Version of true_skill() Python algorithm that handles draws.

trueskill.py

#!/usr/bin/env python2.6
from __future__ import print_function
import sys

from scipy.stats.distributions import norm as scipy_norm

beta = 25/6
gamma = 25/300
epsilon = 0.08

norm = scipy_norm()

def pdf(x):
    return norm.pdf(x)

def cdf(x):
    return norm.cdf(x)

def Vwin(t, e):
    return pdf(t - e) / cdf(t - e)

def Wwin(t, e):
    return Vwin(t, e) * (Vwin(t, e) + t - e)

def Vdraw(t, e):
    return (pdf(-e - t) - pdf(e - t)) / (cdf(e - t) - cdf(-e - t))

def Wdraw(t, e):
    return Vdraw(t, e)**2 + ((e - t)*pdf(e - t) + (e + t)*pdf(e + t)) / (cdf(e - t) - cdf(-e - t))

def true_skill(winner, loser, draw=False):
    muw, sigmaw = winner
    mul, sigmal = loser
    
    c = (2*beta**2 + sigmaw**2 + sigmal**2)**.5
    t = (muw - mul) / c
    e = epsilon / c

    if draw:
        W = Wdraw
        V = Vdraw
    else:
        W = Wwin
        V = Vwin
        
    sigmaw_new = (sigmaw**2 * (1 - (sigmaw**2) / (c**2)*W(t, e)) + gamma**2)**.5
    sigmal_new = (sigmal**2 * (1 - (sigmal**2) / (c**2)*W(t, e)) + gamma**2)**.5
    muw_new = (muw + sigmaw**2/c * V(t,e))
    mul_new = (mul - sigmal**2/c * V(t,e))
    
    winner = muw_new, sigmaw_new
    loser = mul_new, sigmal_new
    
    return winner, loser
  
def update(squads, winner, loser, draw):
    winner_stats = squads.get(winner, (25, 25/3))
    loser_stats = squads.get(loser, (25, 25/3))
    squads[winner], squads[loser] = true_skill(winner_stats, loser_stats, draw)

def rank(squads, squad):
    return squads[squad][0] - 3*squads[squad][1]

def main(argv):
    games = [
        ("PK", "SF",   False), ("RT", "XT",  False), ("ACE", "BS", False), ("SRM", "GB",  False), #day 1
        ("SysX", "XT", False), ("PK", "RT",  False), ("SF", "GB",  False), ("SRM", "BS",  False), #day 2
        ("SysX", "GB", False), ("RT", "BS",  False), ("SF", "SRM", False), ("XT", "ACE",  False), #day 3
        ("SysX", "PK", False), ("RT", "SF",  False), ("XT", "BS",  False), ("GB", "SRM",  False), #day 4
        ("SysX", "RT", False), ("PK", "ACE", False), ("XT", "SF",  False), ("BS", "GB",   False), #day 5
        ("SysX", "SF", False), ("RT", "GB",  False), ("XT", "PK",  False), ("ACE", "SRM", False), #day 6
        ("SysX", "ACE",False), ("RT", "SF",  False), ("XT", "PK",  False), ("BS", "SRM",  False), #day 7
        ("SysX", "BS", False), ("PK", "GB",  False), ("RT", "SRM", False), ("ACE", "SF",  True ), #day 8
        ("SysX", "RT", False), ("PK", "SRM", False), ("XT", "SF",  False), ("BS", "ACE",  False), #day 9
        ("PK", "SF",   False), ("XT", "SysX",False), ("SRM", "ACE",False), ("BS", "GB",   False), #day 10
        ("SysX", "SF", False), ("PK", "RT",  False), ("XT", "SRM", False), ("ACE", "GB",  False), #day 11
        ("SysX", "SRM",False), ("PK", "BS",  False), ("RT", "ACE", False), ("XT", "GB",   False), #day 12
        ("SysX", "PK", False), ("RT", "XT",  False), ("BS", "ACE", False), ("GB", "ACE",  False) ] #day 13

    squads = {}
    
    for winner, loser, draw in games:
        update(squads, winner, loser, draw)
    
    for squad in sorted(squads, key=lambda squad: rank(squads, squad)):
        mu = squads[squad][0]
        sigma = squads[squad][1]
        print("%6s R: %5.2f mu: %5.2f sigma: %5.2f" % (squad, rank(squads, squad), mu, sigma))
    
    return 0

if __name__ == "__main__":
    exit_code = main(sys.argv)
    sys.exit(exit_code)

Got formulas for draws from pages 30 and 33 of:
http://dl.dropbox.com/u/1083108/Moserware/Skill/The%20Math%20Behind%20TrueSkill.pdf