Kudos to MMs

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

  1. nitro

    nitro

    Calibrated FV 1284.23. FV 1363.98.

    This assumes a ~20% chance of recession.
     
    #3531     Aug 31, 2011
  2. nitro

    nitro

    Notice how violently Calibrated FV jumps around. This is an artifact of the uncertainty.
     
    #3532     Aug 31, 2011
  3. Pekelo

    Pekelo

    Stats like this are 78.53% worthless...

    Alright, I have just made that up. They are 100% worthless....
     
    #3533     Aug 31, 2011
  4. nitro

    nitro

    Calibrated FV 1249.31. FV 1378.34.

    Odds of recesion went up a bit from the previous post, at about 30%. So note that even though raw FV is higher, the chance of recession bites deeper into it and therefore calibrated FV is lower.
     
    #3534     Sep 1, 2011
  5. nitro

    nitro

    Who do I use for next quarter GDP estimates? For example, JPM:

    http://www.bloomberg.com/news/2011-...gdp-forecast-to-1-5-on-spending-slowdown.html

    or

    Roubini gives a 60% chance or recession, while Seigel thinks it is probably 10%.

    http://finance.yahoo.com/blogs/dail...-roubini-sees-60-chance-jeremy-144725583.html

    It seems that there are many choices. I have settled on a way where I sort of back it out of the market itself. Thing is, then I can't update it on new indicator numbers news that I posted before. Further, when the concept is fully accepted as a possible reality by the market, the model becomes more accurate, but early on it may lag since only the smart money has the information flow initially, and that is very hard to back out because it is such a weak signal. Still, I feel this is a good way for now, although it is far more volatile than if I did what you are suggesting.

    What do you mean by flat?
     
    #3535     Sep 1, 2011
  6. by flat I mean no growth, it looks like you could use a scoring model, for example if you got out of the mkt when gdp went negative oct 2008 you would have saved your clients a lot of money, you should have entered long q1 2010 so around 1100.
    Obviously this should not be the whole model, but maybe a scored system that updates once a quarter.
     
    #3536     Sep 1, 2011
  7. gdp spy
     
    • gdp.jpg
      File size:
      90.4 KB
      Views:
      50
    #3537     Sep 1, 2011
  8. nitro

    nitro

    #3538     Sep 1, 2011
  9. nitro

    nitro

    One more note on this chart. Look at 2008:Q1. Compare with 2011:Q2. The GDP number is almost identical, whereas every other metric in 2011 is stronger. Yet, look at the SPY level, we are lower in 2011:Q2 than in 2008:Q1.

    Clearly there is something that is not captured by these fundamentals that is captured by price. It would be interesting to dig deeper - I am guessing business cycle, but that is far too vague...

    I really like this chart.
     
    #3539     Sep 1, 2011
  10. nitro

    nitro

    Yet one more note on that chart (you can tell I am enamored). I believe that Rentec's success is nothing more than taking APT to it's logical extreme, perhaps with very careful statistical reasoning.

    Recall that APT sets up giant matrices of factors, and computes which factors affect which stocks. The weighting then is the holy grail. This chart is a miniature version of inputs that would go into an APT matrix.

    http://en.wikipedia.org/wiki/Arbitrage_pricing_theory

    APT is a subset of models called affine,

    http://en.wikipedia.org/wiki/Affine_arithmetic
     
    #3540     Sep 1, 2011