New signal-to-noise ratio for prices

[peterccarl]

Here's a function for examining trendiness over a timeframe using ATR as a proxy for volatility. I think this fits the package, but let me know if you think it goes elsewhere. pcc

[braverock]

This may reopen the conversation about moving the OHLC functionality 'upstream' from quantmod to TTR.

This pull request can't be merged as is, because quantmod requires TTR, so this creates a circular dependency.

Also, CRAN only wants you to check and require another package if that package is in Suggests. otherwise they want you to have either put it in Requires or Imports

[peterccarl]

Sorry, I'll remove those lines since they are unnecessary if the function
is in ttr. Good catch.

postscript: just to be clear, this shouldn't reopen the conversation about OHLC functionality...

On Apr 7, 2016 6:59 AM, "Brian G. Peterson" notifications@github.com
wrote:

This may reopen the conversation about moving the OHLC functionality
'upstream' from quantmod to TTR.

This pull request can't be merged as is, because quantmod requires TTR, so
this creates a circular dependency.

Also, CRAN only wants you to check and require another package if that
package is in Suggests. otherwise they want you to have either put it in
Requires or Imports

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#26 (comment)

[joshuaulrich]

Some other items:

1. ... are not passed to ATR as documented.
2. The try.xts / reclass paradigm should be used so all time series classes will be supported. See other functions for examples.
3. lag.xts should be called explicitly in case try.xts fails and HLC is a matrix.

[peterccarl]

Fixed, thanks for the pointers.

[joshuaulrich]

reclass isn't called before return and quantmod::Cl still creates a circular dependency.

[peterccarl]

Ok, I missed the real circularity you were referring to. I've removed it.

[joshuaulrich]

Merged via 990da83.