Root time in term structure

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 hope you didnt go to bed yet Kev. You can just paste in R studio and hit run. Let me know what you think.
 
#
vol_front = 9.95/(sqrt(2/365)*sqrt(2/pi)*263.5)
vol_back = 12.15/(sqrt(9/365)*sqrt(2/pi)*263.5)
Don't convert to var then convert back to vol, leave as var, otherwise you are just converting back to var in the next step. Use 7 days-to-expiry for the Jun15 expiry -- since this method ignores rates and divs, no reason not to use trading days. Also don't use "t" for days to expiry, use d or dte instead as t in options related formulas usually refers to time in fraction of a year.

S <- 263.5
P1 <- 9.95
P2 <- 12.5
Var1 <- (P1/S)/sqrt(2/pi)
Var2 <- (P2/S)/sqrt(2/pi)
Dte1 <- 2
Dtd2 <- 7

impliedJumpSigma <- vol_jpm(Var1,Var2,Dte1,Dte2)


I rewrote your function in var terms:

vol_jpm <- function(var1,var2,dte1,dte2)
{ varj <- (dte2*var1 - dte1*var2) / (dte2 - dte1)
vj <- sqrt(varj)
return(vj}
}
annualized_vol_jpm <- function(vj)
{
return(vj * sqrt(252))
}
 
I hope you didnt go to bed yet Kev. You can just paste in R studio and hit run. Let me know what you think.

I generally work asian hours, so I am up late. I also prefer R console over R Studio even though I may be the last one still using it.

From your input numbers I get a jump sd of about 4.1% or about 65% annualized.

The general method you are using is ok as far as it goes, but it is pretty crude. The big problem is that earnings expectations manifest themselves in strikes mostly other than atm. The other big problem is that you are assuming base vol in the near and the next contracts is the same -- a very unreliable assumption.
 
The other big problem is that you are assuming base vol in the near and the next contracts is the same -- a very unreliable assumption.

Putting the possibility of another upcoming idiosyncratic event aside, would you normalize this in accordance with the VIX term structure or similar? I would think normalizing past event vol to current conditions is probably a necessary preliminary step as well (although that procedure can get hairy fast).
 
I generally work asian hours, so I am up late. I also prefer R console over R Studio even though I may be the last one still using it.

From your input numbers I get a jump sd of about 4.1% or about 65% annualized.

The general method you are using is ok as far as it goes, but it is pretty crude. The big problem is that earnings expectations manifest themselves in strikes mostly other than atm. The other big problem is that you are assuming base vol in the near and the next contracts is the same -- a very unreliable assumption.
So does it make more sense to trade otm calendars? More volga in the just OTM as well. However I would have to put a calendar on each side. Could you explain a bit more? Thanks for brushing up my code
 
I generally work asian hours, so I am up late. I also prefer R console over R Studio even though I may be the last one still using it.

From your input numbers I get a jump sd of about 4.1% or about 65% annualized.

The general method you are using is ok as far as it goes, but it is pretty crude. The big problem is that earnings expectations manifest themselves in strikes mostly other than atm. The other big problem is that you are assuming base vol in the near and the next contracts is the same -- a very unreliable assumption.
Also a very important question: if I think the vol is underpriced. Do I sell the event expiration and buy the back expiration?
 
If you read this before the close. DOCU JUL/SEP calendar is 2.00. Just got filled. Next earnings not priced in! The July vol will get crushed but Sept vol should remian high. Hope you guys can get in.
 
Hi

When you say next the earnings are not priced in, what do you mean?

Why do you think the calendar went to 2.00 to 2.90 almost 30% in one day, what are the reasons in your opinion? They priced the earnings in one day that weren't priced the day before?

thanks
 
Back
Top