v2.6.0

v2.6.0, 2017-02-26

New strategy, state to action analysis and internal improvements

• A number of type hints added to the library
  #860
  #863
  #864
• New strategy: SelfSteem
  #866
• Add state to action analysis
  #870

Here are all the commits for this PR:
v2.5.0...v2.6.0

v2.5.0

v2.5.0, 2017-02-11

Internal improvements, type hints, documentation and a new strategy

• New strategy: ShortMem
  #857
• A number of type hints added to the library
  #828
  #831
  #832
  #833
  #834
  #835
  #836
  #840
  #846
  #847
  #849
  #850
  #851
  #853
  #854
  #856
  #858
  #824
  #821
  #815
  #814
• internal improvement to how players are cloned
  #817
• Refactor/removal of dynamic classes
  #852
• Run windows CI for py3.6
  #844
• Run mypi on travis
  #843
  #837
• Small update to the readme
  #829
• Docstring fix for random
  #826
• Improve efficiency of neural network strategy
  #819
• Improve efficiency of cycle detection
  #809
• Refactor of a number of tests and test documentation
  #820
• Large refactor thanks to dropping of python 2
  #818

Here are all the commits for this PR:
v2.4.0...v2.5.0

v2.4.0

v2.4.0, 2017-01-05

New machine learning strategies and moran processes on graphs.

• Moran processes on graphs
  #799
• Machine learning strategies
  #803

Here are all the commits for this PR:
v2.3.0...v2.4.0

v2.3.0

v2.3.0, 2017-01-04

Support for py3.6, new strategies, more tournament result information, and
internal improvements.

• Helpful list of short run time strategies
  #792
• Nice Meta strategy
  #794
• New strategies: Mem2, Pun1, Collective Strategy
  #795
• New strategies: Mem2, Pun1, Collective Strategy
  #795
• Python 3.6 supported
  #800
• Keep track of initial play rate in results
  #797
• Fix depreciation warning
  #793
• Moran processes are always stochastic
  #796

Here are all the commits for this PR:
v2.2.0...v2.3.0

v2.2.0

v2.2.0, 2016-12-20

Minor update: ability to pass axes object to plots and internal documentation
build fix.

• Pass axis object to plots
  #791
• Build docs with py3
  #788

Here are all the commits for this PR:
v2.1.0...v2.2.0

v2.1.0

v2.1.0, 2016-12-8

New strategies, ability to fingerprint instances and appearance improvements to
fingerprint plots.

• New Meta strategies
  #777
• Better and better, Worse and Worse 2, Worse and Worse 3
  #786
• Add ability to Fingerprint instances
  #785
• Add plotting options to fingerprint
  #787

Here are all the commits for this PR:
v2.0.0...v2.1.0

v2.0.0

v2.0.0, 2016-12-05

Dropping support for python 2, bug fixes, minor tidy of code (thanks to
dropping python 2 support!) and progress bars for fingerprinting.

• Dropping support for python 2
  #774
• Fix bug in cache
  #782
• Fix bug in stochastic classification of Random player
  #783
• Fix docstrings in fingerprint
  #784
• Use python 3 function caching
  #775
• More progress bars for fingerprinting
  #778

Here are all the commits for this PR:
v1.19.0...v2.0.0

v1.20.0

v1.20.0, 2016-12-05

Dropping support for python 2, bug fixes and minor tidy of code (thanks to
dropping python 2 support!).

• Dropping support for python 2
  #774
• Fix bug in cache
  #782
• Fix bug in stochastic classification of Random player
  #783
• Fix docstrings in fingerprint
  #784
• Use python 3 function caching
  #775

Here are all the commits for this PR:
v1.19.0...v1.20.0

1.19.0

v1.19.0, 2016-11-30

New strategy using a trained neural network.

• Implement the EvolvedANN
  #773

Here are all the commits for this PR:
v1.18.1...v1.19.0

v1.18.1

v1.18.0, 2016-11-24

There are no changes between 1.18.1 and 1.17.1. This release is due to an error
during the uploading to pypi.