[r-t] A new Spliced Surprise Major canon

[Don Morrison]

On Wed, Mar 6, 2013 at 4:42 PM, Mark Davies <mark at snowtiger.net> wrote:
> I've also got a prototype which is meant to "double-up" the ATW of a single
> method, in preparation for replacement of half its leads by a new method
> during a second phase,

Does this replacement have to be by a method with the same lead end order?
And presumably "doubled-up" is more than just every place bell appears
twice, since you need two disjoint sets of leads, each of which is atw, right?
Or am I missing something?

> Well, I'm guessing the stochastic approach is new, although happy to be
> corrected on that front

I'm not sure it's entirely new. Something I did a decade ago was
vaguely similar, though neither as thorough nor leading to anything as
impressive as your results. I assembled a peal of spliced royal, and
then had my software pick two courses, or some selected leads of three
courses, which were essentially Xed out and then replaced using a
depth first, exhaustive search (of orderings of the methods, the calls
entering and leaving those courses being fixed) trying to improve
assorted properties: atw, as the starting composition, assembled by
hand, was missing two or three place bells; method balance; and method
distribution. While the base calling was, compared to its competitors
at the time, relatively little bellfriendly, I wasn't
particularly try to maximize any musical properties with the
stochastic stuff. And, of course, repeat, doing some hill-climbing as
it tried replacing different bits of the evolving composition. Unlike
you, the whole exercise also started with a specific set of methods,
selected for popularity not musical possibilities. The restriction to
replacing two courses, or selected leads from three, was to keep the
resulting search tractable. The end result was

http://ringing.org/main/pages/dfm/royal/spliced#3109

While I submitted it for publication, I don't think the CCCBR
Compositions Committee ever saw fit to do so, but here's the rationale
for I submitted along with it:

"The goals are probably pretty obvious, but in case a reviewer is
interested, they were:

  - more concentration on little bell rollups than has been usual
    in compositions of these four methods in the past
  - minimum length (as have both Kippin's and Allton's)
  - even method balance (again, as have both Kippin's and Allton's)
  - high changes of method, with 109 (versus 101 for Kippin's and
    104 for Allton's)
  - good method distribution, with at most 11 leads between leads
    of any method (versus 13 for Kippin's and 17 for Allton's)
  - and a strong finish, ending with the three 65 course heads,
    followed quickly by the 64523 course, and then back rounds and
    the three 56 course heads."

As you can see, it was really intended to be just a slight,
incremental improvement on the standard callings of those four methods
in common use at the time.



--
Don Morrison <dfm at ringing.org>
"It is difficult to appreciate the complexity of the brain
because the numbers are so huge they go well beyond our
everyday experience (unless you are a cosmologist).
          -- Daniel J Levitin, _This is Your Brain on Music_