Type in the content of your page here.

Class 11 Game Dynamics

Speaker: Andrew Chen, Entrpeneur-in-Residence at Mohr Davidov Ventures

Prior Readings:

How to understand how what makes games addictive


Games get more money than Facebook. (games) web world and game industry are colliding, over 15 billion dollars, bigger than movies
Most have been spent in the console world, Zynga, rockyou, fish, they have a background in web not in games, they are moving away from ea and other huge game companies
Why do people play games (compare to literature)? Refer to some classic reads:
  • Ronny Kohavi (extra readings)
  • Cialdini's persusaion
  • Bartle's "Design of Virtual Worlds"
  • "A Theory of Fun"

Why Games?

Game dynamics are a crucial part of any new consumer web product. This will be very important for new entrepreneurs . Match artists with statistics and math gurus (to provide interest, and then optimize the game virality)
When prompted byAndrew, the class came up with a list of the following reasons that they would plays games:
  • credit/fun
  • challenge
  • killing time/avoiding work
  • compete with others
  • status
  • live a fantasty
  • try new things - cutlurally relevent, fashion, hi
  • socializing
  • low barrier/ easy
  • always more to go for
  • instant gratification/instant feedback
  • self metrics
  • learning
  • filitering
  • destress
  • peer pressure

We then moved to the analysis and literature behind the above reasons.

Quantitative Aspects

measure and optimize that smack
funnel analysis
cohort analysis
ab testing
these two worlds are in battle to figure out hihg preformance producs
bartle and multiuser dungeons (MUD) grad students

Qualitative Aspects

How does the growth happen? Generally, it is through very simple viral prompts (optimized through the psychology and rigours testing AB tests)
Deadly sins are a driving force.
Consider the framework for understanding users as pioneered by the inventor of multiplayer games, Richard Bartle.

external image 2275638906_f7cb9b447e.jpg
(4 buckets) socializer explorer(master the world that live in) achievers(fundamentally motivated by points and levels) and killers(pwns nwbs like it's nobody’s business) (evil but in a good way kill defenseless people) this applies to mass multiplayer online games this is one framework
achiever - i jsut got a high schore
exlpoer - look what i found
socializers - try this with me
killers - i pwned this town

How do you optimize virality?

Viral marketing is when people tell their friends about your product.
Look at the percentage of people that tell their friends, and quantify the likelihood that the friends will click through. Look at it like a times series (it grows exponentially) the idea is to get 100 costumers to sign up more than 100 friends
Conceptualize it as a viral loop (need to make power in exponential greater than 1 to get growth otherwise it will decay to nothing!) Also, you can unroll the loop into a funnel.
The message of how to ask people to add friends is very important. Consider the facebook example.
AB testing is used for optimization of the virality element
go on a step by step basis (so it's a lot of tuning) build the game mechanics into the virality so it occurs at the same time so they don't buld the game and add the virality in it's a hand and hand process
get critical mass (things are now statisically significant) ie get enough users and data
new world hypthesis
old world statistics

May 4, 2010

4:15 Purpose: Understand game dynamics (addiction), and how to make for you (incentives etc)

4:20 AC Landscape / Why games matter ($$)

100M unique users on Zynga / day

New way of product, new way of product design

4:30 ASW Primary goals of gamers (exercise)

4:40 AC PHAME and GAME

P Exponential Growth


H/A Qualitative / Creativity

Readings Andrew Chen


Readings Ronny Kohavi

Quantitative (metrics and experiments)
KDD 2007 http://exp-platform.com/hippo.aspx

Practical Guide to Controlled Experiments on the Web: Listen to Your Customers not to the HiPPO by Ronny Kohavi, Randy Henne, and Dan Sommerfield.

5:25 Prep for Blippy

5:30 End

Dan Goodwin dgoodwin208@gmail.com
Rob Cosgriff cosgriff@stanford.edu
Wrap up points:
Data > Algorithms
Change behavior > technology
Something new that's neither public nor private
Atomization > big review
Openness- companies publish data they have
Companies can't suppress phenomena
Consumption is bottleneck, not production or distribution
Lists -> Flows