|
PureBytes Links
Trading Reference Links
|
Jim,
Not much to reveal at this time...just throwing ideas around.
High continued correlation which has a momentary deviation would/should pose
an opportunity to trade so long as the deviation is momentary and my thought
was to confirm that correlation by using multiple indicies as confirmation.
As for the system, I have some ideas, but nothing concrete at this time.
Thanks for your insights and advice.
Regards
Chris
-----Original Message-----
>From: Jim Johnson <jejohn@xxxxxxxxxxx>
>To: Chris Emery <cemery1@xxxxxxxxxxxxx>
>Date: Saturday, August 31, 2002 9:42 AM
>Subject: Re[2]: Correlation of three data series
>
>
>Hello Chris,
>
>If the two indicators showed a high degree of correlation there is no
>advantage (disadvantage) to using both of them. "High" depends on
>what you are trying to do.
>
>Correlation doesn't occur to me as a way to approach this. you could
>simply write your system using one, both or the other indicator and
>see which works best. then confirm that relationship on a hold-out
>sample.
>
>I've seen a system by Murray Ruggiero that is constantly monitoring
>the correlation between two variables, open interest and volume. When
>that correlation over some lookback period is positive, he buys;
>negative he sells.
>
>I don't mean to get you to reveal your whole approach but you
>description doesn't tell me enough about your approach to suggest much
>more.
>
>
>Best regards,
>Jim Johnson mailto:jejohn@xxxxxxxxxxx
>
>--
>Saturday, August 31, 2002, 9:22:52 AM, you wrote:
>
>CE> Jim,
>
>CE> Its part of an idea I am working on but the need is to look at 2
>correlated
>CE> indicies to determine the purchase /short of a correlated stock. In
>order
>CE> to do this I need to insure that there the proper relationship among
>all
>CE> three prior to taking a trade.
>
>CE> Regards,
>CE> Chris
>CE> -----Original Message-----
>CE> From: Jim Johnson <jejohn@xxxxxxxxxxx>
>CE> To: omega-list@xxxxxxxxxx <omega-list@xxxxxxxxxx>; Chris Emery
>CE> <cemery1@xxxxxxxxxxxxx>
>CE> Date: Saturday, August 31, 2002 8:29 AM
>CE> Subject: Re: Correlation of three data series
>
>
>CE> Hello Chris,
>
>CE> The concept of three variable's correlating isn't one I've heard of.
>CE> Obviously what you suggest (an inter-correlation matrix) is doable and
>CE> useful for understanding the co-variation among three or more
>CE> variables. How you use it depends mainly on what it is you are trying
>CE> to do or learn. Tell us more.
>
>
>CE> Best regards,
>CE> Jim Johnson mailto:jejohn@xxxxxxxxxxx
>
>CE> --
>CE> Saturday, August 31, 2002, 8:14:54 AM, you wrote:
>
>CE>> Dear List,
>CE>> Is there a formula for correlating 3 data series?
>CE>> I guess one could take the correlation of datas 1&2, 1&3 and 2&3,
>weight
>CE>> them according to which combinations do not contain the dependant
>CE> variable
>CE>> but I am unsure if this would be the correct way to go about it. Any
>CE>> thoughts would be greatly appreciated.
>CE>> Regards,
>CE>> Chris
>
>
>
>
|