TW Elim Formulas - Current and Proposed
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

ABCDEFGHIJKLMNOPQRSTUVWXYZAA
1
TW Elim Formulas - Current and Proposed
2
INSTRUCTIONS: Modify the test values below to test the formulas.
3
4
Kills1010Formula 1 (current)60023238.7%
Formula 1 explanation
5
Deaths510Formula 2 (proposed)431.689516738.7%
* Kills divided by deaths
6
Average rating killed
300300
* This is multiplied by the average rating of player killed
7
Games played (GP)550Formula 2 explanation
8
GP required for 100% confidence7575
* Round base rating is the average rating of all players killed that round.
9
Won? (1 if true)11
* For every kill, points rise exponentially (killsÂ˛), normalized to 10. (If round is to 5 deaths, kills are worth twice as much.) 1 kill gives only 1 point. 2 gives 4. 3 gives 8, etc.
10
Win bonus % per player @ start22
* Deaths subtract points exponentially (deathsÂ˛)
11
KOs10
* Winning the round gives a bonus multiplied by the amount of people who began the round. (If 2% and there are 5 players, 10% bonus.)
12
KO bonus %22
* KOs give a bonus for each player KOd to the final rating. (If 2% and you KO 3 players, 6% bonus.)
13
Total kills of streaks broken00
* For each streak broken, +% for each kill of each streak.
14
Streak end bonus % (per kill)11
* For each kill in a streak of 3 or more, a bonus is given
15
Best streak00
16
% bonus per kill for max streak of 3+22
17
Playing to # deaths510
* Playing to a number of deaths less than 10 gives a very slightly reduced rating, as it's slightly easier to go 10-4 in a game to 5 than 20-8 in a game to 10.
18
Num playing at start
55
* Games with a smaller number of starting players reward slightly less than larger games, to reduce the utility of feeding. (Avg rating also prevents this)
19
20
Starting rating:300
21
22
After every round, the round rating is averaged into the current player rating. Then this rating is multiplied by confidence, resulting in the adjusted rating.
23
24
Confidence
25
Players are required to be rated X times/play X games before the bot considers the rating reliable. Before this, their rating is reduced by a confidence factor.
26
Once games played reaches X (proposed value=50), they are at 100% confidence, and suffer no negative penalty to their adjusted rating.
27
28
Formula for confidence with 1000 games:
(Math.log(games) / Math.log(2)) / 10.0
29
Formula for confidence with 50 games:
(Math.log(games) / Math.log(2)) / 6.0
< Currently using this
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100