Unique
3 Returns
Home | Hall of Fame | Reval | Data Library | Buy Books! | Software Tools | Principles | Feedback | Subscribe
Unique
3 Returns
Home | Hall of Fame | Reval | Data Library | Buy Books! | Software Tools | Principles | Feedback | Subscribe
| Summary of Returns for Unique 3 | ||||
| Buy Date | 1 mo Return | 3 mo Return | 6 mo Return | 12 mo Return |
| 1/9/98 | 22.37% | 41.51% | 57.00% | 100.98% |
| 2/6/98 | 12.71% | 24.77% | 14.71% | 93.30% |
| 3/6/98 | 3.13% | -0.28% | -4.11% | 31.45% |
| 4/3/98 | -1.26% | 6.63% | -3.36% | 18.19% |
| 5/1/98 | 0.59% | 27.58% | 56.28% | 91.23% |
| 6/5/98 | 9.07% | -0.23% | 31.84% | 56.94% |
| 7/6/98 | 0.54% | -7.13% | 44.98% | 53.36% |
| 8/7/98 | -13.59% | -0.05% | 29.25% | 7.16% |
| 9/4/98 | -1.87% | 19.02% | 41.39% | 51.26% |
| 10/2/98 | 14.88% | 24.73% | 29.68% | 51.98% |
| 11/6/98 | 7.99% | 31.80% | 45.89% | 72.51% |
| 12/4/98 | 19.50% | 26.31% | 41.74% | 111.44% |
| 1/8/99 | 6.45% | 21.11% | 36.09% | 89.23% |
| 2/5/99 | 0.67% | 8.52% | 14.86% | 90.59% |
| 3/5/99 | 18.14% | 8.88% | 32.61% | 216.99% |
| 4/1/99 | -1.94% | 11.88% | 30.63% | 213.96% |
| 5/7/99 | 4.41% | 5.48% | 51.85% | 101.32% |
| 6/4/99 | 12.16% | 14.13% | 58.01% | 83.16% |
| 7/2/99 | -10.02% | -0.21% | 90.54% | 102.53% |
| 8/6/99 | 10.41% | 27.49% | 95.06% | 84.25% |
| 9/3/99 | -3.99% | 41.78% | 89.15% | 80.57% |
| 10/1/99 | 35.36% | 85.70% | 125.96% | 76.96% |
| 11/5/99 | 16.57% | 39.29% | 47.97% | 41.95% |
| 12/3/99 | 15.60% | 42.19% | 32.47% | -2.63% |
| 1/7/00 | 13.26% | 30.56% | 20.61% | -19.35% |
| 2/4/00 | 34.26% | 10.25% | 20.72% | -25.06% |
| 3/3/00 | -8.54% | -15.23% | 10.92% | -65.54% |
| 4/7/00 | -12.66% | -4.51% | -1.61% | -77.29% |
| 5/5/00 | 11.42% | 25.73% | 17.56% | -61.89% |
| 6/4/00 | 3.77% | 28.77% | -31.16% | -62.90% |
| 7/7/00 | 6.97% | 2.71% | -38.99% | |
| 8/4/00 | 15.54% | 4.04% | -30.65% | |
| 9/1/00 | -16.58% | -41.15% | -60.41% | |
| 10/6/00 | -6.65% | -37.00% | -63.08% | |
| 11/3/00 | -30.57% | -32.26% | -54.24% | |
| 12/1/00 | -1.92% | 9.00% | 17.18% | |
| 1/5/01 | -11.07% | -10.21% | ||
| 2/2/01 | 2.90% | 4.07% | ||
| 3/2/01 | -7.51% | 9.55% | ||
| 4/6/01 | 18.21% | |||
| 5/4/01 | 1.33% | |||
| Arith. Average | 4.64% | 12.44% | 24.93% | 53.55% |
| +3 Sigma | 52.71% | 114.42% | 313.83% | 817.18% |
| +2 Sigma | 34.27% | 71.59% | 170.85% | 379.96% |
| +1 Sigma | 18.06% | 37.32% | 77.27% | 151.16% |
| Geo. Average | 3.79% | 9.89% | 16.02% | 31.43% |
| -1 Sigma | -8.72% | -12.06% | -24.06% | -31.22% |
| -2 Sigma | -19.74% | -29.62% | -50.30% | -64.01% |
| -3 Sigma | -29.43% | -43.68% | -67.47% | -81.17% |
| Estimated CAGR | 56.28% | 45.84% | 34.62% | 31.43% |
| Base Rate v. S&P | 68.29% | 74.36% | 83.33% | 76.67% |
| 3 mo Rolling Base Rate | 74.36% | |||
| 6 mo Rolling Base Rate | 83.33% | |||
| Total Growth | 359.75% | |||
The "estimated CAGR" is a statistical estimate of the compounded annual growth rate of the screen, given the data so far. I've estimated CAGR by taking the geometric average of the returns, and then multiplying that result by itself by the number of periods in a year (e.g., 12 times for the 1-month holding period, twice for the 6-month period). All of the estimated CAGR numbers are (duh) annual (that's what the "A" stands for), so it gives a better apples-apples comparison. The formula is :
"Sigma" uses the geometric standard deviation of the data to show the range of expected returns. About 2/3 of the time, the returns should be +-1 sigma from the geometric average return. For instance, 2/3 of the time, the annual version of Unique 3 should return between -31% and 151%, at least in theory. The returns should be +-2 sigma about 95% of the time, and +-3 sigma approximately 99% of the time. These measurements are accurate only if we assume that the returns have a log-normal probability distribution, and that the data we have is representative of what we can expect in the future.
The "base rate" of a strategy is the percentage of the time it beats some benchmark, such as the S&P 500. I've chosen S&P Depositary Receipts, aka SPDRs, as the benchmark for Unique 3. Realistically, that's the most economical way of investing in the S&P 500, and historical quotes are easily available from Yahoo. (Here are the S&P numbers I used). Base rate is another way of testing the volatility of a screen. A low base rate on a high-performing screen means the screen's numbers are suspect and that the high "average" performance is attributable to one or two outstanding periods. Conversely, a high base rate indicates that the strategy really does "beat the market."
"Rolling" base rate is best explained by example. The 3 month rolling base rate compares how you would have done investing in Unique 3, rebalancing the portfolio every month, for three months. We look at all possible 3-month periods, January-April 1998, February-May 1998, etc., and compare the total return of Unique 3 to the total return of SPDRs in each of those periods. In 74% of those periods, Unique 3 wins. If you lengthen the rolling period to 6 months, Unique 3 wins 83% of the time.
Last Updated on 6/10/2001
By Jonathan Jackel
Email: jonathan@jjackel.com