RSI Correlations

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
DPH
Roundtable Knight
Roundtable Knight
Posts: 115
Joined: Tue Nov 17, 2009 1:07 am
Location: Central Pennsylvania
Contact:

RSI Correlations

Post by DPH » Sat Sep 29, 2018 12:59 am

I had reason to want & see this tonight. Probably useless to most but maybe a few interesting bytes to some.

(5 years approx, 14 period RSI)

Attached....

I hope to stimulate some conversation about atypical & useful correlations anyone has (or wants) to experiment with.

Thanks,
Dean Hoffman
Attachments
RSI Corr Grid.csv
(32.86 KiB) Downloaded 12 times

marriot
Roundtable Knight
Roundtable Knight
Posts: 361
Joined: Thu Nov 20, 2008 3:02 am

Re: RSI Correlations

Post by marriot » Sun Sep 30, 2018 12:42 pm

Years ago, someone here said:
Correlation exist, till there are not.
Well, i think that FV and TY will be always correlated.
IWe sow long period of correletinos between USD and Stock markets.

I would try to enter Long on FV when TY give me a Buy order.
Obviusly there are other way to test correlation.
But the point is how to test it.
It is clear to me that "best "parameters changes with time.
A system optimized from 2006 will give better results then one optimized from 1980.
But best parameters from 2014 will give horrible results.
I suppose that correlation will be the same.
We want to trade what is going to do better, using less loockback that is possible.
But what is the best lookback and what are the parameters that will give the best?

You want to imagine to be in 1990.
You find your best entry-exit and see what's happen in the next one or two months.
Save parameters and results.
Test again starting from 1991 and save again results.
In a couple of years you will have a clear picture.
In other words, i think that the instrument we have today for optimize are outdated.
And it would be time to jump to an higher level.

DPH
Roundtable Knight
Roundtable Knight
Posts: 115
Joined: Tue Nov 17, 2009 1:07 am
Location: Central Pennsylvania
Contact:

Re: RSI Correlations

Post by DPH » Mon Oct 01, 2018 5:06 pm

No doubt correlations are a moving target and why the review mirror has limitations (not useless, but limitations). My interest lies in seemingly different applications of correlation than you might be referring. For example, it’s been reported that some hedge funds are using satellites to correlate how full parking lots in malls are relative to spending (leading indicator?) Or the likelihood of a corporate takeover if Corp A’s private jet flies repeatedly to Corp b’s city and they seem like a potentially good fit (think Moline, Illinois). In other words, I am past simply TY is correlated to FV, I am (trying to) look much deeper. That said, I was able to get an instant 20% MAR ratio jump on a system I had worked to improve for a long time merely by using a unique method to create group risk sectors (beyond simply the standard currencies, meats, metals, grains, and so on). Still lots of gold in them hills in my opinion.

That's why I say the "atypical" types.

Post Reply