Oh, ok, now I understand your question - it was so intuitive that vega notional times change in vol strike is roughly your P&L that I did not get it. Sorry.
Yea, the intuition made sense. I just wanted to know how the actual mechanics, in terms of how the FCM clears / reports the trade, how the exchange settles the contract, would function to back up that intuition. I mean, I need to figure out how I would allocate these trades to my managed accounts at multiple FCMs, and how to account for subsequent gains/losses.
vega.. is the change in the underlying per change in the implied volatility ... if thats how your using it.. how is this not intuitive.. your vegas change when trading options.. you don't get flat with the same amount of vegas as your opened with in a profit or loss scenario in options.. its a ratio it would be rare it would ever be the same. i must be wrong with how yoru using vega... please explain
I'm not talking vega in the generic, theoretical sense. I'm talking about the specifics of how this contract trades, like if I or you wanted to go out and place an order for this thing. You can't submit a buy order for "one contract" at 19.50, the way you might do the VIX (or any other security/future I can think of). Instead, you place an order in vega notional units, but actually get filled in a different unit entirely (futures contracts / variance units). That's the point I was observing. If you fill an order to buy 1000 vega notional, you in most cases will NOT be able to close out that order by selling 1000 vega notional. (And I wonder how how all the trading front-ends that have something called 'flatten position' will react to this?)
I'm going to re-work this example, but from the day 1 example so I understand it. Let's say we bought the Jan 13 contract on the first day it listed, right at the 19.9 variance strike. N = 106, n = 0 (no days have elapsed). realized portion = 0, of course. implied portion = 19.9^2 * (106-1) / 252 = 165.00 Kt = 165.00 * 252/(106-1) = 396... which is of course K0. So futures price on day 1 = Kt - K0 + 1000 = 1000. So far so good!
Back from boat club bar. thanks for clarifying my desire..the only condition I'm OK for is reading you guys at the moment.
I was only wondering if in practice you make any adjustments to get a better pricing. I don't want to go deeper in theory and drown b/c of my ignorance. Just to explain better what I thought: I was referring to the calc of the risk neutral expectation of variance and whether you use a correction or not assuming the process is a diffusion or not. re the realised variance: The formula used is an estimator based on the daily settlement prices. But considering that the underlying trades obviously more frequently intrady, if you take higher frequency returns then you will get a more accurate estimate of the realised variance and probably different. Again my question was from a practical point of view and not academic/theoretic and your reply is sufficient (I can have a look at what "var basis" exactly is)
Just to be clear, in the context of these contracts, realized vol (just like vega) has a very specific definition, based on end of day settlements. We can talk about realized vola estimators all day long, but for this contract, there's no estimation involved at all.
So with annualized volatility at 19.95 (continuing Jan 13 example above)... Daily volatility is roughly 1.25. In terms of impact on futures price, 19.95^2 / 252 = 1.58 * 106/252 = 0.66 rolls off the implied side every day. On realized side, if daily move comes in at roughly 1.25% (19% converted to daily), you get exactly that same 0.66 number. Let's call that our daily var exposure. If you get a two-sigma day, then it's (1.25*2)^2 increase in realized side...4x more than 1-sigma number. So futures go up daily var*3. If it's a 3 sigma day, then it's daily var exposure * 7. If it's 4-sigma, then it's daily var * 15. If it's 0.5 sigma, then you lose 3/4 daily var. Before I calculated for 20000 vega notional we would roughly have 3500 contracts... So a 2x sigma day translates to gains of roughly $7k (if long). A 4 sigma one-day realized move is paper gains of $34k. A half sigma day and you lose roughly $1800. (All of this in addition to the $20k vega notional exposure.) Question: if I wanted just realized and no implied exposure, I guess I should hedge with vix futures? Adjust the number of Vix dynamically as the days roll off the contract?
I understand this (or at least that's what I think). I was only wondering in my OP if in practice, you the traders, make any adjustments. For example let's say you have (in theory) a 100% accurate prediction model for the variance of the SPX500 over the 30 days to maturity of a var swap contract. However what you predict now very accurately will not be (?) exactly the "realised variance" on maturity based on the market definition for the contract Anyway there is a lot of interesting and practical material in this thread and I am actually studying your example posts, so I would hate to derail it.