jsomers

Words = File.new("./pruned-words.txt").read.split("\n")

class String
  def starts_with?(str)
    self[0..str.length - 1] == str
  end
end

# Players can never play "bones", e.g., since the game would end with "bone".
# The next two functions remove all 128,832 such extraneous words from our word list.

jsomers

# The Rules of Mindy Coat
# =======================
# => Four players, opposite sides are in teams.
# => Deal 5 cards each, one-by-one starting left of the dealer.
# => Left of the dealer calls trump looking at those 5.
# => Dealer deals the last 8 cards out (starts on the left).
# => Highest card of the trick wins, unless a trump is thrown, in which case the highest trump wins.
# => Team that takes the most tens wins, unless each takes two, in which case the total number of tricks is counted.
# => (Must follow suit unless you don't have it.)

jsomers

words = File.open("./TWL06.txt").read.downcase.gsub("\r", "").split("\n")
words.select! {|w| w.length == 6}

# Approach:
# 1. Generate an index that maps 2-templates (like "_l___x") to words ("climax").
# 2. Find 2-templates that map to just one word.

def templates(word)
  "Finds all fifteen (6! / 4! * 2!) 2-templates for a given word."
  inds = []; (0..4).collect {|i| (i + 1..5).each {|j| inds << [i, j]}}

jsomers

index = {}

def nxt(n)
  (n % 2 == 0 ? n / 2 : 3 * n + 1)
end

(2..1_000_000).each do |n|
  c = n + 0
  ct = 0
  while n > 1

jsomers

# [n, tot, -AA, 0Ls, -A, 1L-AA, 1L-A, 1LnotA]
lst = [1, 3, 0, 2, 1, 0, 0, 1]
while lst[0] < 30
  n, t, a, b, c, d, e, f = lst
  lst = [n + 1, 2 * t + b - a, c, 2 * b - a + d, t - (a + c), e, f, t]
end
p lst[1] # => 1918080160

jsomers

N = 6
R = 2

def masks
  @masks ||= (0..N - 1).to_a.combination(N - R).to_a
end

def templates(word)
  masks.inject([]) do |templates, indexes_to_hide|
    arr = word.split('')

jsomers

=begin

Most of the work here is in generating a sort of matrix of mutually
compatible "p-bricksets," or single-row arrangements of bricks. There are
something like 3,300 of these total.

Once you have that matrix - what in the code I call the "compatimap" - all you
do is exhaustively "expand" each p-brickset ten times (for the ten rows).

It helps to imagine the space of possible walls as a tree in which nodes are

jsomers

# See http://www.glicko.net/glicko/glicko.pdf

DEFAULT_RATING = 1500.0
AVERAGE_RD = 50.0 # Could be determined empirically.
PERIODS_TO_RESET = 400.0 # If you sit out 400 games your rating is as uncertain as a new player's.
UNRATED_RD = 350.0
MINIMUM_RD = 30.0
C = Math::sqrt((UNRATED_RD ** 2 - AVERAGE_RD ** 2) / PERIODS_TO_RESET)
Q = 0.0057565

jsomers

<!-- TODOS

* More feedback about when the last save was, and what type it was.
* Word count responding to highlights initiated via the keyboard.

-->
<link type="text/css" href="./pad.css" rel="stylesheet" />
<script src="./diff_match_patch.js" type="text/javascript"></script>
<script src="./jquery-1.3.2.min.js" type="text/javascript"></script>
<script src="./field-selection.js" type="text/javascript"></script>

jsomers

$rom_counter = 0
$ram_counter = 16
$symbols = {
  "SP" => "0",
  "LCL" => "1",
  "ARG" => "10",
  "THIS" => "11",
  "THAT" => "100",
  "R0" => "0",
  "R1" => "1",