number of “BUY” recommendations from the Company in the simulation starting from January 1, 2016 for the stock with the corresponding persona. This “Number of BUYs” is what is shown in the graph for that particular stock for that specific persona for the simulation conducted.
Number of SELLs
This is the number of “SELL” recommendations from Upturn in the simulation starting from January 1, 2016 for the stock with the corresponding persona for its respective “BUY.”
This is the average number of days for the “Hold” period for the Company recommendations for the stock with the corresponding persona (Long, or Quarterly or Monthly) considering simulation starting from January 1, 2016. A "hold" period is the time duration when a stock goes from a "BUY" recommendation to a "SELL" recommendation. For the Avg-Invested Days, this also considers all the available “hold” periods along with any “hold” which is unrealized i.e., also considers the days where for a specific “BUY” there has not yet been a “SELL” and divides the entire number of days with “number of BUYs”.
This is for a “SELL” recommendation in the simulation of Upturn recommendations starting from January 1, 2016 for the stock which ended in a profit (i.e., positive return). Note: Refer to Disclosures for Performance Metrics and Simulations which goes into assumptions made in simulation.
This is for a “SELL” recommendation in the simulation of the Company’s recommendations starting from January 1, 2016 for the stock which ended in a loss or negative returns. Note: Refer to Disclosures for Performance Metrics and Simulations which goes into assumptions made in simulation
Portfolio (and Portfolio performance simulation charts and ratios)
Following are terms, definitions and assumptions used when showing the Company’s historic recommendations of a portfolio.
Here a portfolio is defined as a collection of stocks by a user to closely follow and trade. The user can choose similar persona in the portfolio or mix and match each stock with different personas within a portfolio and a portfolio is not restricted to a specific persona alone.
Note: Please review the disclosures for assumptions and different considerations in a simulation or hypothetical returns and ratios. Keeping those assumptions into consideration, this section describes different performance metrics and terms used.
The portfolio simulations are built by having each stocks equally weighted in to the portfolio (this is implemented by using percentage of daily returns) and uses the simulation of the portfolio starting from January 1, 2016. The simulation adds stock returns (positive in case of profits or negative in case of loss) to the portfolio as if the stocks executed a "BUY" or "SELL" based on the Company’s recommendations.
Note: Please read the Disclosures for Performance Metrics and Simulations
Upturn uses Python Empyrical library (or its close implementation)
https://github.com/quantopian/empyrical for the calculation of key ratios and performance metrics of the Portfolio simulation.
Following are the definitions of the key ratios and metrics used in the simulation
This is simulated compounded returns expressed in percentage of the portfolio. When this is shown in charts, the cumulative returns starts with principal value of 1 (normalized value for charts) on January 1, 2016 and how if executed the "BUY" and "SELL" on each of the stocks (with the assumptions as discussed in Disclosures for Performance Metrics and Simulations section below) performed. The charts also compare these cumulative returns with SPY (S&P SPDR 500 ETF). Compounded returns means, in the simulation, the assumption is, the user gained or lost x% returns on its initial principal invested in each stock of all the stocks in the portfolio, that entire principal plus (or minus) the returns was reinvested in the next set of recommendations. Returns can be negative or positive.
This is the mean annual growth rate of returns of the equally weighted simulated portfolio
This is the risk adjusted return of the equally weighted simulated portfolio
This is the risk adjusted return of the equally weighted simulated portfolio
This is the measure of the equally weighted simulated portfolio’s ability to beat the market or how does it compare to the market (here SPY is used to indicate market)
Alpha is typically measured along with beta which measures the overall market risk of the portfolio (here SPY is used to indicate market)
Upturn Insights give holistic metrics of Upturn recommendations specific to personas. The Upturn Insights provides the total number of stocks (in its pool of stocks under recommendations) which has had a “Buy” recommendation on the date. This chart also has SPY (S&P 500) overlaid for that date which shows how S&P 500 stocks were doing.
As shown with the red box, when the market was slowing down (in terms of S&P 500), the number of stocks which had a “Buy” recommendation was very small for all the personas.
Top Performers page shows in an order all the stocks with their respective persona which has had the highest historic returns which is sum of their profits and unrealised profits based on simulated Upturn recommendations. These values are historic and are considered from Jan 1st 2016.
It only includes stocks with Upturn Star rating of 4 (with their personas) and above and only up to 300 stocks.
The same stock can appear twice based on top performance specific to their personas.
Rush Persona for AAPL (Apple Inc)
“Rush” persona, is like monthly persona but with much smaller "BUY" and "SELL" cycles averaging to 2-4 weeks. Some individual "BUY" and "SELL" cycles maybe longer but on an average is expected to be few weeks. In the RUSH persona for AAPL as shown, average invested-days was 27 days (the 27 days was a calculation of the total ‘hold’ time during the simulation which involved 12 ‘BUY’ and ‘SELL’ cycles and one ‘BUY’ without a ‘SELL’ yet divided by then number of the cycles). Some recommendations where over 1 month, some lesser but it is few weeks as an average.
 Please refer to “Disclosures for Performance Metrics and Simulations
Disclosures for Performance Metrics and Simulations Performance Metrics
This section discusses on the assumptions and details on how Performance Metrics are calculated
Broad Assumptions of all Performance Metrics and simulations for all stocks and portfolios
All performance metrics are not specific to an individual user but are simulation (hypothetical returns) of the historic recommendations with the following facts and general assumptions for both stocks and portfolios:
All historic data for stocks is obtained from Xignite Inc.
All performance evaluation, which includes key metrics defined above (cumulative returns, Annualized returns, annual volatility, Sharpe ratio, etc) considers the "BUY" price to be the price at the "Close" of the market on the previous day or the last close price of the stock. However, the “BUY” price for the stock may not be available the next day if the “open price” of the stock has gone up or down. Further, even if market opens for the stock at the mentioned “BUY” price, such price may not be available for a large volume of “BUY” for a user for the next day.
Similarly, all performance evaluation which includes key metrics defined above (cumulative returns, Annualized returns, annual volatility, Sharpe ratio, etc) considers the "SELL" price to be the price of "Close" on the previous day or the last close price of the stock. However, that price might not be available for Selling the stock next day as on the next day the “open price” of the stock might have gone up or down. Further, even if market opens at the mentioned price for “SELL”, that price might not be available for a large volume of “SELL” for a user for the next day.
The term “returns” used in describing metrics may be either a profit or loss
The price used for the stock for metric calculation in the simulation assumes that price is available for either “BUYing” or “SELLing” of a stock and doesn’t take into consideration large volume of transactions for such stock
The simulated performance metric uses January 1, 2016 as the starting date to capture all metrics including recommendation starts both for the stock (the stock chart and returns calculation and for star rating considerations for the stock) and the portfolio evaluation using the above mentioned metrics for the simulated portfolio ratio and returns consideration
This date January 1, 2016, which is considered for all metrics and returns is a date chosen to show at least a few years of historic consideration for the simulation.
All ratios and performance metrics for stocks (such as compounded return, average profit, avg-invested-days, unrealized profits, etc) and portfolios (such as cumulative returns, annualized returns, annual volatility, sharpe ratio, etc) do not take into account taxes for short term or long term gains, cost of buying and selling stocks (the deduction of brokerage or other commissions, and any other expenses that a client would have paid or actually paid) or dividends paid (or reinvestment of the dividends and other earnings) in between which would mean the actual returns would be much lesser or more (in consideration of dividends) than shown in simulation.
The performance metrics of the simulation is to be taken as only directionally representative of past performance and not represent actual trading returns or future performance
Specific to portfolios, if a portfolio is built with 10 stocks chosen, either through a custom user selection of stocks or a stock screener (a stock search tool in Upturn app to search and filter stocks based on Upturn Star rating, cumulative returns, stock price, etc), for all its calculations, for the simulation it is considered the portfolio is equally weighted. That is, all stocks are invested with equal capital. For simplicity of explanation, consider two stocks A and B in the portfolio, if price of stock A is 100 and price of stock B is 10 to start with, equal weighted would mean for every 1 stock of A, 10 stocks of B are bought to start with, and for simplicity this is implemented by taking percentage of daily returns (changes in percentage of every stock movement on a daily basis) divided by the number of stocks in the portfolio
Specific to portfolios, similarly as above, SPY (SPDR S&P 500 ETF Trust) is taken as the benchmark to compare and the percentage of daily returns (changes in percentage of stock price movement on a daily basis) of SPY is taken into consideration
As noted above, users might not build a portfolio which is equally weighted and for a given principal it might not be possible to create an equally weighted portfolio (stock prices could be different with few stocks in hundreds of dollars and few in tens and others in thousands) and all the metrics of the simulation considers an equally weighted portfolio for the sake of simplicity.
The portfolio simulations considers executing on the “BUY” and “SELL” recommendation by Upturn on the stocks in the portfolio for its respective personas. The “BUY” price and “SELL” price considered for the simulations follow the above assumptions mentioned. These prices for “BUY” and “SELL” might not be available the next day for a user to actually “BUY” or “SELL” the stock at.