Correlations revisited ... - page 7

 
yourthebest:

why not? i'm not a coder, Try it! http://math.stackexchange.com/questions/367743/standardizing-a-random-variable-that-is-normally-distributed just an interesting article... i find it fascinating that if you had normally distributed or normalized data AB INIITIO what one could do..

more accuracy in all the indicators and opens up a whole new world.... but how best to standardize/normalize? i'm just thinking out of the box out loud. don't take it too seriously here. I'm the furthest thing from and expert.

Thanks so much

yourthebest

It can be done, but the values of the indicator would be completely different then

 
mladen:

yourthebest

It can be done, but the values of the indicator would be completely different then

better or worse? maybe ccfp is better to apply to .... the correlation indie is all about follow the leader one turns the other will follow....currencies move in zero sum to each other, IE in a cyclical way. when euro goes up something else must come down. Also, price repeats itself. that is potent powerful info to know imho. so, may be better to have it scaled linearly or normally distributed or coming from point of view where all currencies are pegged to an absolute not the floating value of other currecnies...how can we compare euro at value of 1.1 and gj at value of 250? that makes any sense at all? for comparison sakes we have to have apples to apples abi initio.. i'm thinking of CSM also. again simply some ideas that my my mind has wondering about, for your considerations, you obviously would know best what to do ...please do not waste time unless you think it has merit. if it pleases you, and you make something, i'd be grateful to test a copy. Thanks Yourthebest

 
yourthebest:
better or worse? maybe ccfp is better to apply to .... the correlation indie is all about follow the leader one turns the other will follow....currencies move in zero sum to each other, IE in a cyclical way. when euro goes up something else must come down. Also, price repeats itself. that is potent powerful info to know imho. so, may be better to have it scaled linearly or normally distributed or coming from point of view where all currencies are pegged to an absolute not the floating value of other currecnies...how can we compare euro at value of 1.1 and gj at value of 250? that makes any sense at all? for comparison sakes we have to have apples to apples abi initio.. i'm thinking of CSM also. again simply some ideas that my my mind has wondering about, for your considerations, you obviously would know best what to do ...please do not waste time unless you think it has merit. if it pleases you, and you make something, i'd be grateful to test a copy. Thanks Yourthebest

I could not preach what is better or worse, but since the ccfp is using what I told a "strange momentum" I an not sure what would the combination of stochastic values look like. Will have to test it because predicting what the value will look like in a multi symbol combination is never 100% foreseeable

 
mladen:
I could not preach what is better or worse, but since the ccfp is using what I told a "strange momentum" I an not sure what would the combination of stochastic values look like. Will have to test it because predicting what the value will look like in a multi symbol combination is never 100% foreseeable

How can stochastic be combined?

 
sebastianK:
How can stochastic be combined?

I don't think it can

 
morro:
I don't think it can

The stochastic could be inverted for some symbols, but then the sign for symbols must be known

 
sebastianK:
The stochastic could be inverted for some symbols, but then the sign for symbols must be known

If it can not be done automatically, then it is the same as with the it can not. Better not ...

 

It occurred to me that it should be posted here (due to its nature)This is a Spearman rank (auto)correlation that uses floating levels


 
mladen:

It occurred to me that it should be posted here (due to its nature)This is a Spearman rank (auto)correlation that uses floating levels


Dearest MLADEN

Thanks for adding fl,looking pretty.

what do you means for (auto) ?

regards


 
mntiwana:

Dearest MLADEN

Thanks for adding fl,looking pretty.

what do you means for (auto) ?

regards


mntiwana

The "auto" comes from a fact that this Spearman rank correlation is not calculating correlation to some other symbol or some other set of data, but to the data from the same symbol and time frame, making it a correlation it itself, sort of speaking. Real correlation would be if it was comparing data of two symbols, for example