Kudos to MMs

Discussion in 'Chit Chat' started by nitro, Oct 23, 2008.

  1. 1120 seems a very reasonable target, as well.
     
    #3421     Aug 16, 2011
  2. nitro

    nitro

    Added L ES @1185. Will work into a SPY option as the day progresses
     
    #3422     Aug 16, 2011
  3. nitro

    nitro

    Odds 3:1 against seeing 1120 before 1220.
     
    #3423     Aug 16, 2011
  4. Those damn exogenous events... like markets open from nine to four.
     
    #3424     Aug 16, 2011
  5. Please explain how you calculate the odds.
     
    #3425     Aug 16, 2011
  6. nitro

    nitro

    Get VIX. Compute implied move in SPX. Do this for up and down moves (distributions is not symmetric). Feed into Bayes.
     
    #3426     Aug 16, 2011
  7. You're using skew in your calc? If you're going to use skew then you may as use delta (also shit bc of symmetry/moneyness and skew). The result needs to be an equal dvega/dvol for the call and put, but of course the put dvega will be lower. IOW, do not use an implied distribution or use ATM.

    Also IOW, don't back-out SABR.
     
    #3427     Aug 16, 2011
  8. I'd note that there is a much simpler way to come up with the same 3:1 odds:

    current SPX: 1195
    target1: 1220
    target2: 1120

    The distance to target1 is about 3 times shorter than the distance to target2. Using a simple Brownian motion model (a.k.a Einstein's model), the odds of reaching target1 before reaching target2 are 3:1, and this is invariant with respect to time and volatility.
     
    #3428     Aug 16, 2011
  9. nitro

    nitro

    No, imo. That assumes equal risk of downside to upside, and that markets move in equal steps up and down. If markets had no memory, that of course would be true. In other words, if there was a flat options skew, there would be [stat] arb possibilities.

    Markets are really complex. For example, if the German data came out at a slightly different day, the downside today could have been double. The dimensionality is really big, but it is not infinite, and therefore computable. That is why it takes years to learn this business if you don't do careful computer analysis, someone tells you, or you get lucky. Multi-dimensional derivatives and correlations are hard, especially for a human brain that can hardly remember a 7 digit phone number.

    The big epiphany I had was how different players affect markets, and in fact, it shows just how poorly short sellers understand markets or they could sell much higher if they waited a week. What they are saying is they are not willing to risk another exogenous event that would force them to sell lower. That shows a poor understanding as well. Of course, I had the same problem on the down move (lack of understanding of human psychology that I couldn't model then).

    I can safely say that I understand the upside better than the downside at this point. The downside is more complex because fear is far more powerful than greed in the human brain, so the down corrections to a model are more complex, if they can be modeled at all (I had a recent epiphany here too - that I sort of can).
     
    #3429     Aug 16, 2011
  10. noddyboy

    noddyboy

    How do you get upside and downside risk from a single number like VIX? Don't you need to know the skew?
     
    #3430     Aug 16, 2011