Statistics question

B

bozwood

Is there a way to normalize the slope of the linear regression of price data so that different units from different stocks/commodities can be compared?

Thanks in advance,
Jonathan
 
Quote from bozwood:

Is there a way to normalize the slope of the linear regression of price data so that different units from different stocks/commodities can be compared?

Thanks in advance,
Jonathan
More...

I'm certainly no stats pro, but can't you do it in %?

- The New Guy
 
Quote from thenewguy:

I'm certainly no stats pro, but can't you do it in %?

- The New Guy
More...

Heck, I'm not even sure if I understand what you're asking, but if you're asking what I think you're asking and you have access to bloomberg, type HRA <stock> <stock> (or something like that, not sure of the exact syntax) and then there's a setting for something like "regression type" and put that on percentage, then you can compare two seperete products based on their regression in % rather than points.

I think this is where you're going...

- The New Guy
 
Quote from thenewguy:

I'm certainly no stats pro, but can't you do it in %?

- The New Guy
More...

I have been playing around with that. Looks like it may work, but I am not sure if it is mathematically sound. Thanks for the suggestion
 
Quote from bozwood:

I have been playing around with that. Looks like it may work, but I am not sure if it is mathematically sound. Thanks for the suggestion
More...

Personally I put almost everything into % before running calculations and then translate that back into point moves for each product. It's always made more sense to me to do it that way to avoid apples/oranges situations...

- The New Guy
 
I think in e-signal you can do a "Standard Deviation" and a "Standard Error" for Linear Regression.

I'm not right in front of e-signal so I could be wrong.

The only other way to do it is to manually do it in Microsoft Excel.

Hope this helps.
 
Quote from skepticaltrader:

I think in e-signal you can do a "Standard Deviation" and a "Standard Error" for Linear Regression.

I'm not right in front of e-signal so I could be wrong.

The only other way to do it is to manually do it in Microsoft Excel.

Hope this helps.
More...

Actually, I am trying to do it in excel. Any ideas?

Thanks
 
Quote from thenewguy:

Heck, I'm not even sure if I understand what you're asking, but if you're asking what I think you're asking and you have access to bloomberg, type HRA <stock> <stock> (or something like that, not sure of the exact syntax) and then there's a setting for something like "regression type" and put that on percentage, then you can compare two seperete products based on their regression in % rather than points.

I think this is where you're going...

- The New Guy
More...

Just trying to get a comparable slope across different stocks, etc.

Thanks for the ideas
 
Quote from bozwood:

Actually, I am trying to do it in excel. Any ideas?

Thanks
More...

I'm not really sure since I haven't used Excel in so long. I would assume you could do a line graph of prices and if I remember correctly there is a "Standard Deviation" function built into their statistical tool.

I'll try and play with it some either tonight or tomorrow and I'll get back with you on it.

Please let me know if you figure out how to do it in Excel.

Thanks
 
You must log in or register to reply here.
Back
Top