Forums  > Trading  > Better Trading Decisions  
     
Page 2 of 7Goto to page: 1, 2, 3 ... 6, 7 Prev Next
Display using:  

Nonius
Founding Member
Nonius Unbound
Total Posts: 12805
Joined: Mar 2004
 
Posted: 2005-04-29 09:51

fair enough, but let us keep this little puzzle simple, and then we can sup it up to something a little more interesting....

by the way, the percentage of people who get it right IS some knowledge that could help us.


Chiral is Tyler Durden

Johnny
Founding Member

Total Posts: 4333
Joined: May 2004
 
Posted: 2005-04-29 09:51

Absolutely.

 


Johnny Jupiter Radiation Phynancial phorecasting by the stars for the stars

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 09:53
I think I will play in a limited sense. I think I can do a) but I can't do b).

Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

NIP247


Total Posts: 554
Joined: Feb 2005
 
Posted: 2005-04-29 10:01

With the risk of looking like a complete fool, I'm first out with my neck on the chopping block:

Generic trendfollowing model with no curve-fitting generated following stats:

Long trades: Cases: 28, avg: 4,17%, stdev 19,04%, min -15,38%, max 67,52%

Short trades: cases: 27, avg: 4,29%, stdev 10,80%, min -8,02%, max 30,47%

now give me my sentence "It was all random!!!!!"

 


On your straddle, done on the puts, working the calls...

Johnny
Founding Member

Total Posts: 4333
Joined: May 2004
 
Posted: 2005-04-29 10:03
Just to make another blindingly obvious point. One realisation of a random walk might well contain some trend or pattern that appears significant. Therefore it's not a reasonable game to construct one realisation of a process and then ask for opinions on it. In real trading I don't accept trading strategies unless they appear to *work* on a wide range of different markets. I don't want to be making the game unnecessarily complicated, but perhaps we could do something similar by having several (20 or so) realisations of the same process? Just a thought.

Johnny Jupiter Radiation Phynancial phorecasting by the stars for the stars

FDAXHunter
Founding Member

Total Posts: 8372
Joined: Mar 2004
 
Posted: 2005-04-29 10:04
Looks random to me, albeit with a pretty weak RNG.

The Figs Protocol.

Nonius
Founding Member
Nonius Unbound
Total Posts: 12805
Joined: Mar 2004
 
Posted: 2005-04-29 10:07
mkay, you want me to tell you how this was generated?

Chiral is Tyler Durden

NIP247


Total Posts: 554
Joined: Feb 2005
 
Posted: 2005-04-29 10:08
Nonius >> Pleease, I'm dying!

On your straddle, done on the puts, working the calls...

Nonius
Founding Member
Nonius Unbound
Total Posts: 12805
Joined: Mar 2004
 
Posted: 2005-04-29 10:09
if three others say yes, then I'll post the answer.

Chiral is Tyler Durden

RFMontraz
NP Italian Stallion

Total Posts: 2021
Joined: Mar 2004
 
Posted: 2005-04-29 10:13

Knowing that I'll look like a complete fool here's what I get from the RFMTrendFollowingSystemForMongoloids:

One long trade only, only slightly positive. That would lead me to say: random (but the system itself is very basic and analises data in a timeframe that doesn't necesseraly have to work for every market. So if this asset is non random and trends in shorter periods I won't be able to spot it. But the least I tweak the parameters the better - at least that's what I'm comfortable with).

 


Fund Raising and Racketeering, Capital Structure Demolition LLC - What the fuck you mean "not interested" you motherfucker you??

Johnny
Founding Member

Total Posts: 4333
Joined: May 2004
 
Posted: 2005-04-29 10:29

I couldn't see any obvious signs of non-normality, stoch vol or of autocorrelation in either the log returns or the absolute returns. However, there seems perhaps to be some element of trendiness on the *frequency* of around 50 time intervals. It might have been generated by something of the form Return = Drift + Normal Increment where the drift changes sign with probability approx (1/50) on each time step.


Johnny Jupiter Radiation Phynancial phorecasting by the stars for the stars

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 10:34

My answer on prediting, not trading:

y = 5E-09x^3 - 4E-05x^2 + 0.0971x + 75.262
R2 = 0.691

I'd post my Excel graph if I could figure out how to insert it.


Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 10:44

Okay, I put the data on a log scale and did a sixth order polynomial regression over a 1 step series and got this:

y = -4E-19x^6 + 7E-15x^5 - 4E-11x^4 + 1E-07x^3 - 0.0002x^2 + 0.1366x + 77.085

R2 = 0.8324

Using the simple trading rule: "long when one standard deviation below, hold for one standard deviation above, then short, then re-buy when it is one standard deviation below" I cant make any money.


Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

RFMontraz
NP Italian Stallion

Total Posts: 2021
Joined: Mar 2004
 
Posted: 2005-04-29 10:48
Sorry James but you gotta keep in mind that also elements like me are unfortunately hanging around here in the hope of getting a bit smarter (less dumb). What is that you are saying?

Fund Raising and Racketeering, Capital Structure Demolition LLC - What the fuck you mean "not interested" you motherfucker you??

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 10:56

I'm just fitting a curve, that you can then extend into the future. Essentially, it is a 'time series.' I am arguing that the data is best explained by a function of time itself, and that you can set up a simple trading rule around standard deviations above and below this curve.

My time series starts at one (1) and moves in steps of one. It could start at "Jan 1 1985" and move in single days and still have the same curve fitting it.


Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

FDAXHunter
Founding Member

Total Posts: 8372
Joined: Mar 2004
 
Posted: 2005-04-29 11:08
James, are you smoking crack? Polynomials obtained through least squares regression can almost never be used for time series prediction.

The Figs Protocol.

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 11:21
Yep, I know. I was just adding to the chorus of "One realisation of a random walk might well contain some trend or pattern that appears significant." If the random walk has the same time scale for observations, you get my result.

Hard to argue with an R^2 of 0.8324, unless you knew that projecting from same over a one-step series was scary disasterous.

This isn't my sandbox, so I only know the simple sand castle structures that *don't* work from watching other kids gleefully discover then and have it tumble down.

"can almost never be used for time series prediction"

are there ones that can? tell me now!

Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

Johnny
Founding Member

Total Posts: 4333
Joined: May 2004
 
Posted: 2005-04-29 11:23

Time for Nonius to reveal all?

 

 

 

No, Nonius, put that thing away ... Wink


Johnny Jupiter Radiation Phynancial phorecasting by the stars for the stars

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 11:25
"No, Nonius, put that thing away ..."

Given what Nonius often whips out, and what you sometimes grab....

never mind. Wink

Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

FDAXHunter
Founding Member

Total Posts: 8372
Joined: Mar 2004
 
Posted: 2005-04-29 11:26

James: Hard to argue with an R^2 of 0.8324, unless you knew that projecting from same over a one-step series was scary disasterous

Try a 6301 degree polynomial I bet you'd get an even better R^2.


The Figs Protocol.

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 11:28
Big Smile But I think I will try a 50 just for fun, since Johnny found something there.

Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

Nonius
Founding Member
Nonius Unbound
Total Posts: 12805
Joined: Mar 2004
 
Posted: 2005-04-29 11:28
I could fit it perfectly with an infinite series of Genetic Feedback Wanklets.

Chiral is Tyler Durden

James
NP High Priest

Total Posts: 2024
Joined: May 2004
 
Posted: 2005-04-29 11:30
"Genetic Feedback Wanklets"

He's whipping it out!

Johnny, don't touch!

Prior to the publication of the Black-Scholes model in 1973, the quest for a valuation formula that would describe option prices reflected one of the most elusive goals in financial economics. Though much work was done in the 1960s, many of the insights and techniques used to solve the problem were presented or anticipated at the beginning of the twentieth century by Louis Bachelier, an obscure French mathematician. These innovations include the first graphical representation of option pricing, a mathematical description of stock prices utilizing Brownian motion and anticipating the efficient market hypothesis, and the first formal option pricing formula.

Johnny
Founding Member

Total Posts: 4333
Joined: May 2004
 
Posted: 2005-04-29 11:32

James, it's probably best if you could keep your homo-erotic fixations to yourself, don't you think?

 

 


Johnny Jupiter Radiation Phynancial phorecasting by the stars for the stars

Johnny
Founding Member

Total Posts: 4333
Joined: May 2004
 
Posted: 2005-04-29 11:34

So Nonius, you've had your three replies, what's the answer?

 


Johnny Jupiter Radiation Phynancial phorecasting by the stars for the stars
Previous Thread :: Next Thread 
Page 2 of 7Goto to page: 1, 2, 3 ... 6, 7 Prev Next