Raltin has a comprehensive list of features & factors that investors use in machine learning and quant models for investment and risk management. Raltin also utilizes these features to demonstrate to customers and other investors, on the power of these data sets to outperform market benchmarks. We tested thousands of combinations to see what signals show potential for outperformance. One high value and unique signal in our library is Options and Implied Volatility (IV) related metrics. As most investors know, IV is a critical factor in pricing options. However, it's not easy to track IV over time for a given ticker because options have an expiry date. To overcome this expiry issue of options, the VIX was created to track the volatility of the S&P 500 over time. However, a VIX like measure is not available for every ticker. Therefore, Raltin has created an IV measure, akin to the VIX, for all tickers across moneyness and expiry dates. Using this VIX like IV measure, Raltin sorted the top 50 stocks with the highest market cap into five quintiles based on IV. Quintile 1 had the highest 30 day At the Money (ATM) IV and Quintile 5 had the lowest 30 Day ATM IV. The strategy would involve buying the stocks in each quintile at pre-specified rebalancing time (7 days). At the point of rebalancing, the stocks would be sorted again based on IV and organized into quintiles and the strategy would then again be executed on purchasing Quintile 1 stocks. The backtesting period was 2nd Jan 2014 to 27th November 2019. Our results showed that stocks in Quintile 1 (with the highest implied volatility) showed an annualized IRR of 16.63% compared to the SPY (S&P 500) of 11.7%. Quintile 1 outperformed SPY by approximately 5% over the entire period. Raltin Pro and other institutional clients get access to additional data points for every feature and portfolio backtest including: Outcome measures: Sharpe Ratio, Hit Ratio, Information Ratio etc Factor Measures: Median values for each factor and quintile Portfolio Details: What stocks/tickers are most common in several portfolio cycles etc. Please email raltin@raltin.com for more details and demo of our feature and back testing software.
Would be good to see 00/01 and 07/08 years, and also to discern this apart from simply having a high beta portfolio.
Higher information ratio in theory means there are excess returns even after accounting for market covariance. It's surprising because it goes against the resuts for realised vol, eg "Betting against Beta" where low vol stocks outperform high vol stocks. GAT
The conclusions are dangerously over-generalized. For as often as I preach about having relevant time slots (that reflect the *current* market and not some arbitrarily long period stretching back to Hammurabi to "prove" relevance), the study goes 5 years back, but makes repeated *general* statements -- NOT statements with the phrase "over the past 5 years" or "in the recent tech boom" or even "in a bull market environment." I LIKE the execution, but to paint is with absolutes is wrong -- and could cause lots of unsuspecting traders to conclude "that's the way it is" and base all their trading as if the market always goes up, tech leads the way, and IV>market is an indicator of future above-market returns. "Not good."
Thank you very much for your feedback @tommcginnis @globalarbtrader Our team is working on updating our analysis with your feedback points and we will post back here once we completed our updates.