Correlation Analysis

I've been doing research in correlation analysis for 15 years and can honestly say there is absolutely no consistent correlation between tradable instruments. The key word being "consistent".

At points all correlated instruments diverge.
At points all correlated instruments converge.

I am not saying that spreads aren't tradable because they are and are specific. This is no play area, this is the trading grounds for the focused.
 
ProfLogic:

If your meaning of "consistent" is the entire lifespan of your 15 years of correlation study, then I cede your point.

If your meaning of "consistent" is an annual refreshment of your correlation analyses, then I cede 75% of your point.

If your meaning of "consistent" includes fresher and higher frequency correlation analyses updates, then you are flat wrong.

I've had bankable correlations hold up for 3 years, 3 months, or only 3 weeks.

Case in point: three years ago, there was a very high intraday correlation between crude oil and the euro currency. (at that time crude was also very highly correlated with gold on an intraday basis) When that faded away to the point of statistical elimination, it was soon replaced by the Canadian dollar. When that faded away, the next (and current sweetheart) correlation was the Mexican Peso.

Is there a fundamental reasoning behind the statistical fitness? Sure - in the case of Gold, Euro, and Crude it was flight-to-quality in terms of capital flows. With the CD it was all about Canadian Oil Sands, and with the Peso it's all about PEMEX delivery into Cushing OK and the Gulf Coast refinery complex.

Some very good correlations cannot be explained by fundamentals and so I tend to shy away from them - RBOB Gasoline versus Copper being a case in point.

I could not possibly trade the market without correlations. And I have more consulting work than I can manage at the moment because of correlations.
 
Quote from bone:

ProfLogic:
If your meaning of "consistent" is the entire lifespan of your 15 years of correlation study, then I cede your point.
If your meaning of "consistent" is an annual refreshment of your correlation analyses, then I cede 75% of your point.
If your meaning of "consistent" includes fresher and higher frequency correlation analyses updates, then you are flat wrong.
I've had bankable correlations hold up for 3 years, 3 months, or only 3 weeks.
Case in point: three years ago, there was a very high intraday correlation between crude oil and the euro currency. (at that time crude was also very highly correlated with gold on an intraday basis) When that faded away to the point of statistical elimination, it was soon replaced by the Canadian dollar. When that faded away, the next (and current sweetheart) correlation was the Mexican Peso.
Is there a fundamental reasoning behind the statistical fitness? Sure - in the case of Gold, Euro, and Crude it was flight-to-quality in terms of capital flows. With the CD it was all about Canadian Oil Sands, and with the Peso it's all about PEMEX delivery into Cushing OK and the Gulf Coast refinery complex.
Some very good correlations cannot be explained by fundamentals and so I tend to shy away from them - RBOB Gasoline versus Copper being a case in point.
I could not possibly trade the market without correlations. And I have more consulting work than I can manage at the moment because of correlations.
More...

Bone,

I meant permanent correlations.
There will always be fresh correlations being created and others dying. I steer people away from marrying variable strategies which some do.
Flight-to-quality is a perfect example.
As I said, spreads are definitely a profit bonanza if one can stay focused and stay on one's toes.
 
Quote from ProfLogic:I've been doing research in correlation analysis for 15 years and can honestly say there is absolutely no consistent correlation between tradable instruments. The key word being "consistent".
More...

What an utterly meaningless statement.

This is like saying: I've studied global weather for 15 years... and can honestly say there is no place with 100% sunny days. So sun worshipers are screwed.

Thankfully, there are no ultra-long term market correlations that any idiot with a NetBook could exploit... because that leaves the playing field for the Top 2% with the skills to exploit complex systems of correlations that evolve over time. Complexity and, especially, many years of trading experience... is what thins out the herd... and separates the Pros from the Pretenders.
 
Quote from DeeDeeTwo:

What an utterly meaningless statement.

This is like saying: I've studied global weather for 15 years... and can honestly say there is no place with 100% sunny days. So sun worshipers are screwed.

Thankfully, there are no ultra-long term market correlations that any idiot with a NetBook could exploit... because that leaves the playing field for the Top 2% with the skills to exploit complex systems of correlations that evolve over time. Complexity and, especially, many years of trading experience... is what thins out the herd... and separates the Pros from the Pretenders.
More...

No, it isn't like saying there is no place in the world that has 100% sunny days so sun worshipers are screwed.

It's saying there aren't any places in the world that have sunny days everyday but there are places that have sunny days 90% (Aussy desert) of the time so if you MUST be in the sun be sure to pay attention to the forecast everyday and don't ASSUME the sun will shine everyday.
or
Sort of like showing a trader that a stop goes at specific location because if a reversal happens before your target oscillation is breached it is a signal of a change in directional strength and then the traders "feels" they are smarter than the market so they do not exit at that failed oscillation turning a very small loss into a huge loss by hanging on to the trade for months instead of exiting and taking advantage of the reversing strength. Reversing would have produced huge profits as well.
or
Like a cop in Chicago telling someone not to walk from the United Center to the Silversmith downtown because there is a good possibility they will get mugged on the way. They don't listen and makes it to hotel where the idiot says, "See the cop was a fool. I can walk this anytime I want." After the next Bulls game the idiot tries it again and dies in a mugging.
or
A professional baseball player takes batting practice swinging and hitting tens of thousands of balls during his career. During that time he learns what a "perfect pitch" is TO HIM. He is constantly searching for that pitch but he WILL swing at balls that are close to being perfect because he knows he can't survive as a professional only waiting for that one "perfect pitch".

What separates the professionals from the hopeful are the professionals fully understanding environment they are operating in, they understand the rules that govern that immediate environment & finally the consistent results from being able to succeed long term in that environment.
 
Quote from TraderSystem:
:confused: Your post looks confusing. Please explain.
More...

Thanks for pointing it out, TraderSystem.
Apologies, the links did not get posted properly. Have corrected and re-posted the post as below:

TradersStudio does have a provision for measuring correlation between markets and systems and has correlation reports. Please see the following threads:


Interesting examples:


Thanks. :)
 
Quote from DeeDeeTwo:

What an utterly meaningless statement.

This is like saying: I've studied global weather for 15 years... and can honestly say there is no place with 100% sunny days. So sun worshipers are screwed.

Thankfully, there are no ultra-long term market correlations that any idiot with a NetBook could exploit... because that leaves the playing field for the Top 2% with the skills to exploit complex systems of correlations that evolve over time. Complexity and, especially, many years of trading experience... is what thins out the herd... and separates the Pros from the Pretenders.
More...
+1
 
Quote from JaiSreeram:

Thanks for pointing it out, TraderSystem.
Apologies, the links did not get posted properly. Have corrected and re-posted the post as below:

TradersStudio does have a provision for measuring correlation between markets and systems and has correlation reports. Please see the following threads:


Interesting examples:


Thanks. :)
More...

Thanks JaiSreeram for reposting the post. It is clear now. :)
 
You must log in or register to reply here.
Back
Top