2.9310: Sun Jun 10 2018
   - fixed bug found by Robert Urban where the LaguerreFilter filter
     set_days/set_alpha was not working

2.9309: Sun Jun 10 2018
   - fix division by zero bugs in SM/Stochastic; side note:
     What is the correct computation of K and D when H==L or L=0? We can't really know.
     Lane didn't actually give any notes on this afaik. Marcel's notion is to
     just set D and/or K to 100 in such situations — seems reasonable to me.

2.9308: Wed Jun 06 2018
   - It's been a few years since I thought about this project, but Marcel Ebbrecht sent
     a patch to CCI.pm to avoid overwriting the sma in the while loop — an
     artifact that remains untested in t/12_cci.t

2.9307: Fri Feb 12 2016
   - No sense using // in exactly one spot and spoiling the perl5.8 tests for no reason

2.9306: Tue Feb 09 2016
   - patches for https://rt.cpan.org/Ticket/Display.html?id=111834

2.9305: Fri Apr 25 2014
   - I never noticed http://search.cpan.org/perldoc?Math::Business::Stochastic and
     accidentally tried to steal his FCFS namespace.  Woops.  Moved to
     Math::Business::SM::Stochastic.
     Fail.

2.9303: Mon Oct 21 2013
   - lazy test in t/13: https://rt.cpan.org/Ticket/Display.html?id=89645

2.9301: Fri Oct 04 2013
   - added an undocumented (experimental) tag feature,
     descriptors for the models, mainly for graphing purposes.

2.9300_1: Thu Oct 03 2013
   - added ConnorRSI - Still lacks docs and tests
   - added a Bayesian example to the examples directory
   - added a trading strategy example

2.9200: Sun Sep 22 2013
   - added Stochastics

2.9100: Sat Sep 21 2013
   - added a new index (CCI)
   - added an example program showing loading a database with
     MACD and RSI from Forex M1 data.

2.9020: Thu Nov 29 2012
   - Erik Axelkrans discovered a bad bug in the stddev
     computation of bollinger bands.  Thanks!!

2.9015: Sat Jun 09 2012
   - Alan Wee feels Bollinger Bands should support floating point
     arguments.  Probably other modules should do that too...

2.9014: Fri Sep 17 2010
   - fixed a 0/0 error.... clearly since the DI is 0, the DX is
     also 0, even though it's indeterminant form.
   - relicensed as Perl artistic

2.9013: Tue Mar 31 06:58:51 EDT 2009
   - I just noticed that the reason my versions are screwed up is
     that at some point (circa Jul12 2008) I went down in version
     -- foolishly forgetting about floating point versions. Grrz.

2.13: Mon Mar 30 09:54:57 EDT 2009
   - I changed the way the tests worked.  The rand() calls in the
     tests were making the tests fail fairly rarely, but also
     randomly.  I'm hopeful that the rare circumstances where the
     assumptions in the tests fail ... won't happen at all with
     pre-generated data.

2.111: Fri Mar 27 18:35:35 EDT 2009
   - added chorny code to the makefile that makes the repo show
     on cpan.

2.11: Sun Jul 13 13:08:22 EDT 2008
    - fixed the HMA test.  HMA can apparently exceed (or
      under-ceed) input-min-max boundaries.  Huh.

2.10: Sat Jul 12 20:33:50 EDT 2008
    - Gustav <gustavf@gmail.com> found that ParabolicSAR fails to
      accept four tuples that That must have been hard to figure out...
    - I also apparently misspelled tuple everywhere, wow
    - I added a FAQ pointing to Genius Trader

2.9: Wed Jul  2 15:44:30 EDT 2008
    - Added LaguerreFilter (a fascinating take on moving averages
      using digital filter processing)
    - added a test for the LF
    - switched the adaptive diff filter from an average to a
      median (more literally what's in Ehlers' book)

2.8: Sun Jun 30 19:55:50 EDT 2008
    - I axed the start_with functions... Storable's
      freeze()/thaw() should work better anyway.  Choose that.
    - gave instructions on using the dumpers and storable to
      avoid recalcs
    - Added WMA
    - Added the HMA (suggested by John Baker). I'm
      intrigued by it's purported low-lag smoothing abilities..

2.7: Fri Jun 20 16:41:36 EDT 2008
    - my doc rearrange didn't help...  I added NAME sections.
      Might work.

2.7: Thu Jun 19 06:24:27 EDT 2008
    - rearranged the distribution a little... the ::cookbook and
      ::faq pods weren't indexing right on search.cpan.org
    - perl5.6.x needs constant 1.05 (or at least greater than
      1.02) so I can use that nice constant { NAME => 1 } block.

2.6: Wed Jun 18 18:48:20 EDT 2008
    - finished the Parabolic SAR

2.6: Tue Jun 17 07:30:57 EDT 2008
    - skeled a ParabolicSAR module

2.5: Mon Jun 16 19:38:44 EDT 2008
    - screwed up the fix by not realizing that do{} evaluates to
      a scalar in 5.6 (wow).  So my 363 element list was setting
      my @close = do "file"; like it was @close=(363).
      Nice.  Oh well, 5.6 is ancient history.

2.4: Sun Jun 15 22:04:12 EDT 2008
    - fixed a do {} useless context "bug" so the amit test will
      pass on perl 5.6
    - I have been version-ing each module up each time I change
      the dist version, but I'm going to stop doing that now.

2.3: Sun Jun 15 19:01:53 EDT 2008
    - Doc bug...  Forgot to add
      set_alpha(14) = set_days(27)

2.2: Sat Jun 14 20:22:05 EDT 2008
    - RSI(14) really means EMA[27]/EMA[27]
      (indirectly suggested by Amit Dutt)

2.1: Mon Jun  9 13:39:44 EDT 2008
    - responding to a bug report from Greg Jessup
      ... seems the example in the MACD module is frelled in
      various ways.   Fixed.
    - improved the t/03_macd test
    - lost the use version since I'm using a floating point
      version

2.0: Wed Jun  4 19:19:33 EDT 2008
    - added a wantarray return to MACD query
    - fixed minor return() bug in MACD (boolean $m verses boolean
      defined($m))
    - minor doc-fix for RSI
    - unified versions

1.4: Mon Jun  2 06:52:18 EDT 2008
    - more documentation: what are these for?
    - finished the DMI start_with() function

1.4: Sun Jun  1 18:21:49 EDT 2008
    - fixed an irritating pseudo-bug in the ADX logic.
    - fixed a divide by zero bug in ADX... grrz.

1.3: Sat May 31 18:43:46 EDT 2008
    - doc bugs in the Bollinger Bands
    - added $class->recommended instantiaters.
    - sqrt(-2.07093e-20) in Bollinger Bands
    - ATR added
    - DMI added

1.2: Wed May 28 16:13:43 EDT 2008
    - version hell still

1.1: Wed May 28 15:29:56 EDT 2008
    - Yeah, my versions were too low on two modules.  Wow.

1.0: Wed May 28 07:36:23 EDT 2008
    - I combined all the Math::Business::* modules I wrote into
      one distribution
    - I made EMA faster and "better" and fixed the while(insert)
      where insert is 0
    - Refactored SMA to be smarter for long DBI calculations and
      possibly less smart for huge multi scalar inserts
    - Authored the RSI module to my tastes
    - Authored the Bollinger Bands module