Most of the trading systems I have written about have been very similar. Each of the trend following systems attempt to capture big chunks of trends in similar ways. The mean reversion systems I have profiled each offer slightly different ways to execute the same basic mean reversion strategy. While each of these systems offer subtle differences in their approach, the general strategy is usually quite similar.

Of all the systems that I have looked at, the biggest outlier was George Vrba’s Best10 Portfolio Management System. This system wasn’t focused on trend following or mean reversion. It was simply trying to improve on a buy and hold approach to the general market. Because it was so different, this system has stuck out in the back of my mind as something I would love to explore further.

In my research and writing, I generally focus on very simple systems. The reason for this is that if a system is simple enough that my mother can understand the logic behind it, it may convince her to switch from her current buy and hope strategy. I believe that there is an huge market of investors, like my mother, who have no desire to trade for a living, but would love to have a simple way to steadily beat the general market.

The Ivy Portfolio

Last December, Jeff Swanson from System Trader Success wrote about The Ivy Portfolio, which is similar to Vrba’s Best10 System. Swanson’s work was based on a book written by Mebane Faber and Eric Richardson, who studied how Ivy League schools are able to achieve steady and significant returns on their endowment funds. Using what he learned from the book, Swanson built a similar system that would attempt to replicate how those schools are trading.

The concept of Swanson’s system is remarkably simple. He is taking a basket of 5 or 10 ETFs that represent a broad cross section of the market and investing in the ones with the highest relative strength. He formed a simple algorithm to calculate the relative strength of each ETF and then invests in the top three ETFs. He then calculates the relative strength and adjusts the portfolio each month. He also uses the 100 day simple moving average (SMA) as a trend filter to make sure that he is always trading with the trend.

The first step of the system is to rank each of the ETFs in terms of relative strength. Swanson does this by calculating the 20 day return and the three month return. This gives both shorter and longer term perspectives on each of the ETFs. He then weights each of the returns as half of the overall rank. Here is what his formula looks like:

Overall Rank = (20 Day Return * 0.5) + (3 Month Return * 0.5)

Each month, Swanson performs this calculation on each of the ETFs his system trades and then excludes any ETFs that are trading below their 100 Day SMA. He then adjusts his positions by selling any holding that does not rank in the top three positions. He then establishes a position in each of the top three ETFs, provided he does not already have a position in them. Each position accounts for 1/3 of the account equity.

Swanson proposes two different versions of this system. His Ivy Five system trades the following ETFs:

• BND – Vanguard Total Bond Market
• DBC – Powershares DB Commodity Index
• VEU – Vanguard FTSE All-World ex-US
• VNQ – Vanguard MSCI US REIT
• VTI – Vanguard MSCI Total US Stock Market

He also proposed a bigger version of this system that trades these ten ETFs:

• BND – Vanguard Total Bond Market
• DBC – PowerShares DB Commodity Index
• GSG – iShares S&P Commodity-Indexed Trust
• RWX – SPDR DJ International Real Estate
• TIP – iShares Barclays TIPS
• VB – Vanguard MSCI US Small Cap
• VEU – Vanguard FTSE All World ex-US
• VNQ – Vanguard MSCI US REIT
• VTI – Vanguard MSCI Total US Stock Market
• VWO – Vanguard MSCI Emerging Markets

Backtesting Results

Swanson was able to backtest both systems from the middle of 2003 through the end of 2010. During that time, both versions outperformed the S&P 500 by a substantial amount with lower drawdowns. Being able to diversify away from equities and even stay completely out of the market at times gave these systems a tremendous advantage when the S&P 500 crashed in 2008.

Over the course of the backtesting period, the five ETF version of the system averaged an 11.8% annual return compared to only 7% for the S&P 500. The system had a maximum drawdown of 21.3% compared to 55.2% on the S&P 500. It also had a Sharpe Ratio of 0.72 compared to 0.29 on the S&P 500. As you can see, the Ivy Five System significantly outperformed a buy and hold approach with less than half the drawdown.

Since it had more options for diversification, the Ivy Ten System performed even better over the same time period. It averaged an annual return of 14.7%, had a maximum drawdown of -28.7%, and a Sharpe Ratio of 0.82. While the drawdown was a bit higher than the Ivy Five System, it was still way less than the S&P 500, and the overall return was better than the Ivy Five System.

System Analysis

The returns produced by the Ivy Systems are not as spectacular as the Best10 Returns were, but I would argue that the Ivy Systems are far more applicable for a part time trader. The systems also involve a much smaller universe, simpler calculations, and significantly less risk exposure.

These systems are easy to understand, appear to be profitable, and would be fairly simple to implement. Anyone with a high school math education could perform the required calculations and the process could be made even easier with a simple Excel spreadsheet.

My only reservation with these systems is the downside risk exposure that would exist in the event of a Black Swan market crash. If the bottom were to suddenly fall out of a market, I wouldn’t want the systems to wait until the end of the month to recalibrate and go to a cash position. This could be remedied by setting stop-losses at the 100 day SMA filter for all open positions.

Ranking Calculations Example

In order to demonstrate how to calculate the monthly rankings, I buildta simple Excel spreadsheet and looked up the price data for each of the 10 ETFs. I input the current price, the price from 20 trading days ago, and the price from 3 months ago. I also took a quick look at the chart of each ETF to see whether it was above or below the 100 day SMA line. I put a “Y” into the spreadsheet for each ETF that was above the line and an “N” for each ETF that was below the line. The rest was simple math to calculate the returns. Now that I have the Ivy spreadsheet built, the math will be done automatically from here on out.

Backtesting results of a portfolio with 10 ETFs

As you can see, five of the ETFs are currently above their 100 day SMA lines and the other five are below their 100 day lines. The five that are trading below their 100 day lines are automatically excluded from consideration. Interestingly, they were the bottom five in the overall ranking as well.

The top three ETFs in overall ranking are GSG, DBC, and VB. Therefore, if we were starting or reviewing an Ivy Ten portfolio this weekend, it would place one third of its equity into each of those three ETFs. Then we would repeat the same process next month.