THIS IS YOU: "Volatility decreases exponentially with time." That is not even remotely twistible to being correct, tenable, supportable, argument-worthy, or even imaginable under alternative realities, ferGAWDsakes. It *defies* explanation. Post something correct 10 times in a row. I dare you. Show your own trades, from a platform activity sheet. Do ANYTHING to justify your manifest ego or even your [asprirational] moniker. OR, "learn."
Okay this is what I have come up with for when one should initiate a calendar. Lets say BABA has their earnings in a month and today the implied move on earnings is 2 standard deviation below the mean expected move. I could sell the weekly in front of the event and buy the event vol. OR I could buy the event vol and sell some spx vol (theta/gamma/beta weighted) to hedge.
The root time has to do with the options pricing, not necessarily with the volatility. They are different things. Volatility eventually evens out... 4 and 5 year vols will be the same... since there is no reason to think that the 5th year vol should be any different than that of the 4th year... forward vols are the same. So while there usually is a certain term structure in the (implied) volatility... that doesn't have anything to do with root time... But since vol is a derivation of expected movements, root time will have it's effect on the expected underlying move. I will lift a little veil for you guys before I go on holidays. The BS-formula is actually quite easy to derive to a simple formula for pricing a straddle. Basically, it will come down to this: Straddle = S x Vol/100 x sqrt(T)/24... or S/100 x Vol/24 x sqrt(T) S= spot, Vol in points, T = time in days So if keeping vol fixed over time, the straddle 'moves' in root time. S=100, Vol=24.. T=30 > straddle = 5.48... or 5.48% of the spot. T=60 > straddle = 7.75... which is sqrt(60/30) x 30 day straddle. Difference between straddle values is therefore mainly determined through (root) time. And maybe some more if there's a certain term structure in the IV....
Santa I have been putting alot of thought into your statement vega "gains" must out weigh your gamma "losses". So our vega gains will be IVT2*VegaT2 - IVT1*VegaT1. Assuming we do an ATM calendar where volga and vanna are both 0...what am I looking for??? Like how will i be able to tell that the front month wont increase at a faster rate than vegaT1/vegaT2. Like is that even possible to beable to predict the rate of vol change in the term structure when an event happens?
Im working on a project where I trade the earnings vol before the announcement. What I have found is that the earnings expected move changes every day usually not by a lot but sometimes it can get pretty wide. For example LULU had an event vol of 192% the day before the earnings and 1 week prior it was 192% +/- 5% however there was a day when the event vol was only priced at 121%!! Im trying to play this with a calendar spread but it's kind of difficult to understand. How I should be using them because there are so many moving parts. I wouldn't mind sending a PM to anyone who is interested and is okay at reading some R code. Would love some help. Thanks.
What you are really looking for in an earnings calendar is the event priced differently in the two options by a large enough amount to overcome expected transaction costs. Is that how you are analyzing them? Post the code here, publicly. That way others can learn from the discussion also. I, and others, will probably comment on it. I've been reading R since before it existed.
# I will be using AVGO as an example here as they have there earnings tomorrow # Both its current implied move and it's past implied move will be used to show #the spread in which I want to take advantage of. #First I calculate the volatlity of closest expiration to event using the #straddle price. # front month vol = straddle price /(sqrt(t/T)*sqrt(2/pi)*Stock Price) vol_front = 9.95/(sqrt(2/365)*sqrt(2/pi)*263.5) # vol June 8 = .639 #next I calculate the second closest vol. vol_back = 12.15/(sqrt(9/365)*sqrt(2/pi)*263.5) #vol June 15 = .368 #Now I calculate the event vol. #Event vol = Vol - diffuse(ambient vol) #I have created a function called vol_jpm that does this vol_jpm = function(v1,v2,t2,t2) { vj = sqrt((v1^2 - v2^2) * (t1 * t2/(t2 - t1))) vj } #Plug in the number to the fucntion volj = vol_jpm(.639, .368, 2,9) #The vol of the event is 84.7% #We can calcualte the expected move using this. volj*sqrt(2/pi)*263.5*sqrt(1/365) #We get an expected move of 9.21. #Now lets go back a few weeks #May 14. roughly 3 weeks ago. #spreads are only .30 cents wide. small for a $250 stock. # We repeat the same steps to calculate the expected move at that time. vol_front_past = 15.80/(sqrt(25/365)*sqrt(2/pi)*244.76) vol_back_past = 17.55/(sqrt(32/365)*sqrt(2/pi)*244.75) volj2 = vol_jpm(vol_front_past, vol_back_past, 25, 32) #this implies an event vol of only 62%....22% less than the first one!! volj2 * sqrt(2/pi)*244.76*sqrt(1/365) #This implies an expected move of only 6.40. #The 240 calendar on may 14 was priced at 2.50. It ran up to 3.05 withing 2.5 days. #Hope you guys can contribute cheers