Ruby for arranging fencing competitions

Sport fencing competitions are organised in pools, where every fencer
fences every other and the total score is totted up at the end. Sheets
like this are used to keep track of which bouts have been played and
what the score is:
http://www.soton.ac.uk/~fencing/files/pool.pdf
These sheets show an order of bouts for between 4-7 fencers, a typical
number in a pool. An impending competition* got me wondering how one
might use Ruby to generate those lists of bouts. A simple list of what
bouts need to be fought is easy, and I wrote a simple script that
produces an array of the following values for any number of contestants
(in this case 4):

1 - 2
1 - 3
1 - 4
2 - 3
2 - 4
3 - 4

The difficult bit is that fencers need to be given a rest between bouts
if possible, so this order won’t do. Simply randomising this array may
still mean someone having to do more than one fight in a row. Not having
a mathematical background I can’t think of a suitable method to approach
this, and if anyone could suggest some hints I’d be interested to know.

  • I can, of course, use the existing pool sheets for the competition;
    this is just for my amusement.

On 10/16/07, Milo T. [email protected] wrote:

(in this case 4):
still mean someone having to do more than one fight in a row. Not having
a mathematical background I can’t think of a suitable method to approach
this, and if anyone could suggest some hints I’d be interested to know.

  • I can, of course, use the existing pool sheets for the competition;
    this is just for my amusement.

Food for RubyQuiz?

Jesus.

Jesús Gabriel y Galán wrote:

Food for RubyQuiz?

Thanks - I’ll take a look.

On 10/16/07, Milo T. [email protected] wrote:

(in this case 4):
still mean someone having to do more than one fight in a row. Not having
a mathematical background I can’t think of a suitable method to approach
this, and if anyone could suggest some hints I’d be interested to know.

How about:

class RoundRobinPairsGenerator

attr_reader :competitors, :full_size

def initialize(n)
@competitors = (1…n).to_a
@full_size = (n*n - n) / 2
end

def pairings
@pairings ||= compute_pairings
end

private

def more_tired(played, pair)
appearances = played.flatten
pair.sort_by {|comp| appearances.select { |e| e == comp
}.length}.last
end

def compute_pairings
pairings = []
picks = competitors.dup
while pairings.length < full_size
trial_pair = picks[0…1].sort
if pairings.include?(trial_pair)
more_tired = more_tired(pairings, trial_pair)
picks.delete(more_tired)
picks << more_tired
else
pairings << trial_pair
picks << picks.shift
picks << picks.shift
end
end
pairings
end
end
puts “For n = 4”
RoundRobinPairsGenerator.new(4).pairings.each { |p| p p }
puts
puts “For n = 10”
RoundRobinPairsGenerator.new(10).pairings.each { |p| p p }

produces:
For n = 4
[1, 2]
[3, 4]
[1, 3]
[2, 4]
[1, 4]
[2, 3]

For n = 10
[1, 2]
[3, 4]
[5, 6]
[7, 8]
[9, 10]
[1, 3]
[4, 5]
[6, 7]
[8, 9]
[2, 10]
[1, 4]
[5, 7]
[8, 10]
[2, 3]
[1, 6]
[5, 9]
[2, 8]
[3, 6]
[7, 9]
[4, 10]
[1, 5]
[2, 7]
[3, 5]
[4, 8]
[6, 9]
[1, 10]
[3, 9]
[2, 4]
[5, 10]
[6, 8]
[1, 9]
[4, 7]
[2, 5]
[3, 8]
[7, 10]
[4, 6]
[2, 9]
[1, 8]
[6, 10]
[3, 7]
[1, 7]
[3, 10]
[2, 6]
[5, 8]
[4, 9]


Rick DeNatale

My blog on Ruby
http://talklikeaduck.denhaven2.com/

On Oct 16, 1:22 pm, “Rick DeNatale” [email protected] wrote:

number in a pool. An impending competition* got me wondering how one
3 - 4

while pairings.length < full_size
end

For n = 4
[5, 6]
[2, 3]
[6, 9]
[4, 6]

Rick DeNatale

My blog on Rubyhttp://talklikeaduck.denhaven2.com/

Nice.

Just in case you (or anyone else) was interested, I quickly threw
together something that would create the simple pairings (1.upto(x) { |
i| (x+1).upto(x) { |j| … } }) and then attempt to re-order it
afterwards. I didn’t have high hopes for it, and it didn’t seem to
work for any number (well, any number > 2).

I haven’t bothered to work out a mathematical proof or anything, but I
believe it’s impossible to generate the pairings that way and then re-
order them. It really should be done this way, taking the rest period
into account when creating the pairings.

Rick Denatale wrote:

def compute_pairings
pairings = []
picks = competitors.dup
while pairings.length < full_size
trial_pair = picks[0…1].sort
if pairings.include?(trial_pair)
more_tired = more_tired(pairings, trial_pair)
picks.delete(more_tired)
picks << more_tired
else
pairings << trial_pair
picks << picks.shift
picks << picks.shift
end
end
pairings
end
end

That is very interesting - I would never have thought of it.
Many thanks!