Meta:
Choose domain X.
Human Y writes program for agent Agent(Y) in ArtificialMarket(X).
ArtificialMarket(X) is an algorithmic game about
offering and accepting challenges related to domain X.


analogies:
play chess through chess programs.
Instead of playing chess manually, write a program
(a chess playing agent) that plays chess.
Your rating will be based on the rating of
your agent.
challenges: 
I claim this property about chess positions of this form:
There exists f in F For all J in fs(f): prop(f,J).
If you think I am wrong accept the challenge.





SCG is a human meta-game that generates algorithmic games of agents
trading computational challenges in some domain. The goal of SCG is
to drive innovation through fine-grained, fair evaluation of knowledge
that the humans put into their agent.
The competitiveness of humans drives them to produce helpful agents
that constructively attack
the other agents by solving and posing hard computational problems
in the domain and its
niches (subsets of domain problems), effectively improving the agents
for the next contest. The humans are ranked based on the performance
of their agent and the agents are ranked based on how much they
know about the domain. Agents have limited resources: the agent evaluation
is based on how well they apply their resources and knowledge
to solving  the computational problems.