[maemo-community] STV voting strategy

[From: Benson Mitchell]

On Fri, Mar 20, 2009 at 1:47 PM, Dave Neary <dneary at maemo.org> wrote:
>> Algorithm in sh, grep, wc, and dc:
>> $ for i in 1 2 3 4 5 6 7 8 9; do for j in 1 2 3 4 5 6 7 8 9; do grep
>> $i.\*$j.\*0$ ./election.blt|wc -l; done; echo ++++++++p;done|dc
>
> Interesting indeed. What it doesn't measure is the number of times $i
> appeared on the ballot and $j didn't appear at all, which would also be
> a pairwise win for $i. You also give a big advantage to 1, since all of
> the voting lines start with 1 :)

Ah, you're right. I misunderstood the format. No wonder Jaffa was
doing so well in my "results"...

>
> $ for i in 1 2 3 4 5 6 7 8 9; do for j in 1 2 3 4 5 6 7 8 9; do
>  grep "^1.*"$i".*"$j".*0$" ./election.blt|wc -l;
>  grep "^1[^"$j"]*"$i"[^"$j"]*0$" ./election.blt|wc -l;
> done; echo +++++++++++++++++p;done|dc
>
> That gives me results which differ both from the official results *and*
> from your results in the last 2 positions.

True, although it makes the same accuracy as my summing before; 4/5
right, just a different 4.

BTW, this is simpler, and still fixes those issues :P
$ for i in 1 2 3 4 5 6 7 8 9; do for j in 1 2 3 4 5 6 7 8 9; do
  grep "^1[^"$j"]*"$i".*0$" ./election.blt|wc -l;
done; echo +++++++++p;done|dc

(But something tells me a Perl guru could come along and implement
that in 1/2 the characters and making 1/20 the sense...)

Benson