Historical performance results of the TFadapt system for SPY and QQQ are presented. Without any system parameter adjustments, a decrease in drawdown levels and an increase in compound annual returns are obtained as compared to buy and hold. The result is a significant increase in the risk adjusted returns over buy and hold to justify active investment management .
The TFadapt system was introduced in this post and performance examples for several international stock indices were presented in this post. Here performance examples are presented for SPY and QQQ using adjusted data from 01/03/2000 to 12/31/2012.
It should be clear that the test period was not a very good one for buy and hold SPY investors and an even worse one for QQQ investors. It may be seen from the charts below that both ETFs declined sharply after 2000 resulting in very large drawdown levels:
The vertical markers indicate the start of the backtest. The backtest end is at 12/31/2012.
Buy and hold results
Below is a table with the buy and hold results for SPY and QQQ
Table 1. Buy and hold results for SPY and QQQ: 01/03/2000 – 12/31/2012
|Symbol||CAGR||Max DD||MAR ratio|
The results for the TFadapt system as shown in Table 2 below. All tests are for starting capital of $100K and a fully invested equity at each position.
Table 2. TFadapt backtest results for SPY and QQQ: 01/03/2000 – 12/31/2012
|Symbol||CAGR||Max DD||MAR ratio||Win rate||Profit Factor|
For SPY the system generated 7 trades and for QQQ it generated 10 trades. This is a long/short system, i.e. short positiond exit long positions and the other way around. However, the system may not be invested at all times.
Below is a table of monthly results in the case of SPY:
Table 3: Profit table for SPY
Note that 2000 was a losing year because of the choice of the starting date of the backtest the system entered in a long trade near the top of the long uptrend of the 1990s. The only other losing year is 2011 with a -11.1% return. For 2008, the system returned 15.5%.
Below is a table of monthly results in the case of QQQ:
Table 4: Profit table for QQQ
Again, in the QQQ case the first year is negative by virtue of the choice of starting date of the test but although there are more negative years, the losses are much smaller than in the case of SPY and 2011. In 2004 the system had a very small negative return of -0.6% and in 2006 and 2009 negative returns in the vicinity of -3.5%. But unlike the SPY system, it shows positive performance for 2011 and higher returns for 2010 and 2012.
Below are the backtest trade reports for SPY and then QQQ:
The TFadapt system is available for sale.
Disclosure: no relevant position at the time of this post and no plans to initiate any positions within the next 72 hours.
Charts created with AmiBroker