Objective research and reviews to aid investing decisions
 Looking for information on debt consolidation loans? Want to know more about reverse mortgages? |
Robert Bendekgey of AlphaKing requested inclusion in the list of gurus at Guru Grades based on the service's market predictions. "The AlphaKing Research Project was started in 1999 to answer once and for all the age old question of what is the best way to risk one's capital in the stock market." This research culminated in the AlphaKing Trading Indicator (based on Nasdaq Composite index behavior), claimed to encapsulate the following conclusion: "Great fundamentals, when used in conjunction with specific targeted chart patterns, along with targeted beta leverage, while using active defensive and offensive money management techniques on individual stocks once entered, while timing portfolio entry and exit based on stock market trends has proven to provide maximum returns with the least amount of volatility..." Because this indicator is explicit and mechanical, we evaluate AlphaKing as a method and not as a guru. Using the record of the AlphaKing Trading Indicator and contemporaneous daily opening levels for the Nasdaq Composite index over the live out-of-sample period 3/27/06 through 3/28/08 (505 trading days), we find that...
As noted in the footer of their web site, AlphaKing's self-reported trading results "are tabulated using the opening price the day following a new trading signal, and exclude commissions, dividends, or interest paid on cash balances during sell periods." We apply the same assumptions here and use a buy-and-hold strategy for the Nasdaq Composite index as a benchmark.
The following chart superimposes the AlphaKing Trading Indicator (+1 = bullish and -1 = bearish) on the Nasdaq Composite index over the sample period. The indicator has signaled just seven trades (3.5 round trips) during this time and is bullish on 58% of trading days. Visual inspection is inconclusive regarding signal effectiveness. The average daily return while bullish (bearish) is 0.036% (-0.035%), compared to 0.006% for all days in the sample, suggesting some timing value. However, seven signals provides hardly any basis for statistical inference. At the implied pace, it would take over five years (a decade) to generate a sample of 20 (40) signals.
For another perspective, we calculate returns over the intervals between bearish-to-bullish and bullish-to-bearish changes in the AlphaKing Trading Indicator.
The following table summarizes the performance of the Nasdaq Composite index from signal change to signal change over the sample period, beginning with a bullish-to-bearish switch at the open on 5/15/06 and ending with a bullish-to-bearish switch at the open on 11/16/07. Results are inconclusive, with the second and third largest increases in the index occurring while bearish. However, as noted, the number of intervals in the sample is too small for useful statistical inference.
For a third perspective, we consider cumulative returns.
The final chart compares cumulative returns for $1.00 initial investments on 3/28/06 in three alternative strategies using the Nasdaq Composite index (assuming no trading friction and no return on cash):
The strategies based on the AlphaKing Trading Indicator sometimes lead and sometimes lag the index, but with less volatility. At the end of the sample period, the trading strategies lead the index. However, as noted above, the number of signal changes in the sample is not large enough for reliable inference about long-run results. Outcomes of such small samples are very sensitive to start and stop dates.
We applied the above AlphaKing Trading Indicator signals also to the S&P 500 index, with much weaker results, indicating that the signals may lose any predictive power they have when applied to other indexes and, therefore, to individual stocks.
The description of the AlphaKing Research Project, which outlines a five-step development approach, suggests the possibility of substantial data snooping bias (luck) in backtested results. Such bias can derive from borrowed results (of unknown data mining intensity) and from direct intensive testing of a large number of parameter alternatives. Note that Step 1 opens with:
"James O'Shaughnessy showed us how to, and the wisdom of, isolating and combining investment elements that had proven to outperform over the long term, with his high relative strength and low price to sales combination achieving a 24% annual return from 1954 to 1996 (when the research was published in What Works on Wall Street.)"
See our blog entry of 7/14/05 to see what happened to James O'Shaughnessy's backtested results when subsequently implemented in real mutual funds.
Note also that financial markets may well be adaptive (see our blog entry of 5/26/05), such that there is no "once and for all" best way to invest.
In summary, testing indicates that the AlphaKing Trading Indicator may have some value for timing the Nasdaq Composite index, but the duration of the live data sample is much too short for reliable inference.
This review tests the general market timing value of the AlphaKing Trading Indicator for the Nasdaq Composite index and not the performance of AlphaKing stock picks or any other aspects of their services.
For reviews of a few other methods (as well as reviews of some books and information web sites), see Blog Synthesis: Reviews of Books and Web Sites.
