PortfolioBuilder^™ 1.0

Introduction:

PortfolioBuilder™ is a proprietary portfolio analysis tool custom-built specifically to meet the stringent functional requirements specified by one of customers, Madison, Monroe, & Whitaker Investment Services, LLC.  Its primary purpose is to build and analyze portfolio performance, on a day-by-day basis, by combining various markets traded by various automated-systems.

Trading Systems Technologies, Inc. was asked to build this analytical tool because this customer felt that the standard techniques used to perform historical back testing most often provided a poor representation of what actual performance may look like.  Consequently, PortfolioBuilder^™ uses considerably more stringent methods and expanded measures in order to provide a more realistic (less optimistic) review of past performance.  But remember, no matter how stringent the methods used to analyze past performance are, past performance is never a guarantee of future results.

Please review each of these sections for more detailed information as well as examples and definitions of the given measures.

What's Different About PortfolioBuilder™?

There are several key differences which set PortfolioBuilder™ above the rest of other portfolio and system analysis packages.  These include the following:

Day-By-Day Analysis:

Other portfolio and system analytical packages will simply aggregate data on a trade-by-trade, month-by-month, or year-by-year bases.  This did not meet the stringent requirements specified by MMW.  Here's why. 

Assume  investor-A begins trading a particular portfolio on [say] March 15th and that investor-B begins trading the same portfolio March 16th.  The fact is that these two individuals could have significantly different results with the same portfolio.  Why?  Well, if a trade-signal is generated on March 15th, and that trade lasts [say] four month, then investor (A) would have a four month trade on the books that investor (B) would have missed.

This type of scenario can impact every facet of performance and this is why many of the metrics generated by PortfolioBuilder™ are done so on a day-by-day basis.  This type of analysis requires an iterative approach where PortfolioBuilder™ actually begins a calculation on day-one and processes every day forward, then begins the calculation again on day-two and processes every day forward, and repeats for every day of every year of given history.

Worst Initial Loss:

Worst Initial Loss is arguably the most important metric provided by PortfolioBuilder™and, as far as we know, is not readily available on reports generated by any other portfolio or system analytical tool.

Worst Initial Loss is the measure of the worst "dip", or "drawdown" in an account's equity, below the initial-funding, if an investor began trading a given portfolio on any given day within the provided history. 

Time Window Profitability:

Profitability windows look at a portfolio for a given time-segment (the time-window) and simply asks the questions "given this historical data, if I began trading on day-one of this time-segment and stopped trading on the last day of this time-segment, would the account balance be up, or down?".  The analytical process then moves to the next day in the historical data and begins again.  While incrementing the window forward day-by-day PortfolioBuilder™ keeps track of the number of profitable vs. loosing windows and derives the percentage of profitable windows.  Hence, a "12 Month Window Profitability" of 96% implies that 96% of the time the given portfolio was traded for a 12 month period [within the given historical market-data], starting on any day in the given history, the results were profitable.

Inclusion of Market Roles:

We have found that other analytical tools do not include market-roles in their calculations.  If an analytical tool is analyzing [say] 25 years of results for a portfolio of [say] ten markets traded by a system that is "in the market" [say] 90% of the time, there would be thousands of roles (and corresponding commissions and slippage).  These can amount to a significant portion of the profit (loss) and ROI [Return On Investment].

(For those of you new to commodities trading, a "role" is when a contract's expiration date approaches, in order to avoid taking delivery of the commodity, the investor must liquidate their current position and re-enter the market by assuming a contract with a later expiration date.)

Maximum Volatility:

This is another measure that other analytical tools do not calculate.  This metric is a calculation of the largest percentage drop in the accounts balance (second to "Worst Initial Loss") if an investor began trading on the "Start Date" shown (Worst Initial Loss is the largest percentage drawdown using the day-by-day method).  Maximum Volatility is calculated as follows:

current drawdown / initial starting capital + highest profit to-date

This figure is calculated for each "dip" in the account's equity.  The largest percentage dip is what is shown in the report.

Note: The largest "dip", or "drawdown", is not necessarily the largest percentage drawdown.  A $5,000 drawdown in an account of $20,000 is a much larger percentage then a $30,000 drawdown in an account of $300,000.  This is why PortfolioBuilder shows both Maximum Volatility as well as Maximum Drawdown.  Please also note that the incident of Maximum Volatility typically occurs when the account's balance is rather small.  Why?  Because systems typically risk a predetermined dollar-amount per trade despite the account's size.  Thus, a string of losses of x-dollars is a larger percentage of an account's balance when the account is small than the same string of losses when the account's balance is large.

Example Report & Metrics Definitions

Click on Items of Interest

(Click this image to see full report)

Definitions of Metrics:

Gross Profit ACS:

Profits associated with only the winning trades less commissions and slippage.

Back to Top  >

Gross Loss ACS:

Losses associated with only the loosing trades plus commissions and slippage.

Back to Top  >

Net Profit ACS:

Calculated as follows: Gross Profits ACS - Gross Losses ACS

Back to Top  >

Total # of Trades:

Total number of trades taken within this portfolio over the stated time period starting at the "Start Date".

Back to Top  >

Total # of Winning Trades:

Total number of winning trades taken within this portfolio over the stated time period starting at the "Start Date".

Back to Top  >

Total # of Loosing Trades:

Total number of loosing trades taken within this portfolio over the stated time period starting at the "Start Date".

Back to Top  >

Percentage Profitable:

Calculated as follows: Total # of Winning Trades / Total # of Trades

Back to Top  >

Estimated # of Roles:

Calculated on a market-by-market basis based upon each system-markets percent-time-in-market, the number of roles that market experiences per year, and the total number of years the portfolio is traded.  This number for each system-market is then aggregated to derive the total "Estimated # of Roles".

Back to Top  >

Average Winning Trade ACS:

Calculated as follows: Gross Profits ACS / Total # of Winning Trades

Back to Top  >

Average Loosing Trade ACS:

Calculated as follows: Gross Loss ACS / Total # of Loosing Trades

Back to Top  >

Average Trade Profit (Loss) ACS:

Calculated as follows: Net Profit ACS / Total # of Trades

Back to Top  >

Average Annual ROI ACS:

Calculated as follows:
(Net Profit ACS / Suggested Starting Capital) / Trading Period (Years)

Back to Top  >

Worst Initial Loss (all dates):

Worst Initial Loss is arguably the most important metric provided by PortfolioBuilder™and, as far as we know, is not readily available on reports generated by any other portfolio or system analytical tool.

Worst Initial Loss is the measure of the worst "dip", or "drawdown" in an account's equity, below the initial-funding, if an investor began trading a given portfolio on any given day within the provided history. 

(Also see FAQ "What's the difference between 'Worst Initial Loss' and 'Maximum Drawdown'")

Back to Top  >

Worst Starting Date:

The beginning date of the first trade in the series of trades which resulted in the "Worst Initial Loss".  If an investor had taken trades either before or after this particular trade (within the given history) this would result in a smaller "Worst Initial Loss".

Back to Top  >

Maximum Volatility:

This is another measure that other analytical tools do not calculate.  This metric is a calculation of the largest percentage drop in the accounts balance (second to "Worst Initial Loss") if an investor began trading on the "Start Date" shown (Worst Initial Loss is the largest percentage drawdown using the day-by-day method).  Maximum Volatility is calculated as follows:

current drawdown / initial starting capital + highest profit to-date

This figure is calculated for each "dip" in the account's equity.  The largest percentage dip is what is shown in the report.

Note: The largest "dip", or "drawdown", is not necessarily the largest percentage drawdown.  A $5,000 drawdown in an account of $20,000 is a much larger percentage then a $30,000 drawdown in an account of $300,000.  This is why PortfolioBuilder shows both Maximum Volatility as well as Maximum Drawdown.  Please also note that the incident of Maximum Volatility typically occurs when the account's balance is rather small.  Why?  Because systems typically risk a predetermined dollar-amount per trade despite the account's size.  Thus, a string of losses of x-dollars is a larger percentage of an account's balance when the account is small than the same string of losses when the account's balance is large.

Back to Top  >

Drawdown of Max Volatility:

The amount the account's equity "dipped" during this largest-percentage-dip.

Back to Top  >

Average Volatility High-to-High:

The average percentage decline in the account's equity from the time one high is made to the time the next high made.

Back to Top  >

Maximum Drawdown:

The largest "dip" in the accounts equity that would have been realized trading this portfolio over the given "Trading Period (Years)" starting at the "Start Date".

(Also see FAQ "What's the difference between 'Worst Initial Loss' and 'Maximum Drawdown'")

Back to Top  >

Average Drawdown High-to-High:

The average decline in the account's equity from the time one high is made to the time the next high made.

Back to Top  >

Margin Required:

The aggregate of each market's margin traded in this portfolio.  If a single market trades multiple simultaneous positions within the portfolio, this is accounted for as well.

Please Note: Margins may be changed by exchanges without notice.

Back to Top  >

Suggested Starting Capital:

Calculated as follows: (1.5 x Worst Initial Loss) + Margin Required

This would permit the investor to sit through one-and-a-half times the "Worst Initial Loss" and still have enough equity to continue trading all markets within the given portfolio.

Back to Top  >

Return on Starting Capital:

Calculated as follows: Net Profit ACS / Suggested Starting Capital

Back to Top  >

Largest 30 Day Profit:

By taking a 30 calendar-day window (not trading-day window) and incrementing it forward day-by-day, PortfolioBuilder™ derives the "Largest 30 Day Profit" and the "Largest 30 Day Drawdown".  These are the numbers shown here.

Back to Top  >

Largest 30 Day Drawdown:

By taking a 30 calendar-day window (not trading-day window) and incrementing it forward day-by-day, PortfolioBuilder™ derives the "Largest 30 Day Profit" and the "Largest 30 Day Drawdown".  These are the numbers shown here.

Back to Top  >

Average 30 Profit (Loss) ACS:

Calculated as follows: Net Profit ACS / (Trading Periods (Years) / 12)

Back to Top  >

n Month Window Profitability:

Profitability windows look at a portfolio for a given time-segment (the time-window) and simply asks the questions "given this historical data, if I began trading on day-one of this time-segment and stopped trading on the last day of this time-segment, would the account balance be up, or down?".  The analytical process then moves to the next day in the historical data and begins again.  While incrementing the window forward day-by-day PortfolioBuilder™keeps track of the number of profitable vs. loosing windows and derives the percentage of profitable windows.  Hence, a "12 Month Window Profitability" of 96% implies that 96% of the time the given portfolio was traded for a 12 month period [within the given historical market-data], starting on any day in the given history, the results were profitable.

Back to Top  >

Start Date:

The date of the first trade taken by this portfolio.

Back to Top  >

Trading Periods (Years):

The number of years of historical market-data used to analyze this portfolio's performance (the decimal places show one-hundredths of a year).

Back to Top  >

Total Number of System-Markets: