Weekly P-Rankings Page

Starting with 2010, I am setting up a new weekly award: Highest P-Ranking. It is calculated from Sunday 6:00pm to Sunday 5:59pm.

In this Propability-based ranking system every public ranked game is evaluated. It's quite simple how to get a better ranking: Just win against better players ;-)

Every Monday morning 00:00 MET (in summer MEST) the rankings are reset to initial value of 800. The same time the top players are recorded into a statistics page.

You can watch the Rankings any time by visiting this site and takeing a look on the left side P-Rankings block.

The first 3 places every week will get the awards pointing to player list page:

This ranking system is based on ELO computation from chess tournaments.

Some Terms:

Ro = Rating old
n = number of Opponents
Ri = Rating of individual opponents
Rn = Rating new
P(D) = Propability to win vs one player with Ratingdifference D (see table)
P = Position gained
Pe = Position expected

Formula:
Pe = n - SUM(P(Ro - Ri)) + 1
Rn = Ro + 800 * (Pe - P) / (5 + n)
If Rn is less than 0, it is set to 0

  0-3    50  50     69-76   60  40    146-153  70  30    236-245  80  20    358-374  90  10
  4-10   51  49     77-83   61  39    154-162  71  29    246-256  81  19    375-391  91   9
 11-17   52  48     84-91   62  38    163-170  72  28    257-267  82  18    392-411  92   8
 18-25   53  47     92-98   63  37    171-179  73  27    268-278  83  17    412-432  93   7
 26-32   54  46     99-105  64  36    180-188  74  26    279-290  84  16    433-456  94   6
 33-39   55  45    106-113  65  35    189-197  75  25    291-302  85  15    457-484  95   5
 40-46   56  44    114-121  66  34    198-206  76  24    303-315  86  14    485-517  96   4
 47-53   57  43    122-129  67  33    207-215  77  23    316-328  87  13    518-559  97   3
 54-61   58  42    130-137  68  32    216-225  78  22    329-344  88  12    560-619  98   2
 62-68   59  41    138-145  69  31    226-235  79  21    345-357  89  11    620-735  99   1

In this table, you can see three columns the first indicates the difference between the old rating values, the second indicates the winning percentage for the better player, the third the winning percentage of the weaker player.

The table in fact represents a gaussian distribution function.

Hall of Fame:

2010-02-28 18:00:28 to 2010-03-07 17:59:23
1325 Mazza
1297 GYPCOUNT
1280 Pauley

2010-02-21 18:01:58 to 2010-02-28 17:58:16
1487 dollabillz
1382 Smashley
1309 3DA

2010-02-14 18:00:52 to 2010-02-21 17:58:19
1277 THE J-Dude
1237 PJB172
1218 Smashley

2010-02-07 18:03:32 to 2010-02-14 17:57:47
1377 Pauley
1349 Cyn
1262 Highhat

2010-01-31 18:00:37 to 2010-02-07 17:59:38
1276 Big Will E Style
1243 Replicant
1243 Barbie Allstar

2010-01-24 18:00:55 to 2010-01-31 17:59:52
1389 Cyn
1338 Josef
1298 Milltycoon

2010-01-17 18:00:19 to 2010-01-24 17:56:56
1353 Darth Rellek
1259 George874
1215 JoeDude

2010-01-10 18:02:27 to 2010-01-17 17:57:59
1256 Baden
1218 Highhat
1216 Naraku

2010-01-03 18:00:58 to 2010-01-10 17:58:21
1469 George874
1380 Mazza
1373 Naraku

2009-12-27 18:01:56 to 2010-01-03 17:58:56
1409 *Manimal
1343 Smashley
1287 Starman