gamebalancing - stardock.cs.virginia.edu · 2015-04-13 · balancing# •...

54
CS 2501 Game Balancing CS 2501 – Intro to Game Programming and Design Credit: Some slide material courtesy Walker White (Cornell)

Upload: others

Post on 27-Jun-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Game  Balancing  

CS  2501  –  Intro  to  Game  Programming  and  Design        

Credit:  Some  slide  material  courtesy  Walker  White  (Cornell)  

Page 2: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Dungeons  and  Dragons  •  D&D  is  a  fantasy  roll  playing  system  •  Dungeon  Masters  run  (and  someCmes  create)  campaigns  for  players  to  experience  

•  These  campaigns  have  several  aspects  –  Roll  playing  –  Skill  challenges  –  Encounters  

• We  will  look  at  some  simple  balancing  of  these  aspects  

2

Page 3: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

ProbabiliCes  of  D&D  •  Skill  Challenge  •  Example:  The  player  characters  (PCs)  have  come  upon  a  long  wall  that  encompasses  a  compound  they  are  trying  to  enter  

•  The  wall  is  20  feet  tall  and  1  foot  thick  • What  are  some  ways  to  overcome  the  wall?  •  How  hard  should  it  be  for  the  PCs  to  overcome  the  wall?  

 3

Page 4: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

ProbabiliCes  of  D&D  •  How  hard  should  it  be  for  the  PCs  to  overcome  the  wall?  

• We  approximate  this  in  the  game  world  using  a  Difficulty  Class  (DC)  

 

4

Level Easy Moderate Hard 7 11 16 23 8 12 16 23 9 12 17 25 10 13 18 26 11 13 19 27

Page 5: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

ProbabiliCes  of  D&D  •  Easy  –  not  trivial,  but  simple;  reasonable  challenge  for  untrained  character  

• Medium  –  requires  training,  ability,  or  luck  •  Hard  –  designed  to  test  characters  focused  on  a  skill  

5

Level Easy Moderate Hard 7 11 16 23 8 12 16 23 9 12 17 25 10 13 18 26 11 13 19 27

Page 6: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Balancing  •  Balancing  a  game  is  can  be  quite  the  black  art  •  A  typical  player  playing  a  game  involves  intuiCon,  fantasy,  and  luck  –  it’s  qualitaCve  

•  A  game  designer  playing  a  game…  it’s  quanCtaCve  –  They  see  the  systems  behind  the  game  and  this  can  actually  “ruin”  the  game  a  bit  

6

Page 7: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Building  Balance  •  General  advice  

–  Build  a  game  for  creaCvity’s  sake  first  –  Build  a  game  for  parCcular  mechanics  –  Build  a  game  for  parCcular  aestheCcs  

•  Then,  a[er  all  that…  –  Then  balance  –  Complexity  can  be  added  and  removed  if  needed  – Other  levers  can  be  pulled  

•  Complexity  vs.  Depth  

7

Page 8: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

What  makes  a  game  balanced?  • When  evaluaCng  a  game  for  balance,  we  typically  look  at  three  aspects:  –  Fairness  –  Stability  –  Engagement  

8

Page 9: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Fairness  •  A  game  is  considered  fair  if  each  of  an  evenly  matched  group  of  players  has  an  a  priori  equal  chance  of  winning  for  any  given  starCng  posiCon  

•  In  a  normal  fair  game  for  two  players,  each  player  should  win  about  50%  of  the  Cme  with  both  players  playing  at  the  same  level    

9

Page 10: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Fairness  • What  does  it  mean  for  two  players  to  be  equally  matched?  

10

Page 11: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Fairness  • What  does  it  mean  for  two  players  to  be  equally  matched?  –  Similar  heurisCcs  – Ability  to  search  the  outcome  tree  the  same  distance  ahead  

–  Knowledge  of  probabiliCes  

11

Page 12: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

What’s  the  Probability  of  Winning?  •  Consider  older  games  •  Consider  modern  games  • What’s  the  probability  of  winning?  •  How  does  save  games  affect  this  probability?  

12

Page 13: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

The  “Going  First”  Problem  •  A  tradiConal  problem  in  fairness  is  the  “who  goes  first”  problem  

•  Assume  you  have  a  game  in  which  the  player  that  goes  first  wins  2/3  of  the  Cme  

•  How  would  you  fix  this  problem?    

13

Page 14: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

The  “Going  First”  Problem  •  Rotate  who  goes  first  

– Who  lost  last  game?  –  Randomize  – Age  /  skill  

•  Disadvantaged  player  gets  some  extra  resources  

•  Reduce  effecCveness  of  the  first  turn  –  Limited  move  set  

14

Page 15: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Reinforcing  Behaviors  •  As  players  do  things  in  games,  we  want  to  either  reinforce  or  punish  certain  behaviors  to  establish  appropriate  balance  and  pacing  

•  PosiCve  feedback  encourages  a  behavior  to  be  repeated  in  the  future  

•  NegaCve  feedback  discourages  a  behavior  to  be  repeated  in  the  future  

•  AdjusCng  feedback  adjusts  the  game  balance  

15

Page 16: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Reinforcing  Behaviors  •  Consider  basketball  

– When  you  scores,  the  other  team  gets  the  ball  –  This  is  negaCve  feedback  – We  don’t  want  a  team  to  be  able  to  get  ahead  too  quickly  

•  Consider  Mario  Kart  – When  you’re  in  the  lead,  you  get  crappy  items  –  This  is  negaCve  feedback  –  Rubber-­‐banding  is  also  negaCve  feedback  

16

Page 17: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Reinforcing  Behaviors  •  Consider  RPGs  

–  If  you  use  a  sword  a  lot,  it  might  level  up  –  Leveling  up  a  sword  makes  it  hit  harder  or  more  accurately  

–  If  the  sword  is  bejer  than  the  axe,  you’ll  use  it  more  

–  This  is  posiCve  feedback  

17

Page 18: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Reinforcing  Behaviors  •  Both  types  of  feedback  have  their  own  place  in  games  

• We  use  different  feedbacks  to  move  players  along  or  to  increase  challenge  

18

Page 19: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Stability  •  A  game  is  considered  stable  if:  

–  Feedback  is  negaCve  at  the  opening,  slightly  posiCve  at  midgame,  and  very  posiCve  at  endgame  

–  It  has  mulCple  viable  strategies  to  win  (called  stable  Nash  equilibria)  

19

Page 20: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Stability  Curve  

20

Page 21: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Curve  of  Progression  

21

Page 22: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

No  Feedback  Provided  

22

Page 23: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Curve  of  Progression  

23

Page 24: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Too  Lijle  PosiCve  Feedback  

24

Page 25: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Curve  of  Progression  

25

Page 26: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Too  Much  PosiCve  Feedback  

26

Page 27: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Curve  of  Progression  

27

Page 28: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Powerful  NegaCve  Feedback  

28

Page 29: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Curve  of  Progression  

29

Page 30: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Ideal  Game  Progression  

30

Page 31: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

MulCple  Strategies  •  For  good  balance  (and  for  engagement),  there  should  be  mulCple  ways  to  reach  the  win  condiCon  – Doesn’t  necessarily  mean  there  needs  to  be  mulCple  win  states,  but  that  can  be  done  as  well  

• We  can  mathemaCcally  reason  about  winning  outcomes  

•  Def  of  “uClity”  =  anything  used  to  measure  progress  toward  victory  

31

Page 32: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

MulCple  Strategies  •  Player  opCmal  outcome  -­‐  my  uClity  is  as  high  as  possible  

•  Pareto  opCmal  outcome  -­‐  my  uClity  cannot  increase  without  decreasing  another  player's  

•  Equitable  outcome  -­‐  everyone's  uClity  is  the  same  and  as  high  as  it  can  be  

•  Efficient  outcome  -­‐  the  sum  of  everyone's  uClity  is  as  high  as  it  can  be  

•  Nash  opCmal  outcome  -­‐  my  uClity  is  as  high  as  it  can  be,  given  other  players  played  to  their  own  interests  

32

Page 33: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

MulCple  Outcomes  •  Some  of  these  outcomes  are  not  necessarily  feasible  for  all  games  

•  Some  require  at  least  one  player  to  play  to  lose  •  Nash  opCmal  is  the  most  common  as  it  assumes  all  players  are  playing  to  win  and  your  ability  to  win  is  limited  by  how  well  others  play  (to  some  degree)  

33

Page 34: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Thinking  Down  the  Tree  

34

-1 2 -3 0 3 7 -4 -7 -2 4 0 1 -5 6 -6 5

Page 35: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Thinking  Down  the  Tree  • When  players  play  this  game,  they  use  a  minimax  algorithm  working  backward  from  the  opCmal  outcomes  for  their  goal  

•  This  is  a  zero-­‐sum  game,  where  one  score  affects  the  other  

•  The  only  real  outcome  here  is  a  Ce  if  players  are  playing  raConally  

•  Thus  the  game  is  not  balanced  because  there  is  only  one  possible  outcome  

35

Page 36: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Differences  in  Scale  vs.  Kind  • Why  does  this  majer?  •  How  does  this  play  into  engagement?  

36

Page 37: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Randomness  •  Computers  are  actually  horrible  at  being  truly  random  

• Which  is  both  good  and  bad  •  Bad  for  security  purposes  •  Good  for  networked  games  to  have  the  same  state  without  transmimng  that  state  

•  Probability  of  any  value  is  wrijen  as  P(x)  

37

Page 38: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

TradiConal  Randomness  •  Cards  and  dice  are  sCll  used  as  metaphors  for  randomness  

•  (and  non-­‐metaphors…)  • We  sCll  uses  these  items  in  both  board  and  video  games  for  randomness  

38

Page 39: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Cards  and  Dice  •  Card  notaCon:    

–  4S  –  4  of  Spades  – AH  –  Ace  of  Hearts  

•  Dice  notaCon  –  xdn,  where  x  is  the  number  of  dice  to  roll  and  n  is  the  number  of  sides  on  each  dice  

–  2d6  –  roll  2  6-­‐sided  dice,  values  will  be  2-­‐12  –  3d8+4  –  roll  3  8-­‐sided  dice  and  add  4  to  the  result,  values  will  be  7-­‐28  

39

Page 40: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

The  Outcome  Tree  • We  can  calculate  the  probability  of  any  random  event  by  working  out  the  outcome  tree  and  counCng  the  possibility  

• Monte  Carlo  simulaCons  run  the  funcCon  for  a  large  number  of  Cmes  and  using  that  to  determine  percentages  

40

Page 41: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Dice  Expected  Values  •  E(1d4)  =  2.5  •  E(1d6)  =  3.5  •  E(1d8)  =  4.5  •  E(1d10)  =  5.5  •  E(1d12)  =  6.5  •  E(1d20)  =  10.5  

41

Page 42: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

ProbabiliCes  of  Catan  •  Let’s  look  at  the  math  of  Catan  to  figure  out  how  probabiliCes  play  into  the  game  

•  Quick  overview  of  the  rules  of  Sejlers  of  Catan  –  hjp://www.catan.com/service/game-­‐rules  

42

Page 43: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Sejlers  of  Catan  

43

Page 44: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

ProbabiliCes  of  Catan  •  It’s  actually  prejy  easy  to  know  what’s  the  “best”  opCon  –  Just  add  up  the  dots!  

•  Probability  and  randomness  plays  a  HUGE  role  in  Catan  working  “correctly.”  

• What  about  games  in  which  probability  and  randomness  is  the  enCre  game?  

44

Page 45: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Chutes  and  Ladders  

45

Page 46: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Chutes  and  Ladders  •  The  game  is  ALL  RANDOM.  •  But  a  video  game  that  is  ALL  SKILL  can  eventually  get  boring!  –  You’ve  learned  every  pajern  –  You’ve  seen  every  level  and  enemy  – Nothing  varies!  

• We  need  to  consider  games  that  have  some  aspects  of  both!  

46

Page 47: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Why  do  people  gamble?  •  Let’s  face  it  –  gambling  in  Vegas  is  a  losing  proposiCon  

•  Over  Cme,  everyone  loses  money  •  But  in  the  (very)  short  term,  it’s  definitely  possible  to  win  

•  And  besides  –  risk  and  uncertainty  can  be  a  lot  of  fun!  

47

Page 48: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Psychology  of  Randomness  •  Player’s  like  longshots!  

– How  many  Cmes  have  you  gone  for  the  “super  move”  to  win  the  game?  

–  Even  if  it’s  a  low  probability,  players  will  opCmize  for  it!  

•  Player’s  suffer  from  too  much  Monte  Carlo  –  “Oh,  I’ve  gojen  bad  results  for  so  long…  a  good  card/good  roll  has  to  come  up  soon!”  

–  Probability  does  not  care  what  the  last  roll  was,  but  players  will  think  the  game  is  “unfair”  otherwise!  

48

Page 49: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Psychology  of  Randomness  •  But  think  about  it  another  way  –  I  bet  you  remember  those  big  payoff  moments  

•  And  THAT’S  what  gets  you  coming  back  to  a  game!  

49

Page 50: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Other  Forms  of  Risk  •  Imperfect  informaCon  can  add  to  the  challenge/risk  in  a  game  without  as  much  randomness  

•  Perhaps  you  don’t  know  everything  about  the  game  state  –  Either  AI  or  another  player  

•  Perhaps  don’t  know  about  the  game  might  change  

50

Page 51: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Fog  of  War  -­‐  ParCal  

51

Page 52: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Fog  of  War  -­‐  Total  

52

Page 53: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

InformaCon  Types  •  Info  known  to  all  players  •  Info  known  by  one  player  •  Info  known  by  the  game  only  

–  Some  game  state  informaCon,  like  next  card  in  the  deck  

•  Info  unknown  – Next  random  number  to  be  generated  

53

Page 54: GameBalancing - stardock.cs.virginia.edu · 2015-04-13 · Balancing# • Balancing#agame#is#can#be#quite#the#black#art • A#typical#player#playing#agame#involves# intuiCon,#fantasy,#and#luck#–it’s#qualitave#

CS  2501  

Difference  Between  Video  and  Board  •  Table  top  games  rely  on  randomness  to  work  

–  InformaCon  that  the  game  only  knows  can  be  hard  to  manage  

– D&D  does  this  through  a  Dungeon  Master  

•  However,  while  computers  aren’t  as  good  at  randomness,  they  are  fantasCc  at  managing  informaCon  

•  ImplementaCon  and  adherence  to  rules  also  varies  greatly  

54