Kudos to MMs

MADASINHATTER · Sep 7, 2011

Super cala fragilistic ex b ala doses
nitro,just the sound of it is really quite atrocious
when you need to buy you sell
in calibrated doses
super cala fragilistic,help me holy moses..

nitro · Sep 7, 2011

Quote from shortie:

could the model be in fact backward looking? a lag of ~30 days? have you run tests to see if it is a good predictor of past market moves?

It could very well be, but this is a very difficult thing to prove with so many moving parts.

There is a mathematical tool that can aid in discovering lag relations, the Auto-correlation, but i don't have historical data. That is why this is so hard.

This is a very whippy market, and trend systems tend to do badly in them. Still, I am learning tons.

Maverick74 · Sep 7, 2011

Quote from nitro:

It could very well be, but this is a very difficult thing to prove with so many moving parts.

There is a mathematical tool that can aid in discovering lag relations, the Auto-correlation, but i don't have historical data. That is why this is so hard.

This is a very whippy market, and trend systems tend to do badly in them. Still, I am learning tons.

Nitro. This is absurd. Why don't you just paper trade this thing. You obviously don't have all the kinks worked out. Just paper trade it until you know it's consistent.

arbs-r-us · Sep 7, 2011

Quote from MADASINHATTER:

Super cala fragilistic ex b ala doses
nitro,just the sound of it is really quite atrocious
when you need to buy you sell
in calibrated doses
super cala fragilistic,help me holy moses..

This is AWESOME!

arbs-r-us · Sep 7, 2011

Quote from nitro:

It could very well be, but this is a very difficult thing to prove with so many moving parts.

There is a mathematical tool that can aid in discovering lag relations, the Auto-correlation, but i don't have historical data. That is why this is so hard.

This is a very whippy market, and trend systems tend to do badly in them. Still, I am learning tons.

You don't have a trend system!!! IT'S A SICK, TWISTED, ABSURD MEAN REVERSION TWIST OF SANITY!!!!

Maverick74 · Sep 7, 2011

Quote from MADASINHATTER:

Super cala fragilistic ex b ala doses
nitro,just the sound of it is really quite atrocious
when you need to buy you sell
in calibrated doses
super cala fragilistic,help me holy moses..

Yes, you are very talented. LOL.

Pekelo · Sep 7, 2011

I was right about Nitro 3 years ago, and recently a month ago:

Quote from Pekelo:

I suggest a -1 multiplier.

His system is actually gold, pure gold. He just needs to evaluate the "edge" with an inverse look. Thus when the FV goes BELOW the market, he should go LONG, and vica versa.

In the last 4 weeks the market had one of its worse August, the FV difference was hugely positive. On the day when it went negative the market had one of its biggest rally. Quite simple really...

No charge for the advice....

Maverick74 · Sep 7, 2011

Quote from Pekelo:

I was right about Nitro 3 years ago, and recently a month ago:

His system is actually gold, pure gold. He just needs to evaluate the "edge" with an inverse look. Thus when the FV goes BELOW the market, he should go LONG, and vica versa.

In the last 4 weeks the market had one of its worse August, the FV difference was hugely positive. On the day when it went negative the market had one of its biggest rally. Quite simple really...

No charge for the advice....

That's two data points. Not exactly statistically significant.

stock777 · Sep 7, 2011

by that standard, you'd want to hire a few 'tards to trade for ya.

you should have no problem finding some on et. throw a dart.

nitro · Sep 8, 2011

Calibrated FV, 1140.88. FV 1290.68.

Remember, the difference between FV and Calibrated FV (CFV) is that CFV attempts a correction term to FV that attempts to value the probability of a recession. When the chance of recession is 0, CFV will converge or even go above FV, what people will term "Risk on".

Some people have suggested that the best estimate of "FV" is not the ends, but the average of the two, or "FV" = (FV - CFV) / 2. That is an interesting idea. Recall that a simple moving average if a FIR

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

and we can back out the weights by the following equation:

SPX = xFV - yCFV

where x, y are the weights of some FIR average. That is more sophisticated than a simple MA, and may be worth backing out every day.