MarkovTextAdventure

MarkovTextAdventure copied to clipboard

Markov Text Adventure Analysis

This is a Markov analysis of the The Teeny Tiny Mansion. Please see Evan Miller's blog for the basic information on what this is supposed to do.

Usage:

julia MarkovTextAdventure.jl

or

python MarkovTextAdventure.py

Typical output:

Scenario 1: Green-key bug and help-me bug
 => There is a dead end :-(
   => Calculating how bad it is...
   => Still calculating...
 => This bug affects 78.46% of button-mashing players
 => Valid states: 1680
 => MAX_FINISH_DEPTH: 14
Scenario 2: Green-key bug only
 => There is a dead end :-(
   => Calculating how bad it is...
   => Still calculating...
 => This bug affects 25.00% of button-mashing players
 => Valid states: 1664
 => MAX_FINISH_DEPTH: 14
Scenario 3: Help-me bug only
 => There is a dead end :-(
   => Calculating how bad it is...
   => Still calculating...
 => This bug affects 71.28% of button-mashing players
 => Valid states: 1680
 => MAX_FINISH_DEPTH: 14
Scenario 4: No bugs (hopefully)
 => 59 recurrent classes and no dead ends! Hooray!
   => Verifying the result...
   => Still calculating...
 => This bug affects -0.00% of button-mashing players
 => Valid states: 1664
 => MAX_FINISH_DEPTH: 14

Python differences

The Python port (contributed by hackerb9 has some differences from the Julia version.

Changes:

• MarkovTextAdventure.py runs on Python (2 and 3).
• Uses Arnoldi Iteration and sparse arrays.
• Runs faster, requires less memory.
• Now works on CPUs from the 1990's which lack SSE2 (woohoo!).
• Does not yet implement % button-masher code.

To do:

• Calculate % of button-mashers affected.
• Make matrix more sparse by culling actions that go "backward" .
• Allow Alice and Bob to be eaten by a GRUE.

NB: This is a work of ergodic fiction, suitable for appropriate age groups only.