# What are some common methods for calculating short term historical volatility (i.e. look back 5 minute time periods)

I'm interested in quantifying the impact of short term price volatility on a particular strategy I'm running. So far I'm simply calculating the standard deviation of log returns, but I'm a bit unsure on where to take it from here. Additionally - any textbooks/further reading on this sort of thing appreciated.

• What is the frequency of your data? It sounds like you're looking for realized volatility estimators. In general, if you have a sub-minute frequency it is often a good idea to consider noise-robust estimators, since price-data (TAQ data) tend to be noisy (for equities). Here, liquidity of your considered instruments, also have an impact on your estimates and sampling scheme.
– Pleb
Oct 7 at 11:15
• I am using sub-minute frequency, thanks, I'll look into these key words. Oct 7 at 17:27
• For noise-robust realized estimators you can take a look at the Realized kernel estimator of Barndorff-Nielsen et al. (2009) or the Preaveraging estimator of Jacod et al. (2009). They both exist in a multivariate version and are implemented in the (very detailed and thorough) R package called highfrequency. They recently updated their documentation of the package, which is available here.
– Pleb
Oct 7 at 18:04
• Reading their documentation will give you loads of references and applicable estimators besides the ones I've talked about above. I hope this helps :-)
– Pleb
Oct 7 at 18:06
• Thanks! This looks great. Much appreciated Oct 8 at 12:04