Rankings, you're doing it wrong

Slide 1

Slide 1 text

Rankings: you’re doing it wrong.

Slide 2

Slide 2 text

João Moura github.com/joaomdmoura twitter.com/joaomdmoura speakerdeck.com/joaomdmoura

Slide 3

Slide 3 text

No content

Slide 4

Slide 4 text

largest city in Brazil southern hemisphere

Slide 5

Slide 5 text

No content

Slide 6

Slide 6 text

No content

Slide 7

Slide 7 text

RubySource

Slide 8

Slide 8 text

No content

Slide 9

Slide 9 text

No content

Slide 10

Slide 10 text

Why? where and how?

Slide 11

Slide 11 text

#Everything is Software

Slide 12

Slide 12 text

Young Entrepreneurs are developing new applications for young audiences

Slide 13

Slide 13 text

Gartner Research: Over 70% of Global 2000 Organisations Will Have at Least One Gamiﬁed Application by 2014 (source: Christy Pettey, Gartner press release, http:/ /www.gartner.com/newsroom/id/1844115)

Slide 14

Slide 14 text

Gartner Research: By 2014, 80% of Current Gamiﬁed Applications will fail to Meet Business Objectives Primarily Due to Poor Design (source: Christy Pettey, Gartner press release, http:/ /www.gartner.com/newsroom/id/2251015)

Slide 15

Slide 15 text

Objective order and distinguish

Slide 16

Slide 16 text

Factor how skilled relevant

Slide 17

Slide 17 text

Ranking Behavior Relevancy

Slide 18

Slide 18 text

Ranking Behavior Relevancy

Slide 19

Slide 19 text

! Problem wrong relavancy!

Slide 20

Slide 20 text

“I want to be the best user ever!”

Slide 21

Slide 21 text

1 2 3 1000 2000 1500

Slide 22

Slide 22 text

1 2 3 1000 2000 1500

Slide 23

Slide 23 text

1 2 3 1000 2000 1500 10

Slide 24

Slide 24 text

1 2 3 4 1000 2000 1500 10

Slide 25

Slide 25 text

1 2 3 4 1000 2000 1500 10

Slide 26

Slide 26 text

1 2 3 4 1000 2000 1500 10

Slide 27

Slide 27 text

Consequence objective X reality

Slide 28

Slide 28 text

Humans Want to prove themselves

Slide 29

Slide 29 text

Skill != history

Slide 30

Slide 30 text

Rankings isn’t child's play

Slide 31

Slide 31 text

Possible algorithms

Slide 32

Slide 32 text

Algorithm is an finite sequence of well defined instructions, that can be executed mechanically in an finit amount of time with an finit effort.

Slide 33

Slide 33 text

Elo rating

Slide 34

Slide 34 text

Rating game

Slide 35

Slide 35 text

Rating game Ra = Rb

Slide 36

Slide 36 text

precision Nø matches

Slide 37

Slide 37 text

Microsoft yeah, Microsoft

Slide 38

Slide 38 text

Microsoft yeah, Microsoft

Slide 39

Slide 39 text

Slide 40

Slide 40 text

TRuESKILL

Slide 41

Slide 41 text

True Skill Elo rating on steroids

Slide 42

Slide 42 text

distrust SKILL rating {

Slide 43

Slide 43 text

distrust USER game SKILL

Slide 44

Slide 44 text

2000 100 1500 500

Slide 45

Slide 45 text

2000 100 1500 500

Slide 46

Slide 46 text

distrust SKILL

Slide 47

Slide 47 text

distrust SKILL

Slide 48

Slide 48 text

distrust SKILL

Slide 49

Slide 49 text

x distrust SKILL challenge

Slide 50

Slide 50 text

WIN LOST difficulty

Slide 51

Slide 51 text

Microsoft Yes, it’s patented

Slide 52

Slide 52 text

Sigma Ruby, Ruby, Ruby ̇

Slide 53

Slide 53 text

Skill Distrust Expectation Assertiveness Considerate

Slide 54

Slide 54 text

100.times { user_1.won (user_1.rating - user_2.rating) user_2.lost (user_2.rating - user_1.rating) user_2.won (user_2.rating - user_3.rating) user_3.lost (user_3.rating - user_2.rating) user_3.won (user_3.rating - user_4.rating) user_4.lost (user_4.rating - user_3.rating) }

Slide 55

Slide 55 text

2.times { user_5.won (user_5.rating - user_2.rating) user_2.lost (user_2.rating - user_5.rating) } 2.times { user_5.won (user_5.rating - user_1.rating) user_1.lost (user_1.rating - user_5.rating) } expect(User.ranking.first).to eq user_5