I am debuting a new system that I have developed based on Grossman's GMR posted here. I wanted to make a similar system for myself that was easier to implement. I don't have fancy trading tools like Quantshare -- I just use Yahoo finance for data and Excel for processing.
My variation on the GMR is backtested to 2004 with remarkable results. Returns versus the S&P (dividend adjusted) are:
|Year||My System||S&P (div adjusted)|
I have tested a number of systems but this is the first one that (1) beats the S&P every year for at least 10 years (2) does not exhibit diminishing returns in the later years (3) provides market-beating results in bull markets and (4) is practical and can actually be used.
I am officially going live now. The system sent a cash signal on December 31, 2013. How prescient! January was bad. I'll chalk that up as a win -- not losing is definitely a win when the rest of the market is going down.
For Jan 31, 2014, the system gave a signal for SSO and we bought at 95.17 on Monday morning, first thing. Well, I should've waited. Anyway most of Monday's loss has been recovered today. And the month is young...
Some more stats and features of my system:
- Rotates SSO (27%), FEZ (12%), EEM (6%), ILF (28%), EPP (3%), EDV (19%), and SHY (5%). The percents show the portion of time spent in each position.
- Uses monthly data only. This makes data management and processing very easy on Excel.
- 42% CAGR since 2004. Never a down year, always beat the S&P.
- Maximum drawdown only 16.1% (Apr 2010). Annual standard deviation is 22.5%.
- Even with a horrible-case 1% slippage per trade, CAGR was still 35.8%. This is because the system doesn't trade all that often: 121 trades since 2004. Thus, you do not need to be so careful about your entry and exit points, and market liquidity problems will not kill you if you accidentally forget to click "limit order." In other words, it is easy to implement.
- It relies on 6-month volatility, not 20-day volatility like Grossman's system. There is no need to acquire and calculate daily data. Volatility is also a positive correlation in this system -- higher volatility will increase the ranking. I have struggled to explain why this is, but my best guess is that it marks the beginning of an uptrend, rather than the end of one, as you would see with lower volatility.
- you trade once per month. The signal comes after the close of the last trading day of the month. You can then sell the previous position and buy the new one (if the signal has changed; it doesn't always) at the open on the next trading day. If there is a big gap, you can decide to skip the trade or enter a limit order at the "signal price" or just jump in. I showed above that even a 1% slippage from the "signal" price doesn't do too much harm.
Disclosure: I am long SSO.