Kudos to MMs

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

  1. nitro

    nitro

    All I do is lose money. I am second guessing everything.
     
    #1321     Mar 24, 2010
  2. nitro

    nitro

    "NFV" goes out at ~1161.10. SPX goes out at 1167.72.

    Convergence is happening. Even 6 point difference is probably within the realm of model noise/error.
     
    #1322     Mar 24, 2010
  3. nitro

    nitro

    "NFV" ~1164.73
     
    #1323     Mar 25, 2010
  4. I think you made an error this morning. "NFV" is really ~1164.73898979238675910239847789781203984710293874
     
    #1324     Mar 25, 2010
  5. nitro

    nitro

    Under no circumstances is it correct to short here, imo.

    If you must short (and you are a masochist), wait to 2:55 CT, and then decide if you want to short. Then stay up all night and try to make a few handles into tomorrow while deciding whether to ride the short. Either way, flat is the right position into Friday's close, imo.

    There is really not much edge to shorting at this point anyway according to "NFV". "NFV" is pretty good at keeping pace with SPX (so far), and the new edge band appears to be around >20 points of divergence. As I write this, we are at about -13.
     
    #1325     Mar 25, 2010
  6. the way i understood it is that your system consists of two parts:
    1) hard-coded formulas for FV and entry/exit rules
    2) nitro's trading experience on top of 1) that may override/adjust 1) on the spot.

    part 1) can be automated and is generally should be profitable
    part 2) can't be automated (yet) but presumably is adding value to the system

    part 1)+2) can't be backtested objectively, but part 1) can be.

    my question: has part 1) had so much troubles before in backtests and during real trading. if yes, then when those tough times happen? 2003, 2009?
     
    #1326     Mar 25, 2010
  7. nitro

    nitro

    Those are really good comments. Now that I think I finally have a model that comes close to modeling SPX "FV", I am in the process of getting data to test 1). It is a lot harder than I thought it would be. The periods that you recommend are also interesting since they have similar "fundamentals" in the form of low IRs following debacles. Presumably, these regimes are the ones that give mean-reversion systems the biggest headaches, but I will hopefully be able to test the entire period from 2000 to today.

    Interestingly, on the long term chart frames, I don't think we have had a consistently mean reversion regime for any period of time since 1982 on the daily charts. This goes contrary to known outstanding firms like Rentec that claim to be contrarians. That means that conjecturelly (is that a word?) markets aren't fractal (because Rentec makes most of it's money intra day and therefore there must be lots of mean reversion going on intraday). An interesting observation if true.
     
    #1327     Mar 25, 2010
  8. nitro

    nitro

    Up to the second "NFV" ~1165.01. SPX ~1167.23.

    Almost 100% convergence. I didn't think I would see it in my lifetime.
     
    #1328     Mar 25, 2010
  9. nitro

    nitro

    100% wrong. I honestly didn't think there was any edge short, but 12 handles ES and 22 NQ down handles later...

     
    #1329     Mar 25, 2010
  10. nitro

    nitro

    "NFV" ~1168
     
    #1330     Mar 26, 2010