[Music] Alphas by holding frequencies and delays as a reminder World Quant defines an alpha as a mathematical model that seeks to predict the future price movement of various financial instruments I'm nitish Manny your partner in learning about Quant Finance research I'm the chief strategy officer at World Quant in my role I help to Define our firm's strategy and drive several cross functional business initiatives including new avenues for growth and innovation in this video we will focus on the role played by holding frequencies and data delays holding frequency is the total length of time an investor
expects to hold a portfolio or security the holding period has has implications on the risk returns and transaction costs data delay refers to the timestamp before which the alpha is allowed to use the data in its pack test more recent data has upto-date information which may benefit the alpha performance but data delay can help ensure there are no look ahead biases an important idea we will explain in depth in this video If an alpha uses data from a day before the date of back test I will call it a delay 1 Alpha and if the
alpha uses today's price up to a chosen time during the day for simulation to run after that time I will call it a delay zero Alpha because it uses the same day information we will also put our ideas to the test on world Quant brain our cloud-based platform where we can get realtime feedback if you would like to try this as I explain these ideas log into brain [Music] now Alphas can be classified into three categories based on the holding period the medium to low frequency Alphas these are alphas with holding periods that range from
a few days to months and sometimes even longer intraday trading these update positions much more frequently ranging from every minute to a few hours they are typically based on price volume signals to capitalize on shortlived price patterns they can also use news data earnings announcements or faster intraday data like sentiment or options high frequency trading or hft Alphas these hold positions from Nan to a few seconds hftd signals can be Market making strategies that provide Market liquidity or Price Arbitrage strategies that capitalize on price discrepancies between related assets or markets they can also take short-term
directional trades based on price patterns like short-term momentum it is important to prevent biases in the data during back testing look ahead bias happens when you use future information in your analysis that wouldn't have been known or available during the period being analyzed this bias can lead to over optimistic performance predictions that don't hold up in reality skewing over expectations and potentially leading to costly surprises to avoid the look ahead bias quants use delayed data in the back test delay one Alphas use yesterday's price delay zero uses today's price up to a chosen time during
the day theoretically delay zero Alphas are expected to perform better than delay 1 but have a lower trading capacity due to shorter time available to trade we have reviewed several Delan Alphas in our previous videos so let's shift to a delay zero example we will use volatility data focusing on at the money implied volatility of call options expiring within 4 months and their Parkinson's volatility Parkinson's volatility measures realized volatility using the the high and low prices within a day it captures large intraday price variations even when the previous day close and the current day close
price change is small the strike price is where the call option starts making a profit and implied volatility represents the expected future stock movement highly influenced by option demand add the money implied volatility Aggregates implied volatility ities of call options where the strike price equals the Stock's current price and the option expires within 4 months our Alpha idea captures call option demand using the implied volatility of its at the money call options expiring up to 4 months ahead scaled by Parkinson's volatility over the past 4 months if the implied volatility to Parkinson's volatility ratio is
high the hypothesis is that we expect High future stock returns and vice versa now let's implement this idea on world Quant brain using its proprietary Expressions language in this example we simply take the ratio of the implied volatility of add the money call options to Parkinson's volatility for the stock using the data fields on the brain platform these delay zero values are are available by the cut of Time Each Day in the markets for the alpha to generate positions the results show a consistent sharp of about two across years of 14% gross returns 1.2 returns
to draw down ratio 29% turnover and decent coverage across the selected Universe in this video we learned about three holding periods in Quant Finance the importance of delayed data to prevent look ahead biases and reviewed a delay zero Alpha example in the next video we will discuss the power of diversity of Alphas step by step I hope that you are powering up your knowledge and testing out your ideas with me let's go Quant see you in brain [Music] h
