Quant Investing: Rule-Based Strategy using Realized Volatility
This report introduced two measures of realized volatility, RV30d and RV5min, and dives into their own characteristic.
| Description | Past | Current | Future Expectation | |
|---|---|---|---|---|
| VIX | a 30-day forward projection of volatility | O | O | |
| RV30d | historical measures using daily returns over the prior 30 calendar days | O | O | |
| RV5min | realized measures using five-minute and intraday returns | O |
By using peak comparison and cross correlation in high volatile market, we found that for both RV30d and VIX, RV5min has its potential leading effect and could be utilized in the construction of rule-based trading strategy.
A sensitivity analysis is then introduced to calibrate this rule-based strategy with two parameters, and surprisingly we found that such strategy shows a good robustness over ten-year Walk Forward Optimization.
| Slow Fast Line | Mean of Ranking | Std of Ranking | Total Occurrences | ||
|---|---|---|---|---|---|
| RV30d,RV5min | 8 | -0.33 | 3.4 | 1.505545 | 10 |
| RV30d,RV5min | 8 | -0.32 | 2 | 1.154701 | 10 |
| RV30d,RV5min | 11 | -0.33 | 2.4 | 0.843274 | 10 |
| RV30d,RV5min | 9 | -0.32 | 6 | 2 | 9 |
| RV30d,RV5min | 10 | -0.32 | 6.25 | 2.866058 | 8 |
Finally, according to “Calm” performance in backtesting results, the increase of Sharpe Ratio is about 70% compared to buy-and-hold SPY, and the decrease of Max drawdown is about 62%.
| Start date | 2008/1/14 | ||
|---|---|---|---|
| End date | 2021/2/24 | ||
| Total months | 156 | ||
| Calm | Turbulent | SPY | |
| Annual return | 0.170475 | -0.048152 | 0.114114 |
| Cumulative returns | 6.778178 | -0.474347 | 3.088626 |
| Annual volatility | 0.161742 | 0.131747 | 0.208741 |
| Sharpe ratio | 1.054236 | -0.308327 | 0.622269 |
| Calmar ratio | 0.869337 | -0.085489 | 0.221661 |
| Stability | 0.957849 | 0.121534 | 0.9525 |
| Max drawdown | -0.196098 | -0.563249 | -0.514815 |
| Omega ratio | 1.227241 | 0.806868 | 1.135207 |
| Sortino ratio | 1.543685 | -0.410592 | 0.875957 |
| Skew | 0.234356 | -0.457887 | -0.045594 |
| Kurtosis | 11.153861 | 79.610678 | 15.246903 |
| Tail ratio | 0.938033 | NaN | 0.900183 |
| Daily value at risk | -0.019701 | -0.01676 | -0.025783 |
| Alpha | 0.096836 | -0.088317 | 0 |
| Beta | 0.601017 | 0.398983 | 1 |
[1] Edwards, T., & Preston, H. (2017). Reading vix R: Does vix predict future volatility. S&P Global Working Document.
[2] Jeffrey Lin (2021). Technical Report 202101. Gamma Paradigm Research.
[3] Bekaert, G., & Hoerova, M. (2014). The VIX, the variance premium and stock market volatility. Journal of Econometrics, 183(2), 181-192.
[4] Liu, L. Y., Patton, A. J., & Sheppard, K. (2015). Does anything beat 5-minute RV? A comparison of realized measures across multiple asset classes. Journal of Econometrics, 187(1), 293-311.
[5] Y.C. Lin (2020). Technical Report: HMM on winner portfolio. Gamma Paradigm Research.
YC Lin
AVP Quantitative Finance, Gamma Paradigm
LinkedIn
Contact us at info@gammaparadigm.com if you want to know more.
