Noiseless Return Methodology
Reason
Investing in the stock market is a treacherous process, full of ups and downs and a sea of information that constantly bombards ones senses. Following a buy and sell discipline of a standard technical indicator can take away the emotion of the market, but following such a strategy when the market is in a sideways churn produces very poor results. Trying to poor over fundamental data both at the macro and micro level is a long and tedious process and is always subject to interpretation. There has to be a way to look at a stock's trading history, much the same way as a technical indicator, and use it to extrapolate if the stock will continue to out or under perform.
Hypothesis
Looking at stock performance verses the stock market should give us a way to find out which stocks are currently outperforming and the strength of that out performance. An outperforming stock should continue to outperform, though there should be ways to tell if you are too late to the party. If we position ourselves in the stocks that are most likely to outperform our returns should be higher than the overall market and a dart throwing monkey.
Building the Indicator
The basic indicator of performance verses the market is a simple moment indicator. The timeframe of the indicator can be varied as can the market you test against. The premise is the same: find out how quickly the stock is moving in relation to the stock market. In effect, this takes the random noise produced by the stock market out of the equation.
Great, but now what?
Once you take the noise of the market away, the result is a smooth oscillator that swings based upon if the stock is outperforming or underperforming. There are two options when looking at this data: study the value of the out/under performance itself, or examine the rate of change or the slope of the values. Since we are interested in trying to spot outperformance as quickly as possible, it makes more sense to study the rate of change over the value itself.
Refining the Data?
With a relatively noise free indicator in our hands, can we examine this data further to help refine the process?
Performance Area - A simple measure of the area of the graph that is above zero verses the total area of the graph.
Volotility - The maximum reading of our slope divided by the maximum reading minus the minimum reading.
Swing Time - The number of days since our last swing (slope crossing zero) vs. the average number of days between swings.
The Test
We run our calculation on the stocks in the S&P 500 verses the index itself weighing Performance, Volatility, and Swing Timing differently (See: weighing table below) for four different time periods. At the start of each quarter we would buy the top 12 stocks based on their rating. The stocks would be held onto for exactly one year. The test period is from 1998-Q4 until 2004-Q3, this includes the tail end of the .com bubble and the crash that followed.
Time Periods
| Label | Meaning | Time Frame |
|---|---|---|
| ST | Short Term | 4-12 Weeks |
| MT | Medium Term | 2-6 Months |
| MLT | Medium/Long Term | 4-12 Months |
| LT | Long Term | 9+ Months |
Weighing
| Label | Meaning | Vol | Perf | Swing |
|---|---|---|---|---|
| +EQ | Equal Weight | 1 | 1 | 1 |
| +Vol | Favor Volatility | 3 | 1 | 1 |
| +Perf | Favor Performance | 1 | 3 | 1 |
| +Swing | Favor Swing Time | 1 | 1 | 3 |
Results
Below is a table of the results: green indicates that the system outperformed a dart throwing monkey (the group), red indicates that the system underperformed a dart throwing monkey. Since we held onto the top 12 stocks for one year, we should expect worse results from the short term time frames, with stronger results from the long term time frames.
| Period | Index | Group | ST+EQ | ST+Swg | ST+Vol | ST+Per | MT+EQ | MT+Swg | MT+Vol | MT+Per | MLT+EQ | MLT+Swg | MLT+Vol | MLT+Per | LT+EQ | LT+Swg | LT+Vol | LT+Per |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1998-Q4 | 15.83% | 39.85% | 8.24% | 8.24% | 10.58% | 2.73% | 156.84% | 170.32% | 116.60% | 220.87% | 230.44% | 109.43% | 166.78% | 231.32% | 93.94% | 70.62% | 39.76% | 196.53% |
| 1999-Q1 | 5.33% | 19.59% | 181.45% | 172.88% | 192.01% | 179.93% | 54.00% | 87.35% | 63.68% | 83.63% | 166.28% | 166.55% | 166.42% | 208.17% | 80.84% | 101.59% | 70.91% | 89.27% |
| 1999-Q2 | 11.98% | 36.25% | 52.82% | 20.51% | 55.04% | 92.09% | 182.09% | 189.83% | 235.82% | 181.82% | 69.79% | 33.57% | 117.43% | 108.62% | 169.36% | 132.91% | 238.59% | 162.39% |
| 1999-Q3 | -10.14% | 22.64% | -5.29% | 6.70% | 0.83% | 1.73% | -0.19% | -3.02% | 9.06% | 7.74% | -14.73% | -11.84% | 1.52% | -4.17% | 5.00% | 28.22% | 17.13% | -15.93% |
| 1999-Q4 | -22.57% | 11.94% | -29.41% | -12.17% | -16.37% | -39.67% | -44.97% | -12.05% | -47.26% | -45.27% | -37.83% | -11.47% | -34.68% | -53.00% | -37.77% | 1.46% | -35.85% | -37.57% |
| 2000-Q1 | -15.83% | 18.75% | -16.19% | -35.52% | -16.19% | -28.43% | -21.16% | -17.65% | -10.41% | -19.59% | -16.70% | -15.22% | -22.11% | -32.20% | -12.22% | -10.83% | -36.74% | -13.14% |
| 2000-Q2 | -27.70% | -5.61% | -19.52% | -14.93% | -14.52% | -38.47% | -19.73% | -27.95% | -22.97% | -26.96% | -8.21% | 5.71% | -2.83% | -27.45% | -34.68% | -32.05% | -45.24% | -46.71% |
| 2000-Q3 | -13.04% | 3.13% | 8.71% | -3.42% | -4.25% | 5.49% | 3.18% | 3.99% | 3.95% | -4.66% | -3.33% | 1.64% | 2.60% | -9.71% | -0.17% | -3.44% | 6.27% | -5.60% |
| 2000-Q4 | -1.19% | 13.32% | 13.90% | 7.87% | 9.56% | 2.08% | -1.49% | 4.14% | -4.54% | 8.09% | 16.97% | 17.08% | 20.35% | 8.86% | 34.30% | 24.80% | 24.80% | 27.97% |
| 2001-Q1 | -20.89% | -6.64% | -6.97% | -0.22% | -3.30% | -10.97% | -30.54% | -35.01% | -30.88% | -35.94% | -19.98% | -7.36% | -12.38% | -20.98% | 6.68% | 5.67% | 6.44% | -12.39% |
| 2001-Q2 | -18.36% | -5.19% | -16.30% | -20.47% | -15.53% | -17.46% | -11.54% | -9.07% | -6.33% | -18.25% | 0.52% | -8.12% | -1.09% | -9.79% | -2.70% | 8.30% | 11.33% | -0.93% |
| 2001-Q3 | -23.37% | -15.31% | -10.07% | 0.43% | -21.62% | -26.84% | -30.61% | -24.69% | -29.04% | -25.92% | -41.22% | -42.66% | -32.45% | -35.92% | -16.21% | -15.81% | -14.59% | -24.49% |
| 2001-Q4 | -25.12% | -22.04% | -20.98% | -20.01% | -19.58% | -27.80% | -41.40% | -44.86% | -41.40% | -44.60% | -23.24% | -24.81% | -22.96% | -24.45% | -5.85% | -5.81% | -8.40% | -2.49% |
| 2002-Q1 | 1.41% | 6.93% | 8.29% | 7.24% | 13.92% | 23.56% | 7.02% | 19.48% | 7.55% | -1.19% | 12.06% | 15.52% | 3.32% | 23.29% | -8.61% | -4.49% | -6.27% | -6.11% |
| 2002-Q2 | 20.09% | 37.57% | 41.08% | 21.31% | 22.07% | 33.77% | 33.47% | 28.16% | 36.17% | 36.49% | 37.18% | 36.96% | 32.09% | 21.38% | 19.86% | 22.52% | 19.16% | 23.05% |
| 2002-Q3 | 26.38% | 42.44% | 57.06% | 40.15% | 62.20% | 39.68% | 40.66% | 51.34% | 53.98% | 46.03% | 40.35% | 36.60% | 53.13% | 41.38% | 50.09% | 49.82% | 61.05% | 37.89% |
| 2002-Q4 | 31.88% | 54.21% | 45.18% | 50.48% | 42.41% | 41.89% | 54.82% | 52.90% | 70.99% | 49.80% | 38.90% | 57.89% | 54.28% | 32.82% | 37.63% | 31.88% | 66.77% | 31.21% |
| 2003-Q1 | 14.93% | 27.52% | 37.78% | 24.32% | 25.96% | 38.45% | 29.07% | 42.46% | 34.96% | 20.71% | 37.15% | 32.30% | 36.36% | 31.82% | 37.28% | 60.22% | 34.42% | 27.42% |
| 2003-Q2 | 11.13% | 20.88% | 40.04% | 42.67% | 36.98% | 39.99% | 36.93% | 32.25% | 45.72% | 46.02% | 28.02% | 23.65% | 27.59% | 22.42% | 17.06% | 15.63% | 10.18% | 18.41% |
| 2003-Q3 | 8.99% | 18.39% | 24.06% | 21.45% | 17.31% | 20.38% | 40.02% | 40.27% | 30.68% | 52.33% | 18.61% | 20.49% | 21.74% | 15.37% | 36.01% | 42.83% | 33.27% | 24.55% |
| 2003-Q4 | 3.60% | 11.12% | 25.24% | 20.67% | 24.00% | 32.44% | 30.83% | 23.80% | 24.54% | 8.82% | 24.27% | 18.17% | 24.57% | 35.32% | 5.98% | -2.42% | 16.98% | 5.23% |
| 2004-Q1 | 7.80% | 14.68% | 24.67% | 10.20% | 21.62% | 30.01% | 34.53% | 28.62% | 36.89% | 28.14% | 31.45% | 30.58% | 34.01% | 38.28% | 16.61% | 13.11% | 22.04% | 32.28% |
| 2004-Q2 | 7.55% | 12.99% | 3.32% | 5.73% | 12.94% | 11.56% | 31.61% | 35.91% | 25.13% | 25.76% | 15.50% | 22.20% | 13.27% | 20.09% | 27.88% | 25.75% | 24.70% | 31.78% |
| 2004-Q3 | 0.42% | 2.23% | 9.92% | 11.21% | -1.37% | 11.34% | -3.27% | -0.76% | -1.74% | 8.25% | 9.84% | 10.51% | 5.71% | 5.72% | 5.59% | 1.98% | 7.08% | 6.97% |
| Average | -0.45% | 14.99% | 19.04% | 15.22% | 18.11% | 17.40% | 22.09% | 26.49% | 25.05% | 25.09% | 25.50% | 21.56% | 27.19% | 26.13% | 21.91% | 23.44% | 23.49% | 22.90% |
| Records | 15-9 | 9-15 | 9-15 | 10-14 | 12-12 | 12-12 | 13-11 | 9-15 | 13-11 | 13-11 | 15-9 | 10-14 | 13-11 | 12-12 | 16-8 | 10-14 | ||
Conclusion
Using The Noiseless Return Rating to find stocks that are currently the strongest outperformers does indeed produce strong returns. However, as with any technical indicator that solely uses stock price data, these results are inferior during major shifts in the market.