<t>Thanks Alan, I'm now going to acquire 5 minute data for the underlying and see if I can account for these seasonal effects. I've been going through Bollerslev & Andersen (1997/8), is there anything else you could recommend me on intraday volatility estimation/forecasting? A lot of literature ...

<t>This seems quite possible, using a very naive extrapolation of the vol term structure I got something which followed their live odds quite closely which works fine at the weekly scale. At smaller scales, it's less and less reliable, and the proportion by which I need to modify my IV to approximat...

<t>Thanks for your reply Alan, I'm investigating how some gambling companies price their fixed odds very short term markets on stock indices. These are fixed-odds bets which expire every 5 minutes and allow the client to bet on the market to finish above / below various strikes close to the money, i...

<t>Hello, hope this is the best place for this!I was wondering if anyone could point me in the right direction in terms of interpolating/extrapolating the volatility surface to small time, for the pricing of very short term options (5-10 minutes!)Should I be investigating the asymptotic expansions o...

<t>McWulf - I followed the paper through and did my own monte carlo tests, have a spreadsheet if you'd like to see, it all works out very well. If you assume no trading is a 0 return you'll end up with a heavily zero biased correlation estimator. Now just looking for ways to exponentially weight it ...

<r>I'm calculating some covariance matrices, the stocks involved vary in liquidity. I have found this paper here <URL url="http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1020943"><LINK_TEXT text="http://papers.ssrn.com/sol3/papers.cfm? ... id=1020943">http://papers.ssrn.com/sol3/papers.cfm?abstr...

<t>Hello. I was grateful for MHill's succinct reply, although I'm fuzzy about how to calculate the covariance of timeseries with different timescales between them. I've read something about looking at overlapping periods of both series... I posted in a new thread just to have it in another topic/tit...

Hi all,Quick question - how do I model correlation/covariance of time inhomogeneous time series, due to illiquidity? Say I have a liquid time series sampled daily, and an illiquid product has zero returns due to no volume, do I have to do some interpolation/time scaling?Kind Regards,J

Thanks MHill for the clarification

Thanks for the help, works much better in backtests! However doesn't (2%)^2 * 5 = (10%/sqrt(5))^2 = 0.002?

MHill, is that essentially the same as linearly interpolating to create a homogeneous time-series from?

<t>Hi all,What is the best method for modelling returns of illiquid stocks which don't trade for days at a time? I'm looking to measure daily univariate risk and volatility. I've tried fitting skewed student t distributions but I can't get ghyp to converge when the timeseries of returns is mainly ze...

Thanks for the help. The quick fix seems to have worked well.

<t>Hi all,I'm using some historical option price data to explore hedging models - I solved BS for implied vol using newton raphson hoping to calculate greeks, however when time to expiry is small I'm experiencing problems calculating IV for ITM calls. Is there any way around this problem, or is ther...

<t>Thanks for the tips. Do you have any pointers as to where to read further into a matrix method? I'm assuming that I can merge all delta exposure into my var book and monitor the other greeks. Is it standard procedure to simply limit individual greek exposure for each instrument and greek, or is t...