Playful game comparison and Absolute CGT
Urban Larsson, Technion - Israel Institute of Technology, coauthors Richard N. Nowakowski and Carlos P . dos Santos
GAG-2017, Lyon 1
Playful game comparison and Absolute CGT Urban Larsson, Technion - - - PowerPoint PPT Presentation
Playful game comparison and Absolute CGT Urban Larsson, Technion - Israel Institute of Technology, coauthors Richard N. Nowakowski and Carlos P . dos Santos GAG-2017, Lyon 1 Thanks to organizers We develop a framework for many classes
Urban Larsson, Technion - Israel Institute of Technology, coauthors Richard N. Nowakowski and Carlos P . dos Santos
GAG-2017, Lyon 1
combinatorial games:
restrictions on the games: dicot, dead ending, guaranteed scores, etc
Milley, Ettinger, Stewart, Santos, Nowakowski, Larsson, Dorbec, Sopena et al.
wish to unify theory
alternating perfect play, a given winning condition, disjunctive sum, etc
prefer G before H?
any game in the same universe
structure and game comparison simplifies to play G-H
computation
winning conventions
recursively, starting with each adorned empty set of
with a neutral element, ‘0’
no options
the outcome depends on who moves last
games?
assign a binary result to G
terminal situation
postponed until G+H ends
to -1
to +1
evaluates to a
= (oL(G), oR(G)), where
left options of G
parental and dense
the game {G|H} is also in the universe
then there is a game H such that the o(G+H) = x
comparison is ‘constructive’; we use a normal play analogy:
provisonal game (LPG), is played as follows
absolute universe, G ≥ H if and only if Left wins [G,H] in normal play (!) playing second
GRL such that GRL ≥ H, or there is HR such that GR ≥ HR
HLR such that G ≥ HLR
uses the downlinked idea developed by Ettinger and Siegel
such that oL(G+T) < oR(H+T)
downlinks H (easy)
all HR, G not ≥ HR (hard, uses dense and parental)
players have an option
ending misere, etc) game comparison is also constructive: see Richard’s and Rebecca’s talks
contains all the good ideas got rejected twice. It is probably the strongest paper I wrote.)
any absolute universe. It seems that guaranteed scoring play could have interesting categorical structures. Similar to normal play it satisfies a certain closure property. (Dicot absolute universes do not satisfy closure properties.)
extensions (they are between dicot and dead ending).