AlphaDroid Logo
   Portfolio       Bonds Aye      Style Box Fox       Factor Faves      Global Domination      Sector Nectar   
  Print Page  

Artificial Intelligence Portfolio
Merlyn's Magic - 100:0

Merlyn.AI
 Unleash the Wizard


Portfolio Overview

Aggressive AI - Asset Class Weights:  100% Stocks, 0% Bonds

Conceptually modeled after the classic Aggressive risk portfolio's 100:0 allocation split with an important difference: This portfolio is composed of four divergent underlying momentum strategies, each of which selects one ETF for the portfolio to hold. The names and allocation weights of the four underlying strategies are:

  Sector Nectar: 45%           Global Domination: 25%             Factor Faves: 15%             Style Box Fox: 15%

Merlyn.AI Improves and Simplifies Investing
Wikipedia defines AI (Artificial intelligence) as the ability of a machine to perceive its environment and take actions to maximize its chance of success at some goal. For investors, that usually means higher returns, lower risk, and less effort.

   •  Merlyn.AI intelligently analyzes and automatically selects from hundreds of ETFs on a monthly basis.
   •  Merlyn.AI employs genetic algorithms to continuously evolve strategies as the markets and ETFs change.
   •  Merlyn.AI adaptively tunes itself to help ensure your trades will be neither too early nor too late to the party.
   •  Merlyn.AI integrates StormGuard-Armor and Bear Market Strategies to avoid a large loss in the next market crash.

 
Merlyn.AI Employs Three Forms of AI

Merlyn.AI incorporates: (1) "adaptive algorithms" to adaptively tune the filters and thresholds employed to determine subsequent buy/sell decisions (Forward-Walk Progressive Tuning); (2) "Fuzzy Logic" to evaluate a composite of 12 measures of the market's character to determine current investment safety (StormGuard-Armor); and (3) "genetic algorithms" to evolve the set of candidate funds in a population of momentum strategies to remove the possibility of hindsight selection bias.  Merlyn.AI Strategies utilize these adaptive and genetic algorithms to both identify suitable candidates and to determine which of them has the best probability of doing well next month. In the  Portfolio performance chart below, particularly note the Two-Year rolling chart. The underlying Strategy charts appear later.

 Model Hypothetical Performance

Prudent NPF Portfolio

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA

Active Risk Reduction

Risk is Not a One-Dimensional Problem Cured by a Single Dose of Diversification.  The passive risk reduction provided by Modern Portfolio Theory's 1952 buy-and-hold diversification method was only just the beginning. Today, momentum in market data is well-established, and the cross-disciplinary sciences of differential signal processing and matched filter theory facilitate multiple forms of active risk reduction that forever change the game. They include; (1) risk avoidance as a byproduct of its tactical momentum algorithms designed to identify the trend leader from among a set of candidate ETFs, thus inherently avoiding laggards; and (2) market crash risk avoidance through determining the market is no longer safe (three-factor test of StormGuard-Armor) and automatically switching to an integrated Bear Market Strategy that selects the best-performing ETF from among multiple safe harbor investment candidates.


Don't Just Shuffle Risk  --  Actually Reduce Risk

Break Through the Efficient Frontier

It's only by owning the trend leader and avoiding the laggards that one can simultaneously improve returns and reduce risk. While buy-and-hold performance under Modern Portfolio Theory is bounded by the Efficient Frontier (gray curve), Merlyn's Magic 100:0 Portfolio and underlying Strategies break through the Efficient Frontier into the upper left portion of the risk-return chart below. They are plotted alongside the Industry's set of Consensus Asset Allocation Portfolios  1  through  5  detailed in our white paper Satisfying the Prudent Man. Merlyn's Magic Portfolio contains four underlying strategies, each of which represents a different asset class focus and allocation consistent with the portfolio's target definition. Each underlying strategy examines its set of candidate ETFs and determines which one, and only one, is its trend leader and should represent it in the overall Portfolio.

By combining a set of generally divergent Strategies in a portfolio, the return is the weighted average of the underlying Strategies, but the risk is further reduced by the partial uncorrelated character of the divergently focused Strategies.

Merlyn's Magic 100:0 Portfolio

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA

Performance Drivers 

StormGuard−Armor:  Triple Crash Protection
Better market-crash protection is really about determining when the market is no longer safe, which is a very different problem from simply measuring its direction.  StormGuard−Armor achieves superior performance by combining price-trend, high/low sentiment, and intuitional momentum measures using advanced fuzzy logic principles to determine when it's time to exit the market. When institutional and value investors are nervously standing at the door ready to run, it's more likely world events will trigger a move lower than higher.  Safety is job one.  

Integrated Bear Market Strategy:  Downside Profit Potential
When StormGuard−Armor indicates the market is no longer safe, a Bear Market Strategy automatically takes charge and selects the trend leader from a set of safe harbor candidates such as money market funds, bond funds, gold funds, and U.S. treasuries. The performance improvement provided by a Bear Market Strategy is illustrated in the chart (right). Unfortunately there has been no single reliable safe harbor asset class (other than cash) over the last century. However, between cash, treasuries, bonds, and gold, an integrated Bear Market Strategy is likely to find that one of them can provide a degree of profit potential during bear markets.

Polymorphic Momentum:  Our Performance Pedigree

MPT reliably achieves average returns with reduced risk -- but doing better requires changing the game. Momentum in market data supplements MPT's long-term statistical measures by indicating how to allocate more assets toward trend leaders and away from trend laggards. Extracting the momentum signal from noisy market data is the whole game. We credit these astute agents of change:

Narasimhan Jegadeesh, Emory University Professor of Finance, and Sheridan Titman, University of Texas, Austin Professor of Finance published "Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency" in 1993, which is widely considered the seminal academic paper proving momentum exists in stock market data.

Eugene Fama, Nobel Laureate, University of Chicago Professor of Finance and Kenneth French, Dartmouth College Distinguished Professor of Finance confirmed in their 2008 academic paper "Dissecting Anomalies" that momentum truly is "the premier market anomaly" that is "above suspicion." It's a very powerful statement considering Eugene Fama developed the Efficient Market Hypothesis and now acknowledges momentum is an exception.

Claude Shannon, Bell Labs, National Medal of Science, Kyoto Prize, Medal of Honor-IEEE, Alfred Noble Prize. In his legendary papers of 1948 "A Mathematical Theory of Communication," and 1949 "Communication in the presence of Noise" Shannon proved that signal-to-noise ratio determines a system's probability for making a good decision. This fundamental electronics communications principle is also the primary contributor to better strategy performance.  

J. H. Van Vleck, Nobel Laureate, National Medal of Science, Harvard Dean of Engineering, showed in his 1946 paper "A Theoretical Comparison of the Visual, Aural, or Meter Reception of Pulsed Signals in the Presence of Noise" that a matched filter design produces the optimum signal-to-noise ratio when extracting a signal from a noisy source. It turns out that the optimum momentum filter is quite different from any of the SMA filters academic researchers use.

Samuel H. Christie, Royal Society Fellow, first described his "diamond method" of measuring in his 1833 paper "The magneto-electric conductivity of various metals." Its value was later recognized by Sir Charles Wheatstone and become known as the Wheatstone bridge. Differential signal processing reduces common mode system noise. While seemingly arcane, the strategy design implications are profound for the decision framework and its candidate funds.

These signal processing innovations have become the technology backbone that has enabled Ethernet, Wi-Fi, and smart phones to perform so well. AlphaDroid's algorithms incorporate this same technology to better extract the momentum signal from noisy market data to improve the probability of making a better investment decision.  Additionally, our award-winning Automated Polymorphic Momentum algorithm was developed to automatically adapt the momentum filter's shape and duration to (1) the character of the ever changing market, and (2) the character of the strategy's candidate funds -- bonds, treasuries, sectors, countries, and commodities are clearly quite different from one another.

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA

Merlyn  Strategies  (Note: The Bonds Aye Strategy is not used in the aggressive Merlyn's Magic Portfolio)


Sector Nectar Investment Summary
This Strategy seeks to provide investment results superior to the Standard and Poor's 500 Index (S&P 500) by utilizing Artificial Intelligence algorithms to identify and select one market sector ETF each month that exhibits momentum and volatility characteristics indicative of outperforming its peers in the subsequent month. The ETF selection is made from a field of over 145 sector and index ETFs, including these fund classes: materials, energy, financial, industrial, staples, discretionary, healthcare, bio-pharma, technology, semiconductors, factors, and style box indexes. The Strategy invests 100% of its assets in the selected ETF. The Strategy is not broadly diversified.     

Prudent Momentum Portfolio


Global Domination Investment Summary
This Strategy seeks to provide investment results superior to the FTSE Global All Cap Index by utilizing Artificial Intelligence algorithms to identify and select one market index ETF each month that exhibits momentum and volatility characteristics indicative of outperforming its peers in the subsequent month. The ETF selection is made from a field of over 100 global market index ETFs, including these market index fund classes: World, Europe, EAFE, Asia Pacific, Emerging Markets, International Dividends, and North America. The Strategy invests 100% of its assets in the selected ETF. The Strategy is not broadly diversified.

Prudent Momentum Portfolio

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA

 


Factor Faves Investment Summary
This Strategy seeks to provide investment results superior to the Standard and Poor's 500 Index (S&P 500) by utilizing Artificial Intelligence algorithms to identify and select one market index ETF each month that exhibits momentum and volatility characteristics indicative of outperforming its peers in the subsequent month. The ETF selection is made from a field of over 135 market index ETFs, including these market index fund classes: value, growth, size, dividends, earnings, quality, and global. The Strategy invests 100% of its assets in the selected ETF. The Strategy is not broadly diversified.        

Prudent Momentum Portfolio


Style Box Fox Investment Summary
This Strategy seeks to provide investment results superior to the Standard and Poor's 500 Index (S&P 500) by utilizing Artificial Intelligence algorithms to identify and select one market index ETF each month that exhibits momentum and volatility characteristics indicative of outperforming its peers in the subsequent month. The ETF selection is made from a field of over 90 market index ETFs, including these market index fund classes: value, growth, small-cap, mid-cap, large-cap, equal weight, and other combinations thereof. The Strategy invests 100% of its assets in the selected ETF. The Strategy is not broadly diversified.    

Prudent Momentum Portfolio
 

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA

 


Bonds Aye Investment Summary
This Strategy seeks to provide investment results superior to the Bloomberg Barclays US Aggregate Bond Index (AGG) by utilizing Artificial Intelligence algorithms to identify and select one fixed income ETF each month that exhibits momentum and volatility characteristics indicative of outperforming its peers in the subsequent month. The ETF selection is made from a field of over 70 fixed income ETFs, including these bond fund classes: aggregate, short-term, mortgage, municipal, corporate, junk, treasury, and TIPS.  The Strategy invests 100% of its assets in the selected ETF. The Strategy is not broadly diversified.  

Prudent Momentum Portfolio

 

 

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA

 

Modeling Information

Riskalyze Risk Number fo AlphaDroidTactical Model Risk  vs.  Current Holding Risk
The Riskalyze Risk Numbers in this document reflect hypothetical long-term portfolio performance. However, some AutoPilot web pages display a short-term portfolio Risk Number that assumes (incorrectly) that current ETFs will be held for another 6 months ... without the benefit of trades that tactically avoid trend laggards and bear markets. Fortunately, some pages have an option to display the Risk Number in SMA mode (separately managed account) that includes the full portfolio history with the benefits of tactical momentum and which result in the same lower Risk Number found here. The above example values are for the Prudent Momentum 60:40 Portfolio in May of 2017 when it held MDYG, IJJ, JNK, and HYG.

Forward-Walk, Backtesting, and Inception

AlphaDroid employs Forward-Walk Progressive Tuning — the industry gold standard for performance modeling. A first period of time is used to tune the parameters of the algorithm for use in making decisions during a subsequent period of time. The first period is called backtesting and is inherently flawed because tuning is done in hindsight of the data. The subsequent period is called forward-walk testing because it applies the tuning to an "out-of-sample" set of data as the algorithm walks forward in time, thus denying random events of any special tuning advantage. As the algorithm moves forward through the out-of-sample data it periodically pauses to re-tune itself using only past data so that it can adapt to changes in market's character or adjust to the character of new funds participating in the strategy.  Each performance chart is thus a composite of performance estimates, including (a) an initial tuning period determined by backtesting, (b) a subsequent period of Forward-Walk Progressive Tuning, and (c) a final period starting at the strategy's Inception Date (when its design was complete) and then walking forward in real time from there.
 

Trade Signals, Data Source, and Performance Modeling

Underlying strategies typically produce 3 to 5 month-end trades per year, as new trend leaders emerges.  Trade signals are generated following the close of the last trading day of the month. Performance modeling assumes trades are executed at the close of the subsequent market day. End-of-day market data is provided by FastTrack and is back-adjusted for splits, dividends, and capital gains distributions. The charts in this document are updated daily. Portfolio Benchmarks are derived from Fidelity Asset Manager mutual funds to represent corresponding classic asset allocation portfolios. Hypothetical model performance does not include the costs of trading fees, portfolio management fees, account management fees, or financial advisor fees.

About Us

We Believe:  High-performance investment software should be for everyone, not just big Wall Street firms. SumGrowth Strategies, LLC was founded in 2009 to make simplified high performance momentum algorithms available to independent investors through SectorSurfer and later to independent advisors through AlphaDroid. It began in 1992 when founder Scott Juds first performed a few spreadsheet experiments to determine if trends existed in market data as he was searching for a better way to invest his IRA funds. As implied by the Performance Drivers section of this document, creating a comprehensive solution turned out to be a complex, multi-faceted problem resulting in years of incremental developmental improvements, typical of all impressive modern technology. Our mission is to develop and market high-performance investment algorithms, as tools and models, to support the needs of investment advisors and individual investors alike. Meet our team members HERE and see our Innovation Timeline HERE.

Disclosures and Disclaimers

The Company: SumGrowth Strategies, LLC of Seattle, Washington (the Company) is not a registered investment advisor, does not provide professional financial investment advice specific to anyone's life situation, makes no custom strategies or portfolios for anyone, and has no fiduciary relationship with its portfolio model subscribers. The Company develops and markets high-performance investment tools and models. 

Our Models: Merlyn's Magic Portfolio Models employ the methods and algorithms described in the Performance Drivers section of this document in combination with funds having asset class allocations as specified by the Model. These Models are published online daily by the Company for use by its subscribing financial advisor clients. Historical performance is generated by applying the current models to the specified historical period using exchange-provided price data, not from actual traded accounts. Thus, trading performance displayed is hypothetical, only represents historical conditions of the market, and does not include the costs of trading fees, portfolio management fees, account management fees, or financial advisor fees. There is no guarantee that such performance will be achieved in the future and there is no representation being made that any account will or is likely to achieve profits or losses similar to those shown. Investing in securities involves risk of loss that clients should be prepared to bear. An investment's objective, risks, charges, and expenses must be carefully considered before investing.

The Data:  Market data occasionally contains errors or inaccuracies and may also be changed or updated without notice. The Merlyn's Magic Portfolio Models are provided "as is" without warranty of any kind. SumGrowth Strategies, LLC, its affiliates and employees make no representation or warranty, expressed or implied as to the suitability, effectiveness, accuracy, availability or completeness of its investment models, and specifically disclaim all other warranties, expressed or implied, including but not limited to implied warranties or fitness for any particular purpose. Neither SumGrowth Strategies, LLC, nor any of its affiliates or employees shall be liable for any direct, indirect, incidental, special, punitive, or consequential damages that result in any way from use, non-use, reliance upon the information, or that may result from mistakes, omissions, interruptions, deletions of files, defects in market data, operational delays, transmission delays, failure of equipment, or failure of performance. The sole and exclusive remedy for dissatisfaction with any information or service provided by the Company is to discontinue using said information or service.

  www.AlphaDroid.com     Merlyn@AlphaDroid.com     SumGrowth Strategies, LLC  Seattle, WA