[Kde-games-devel] Review Request 111705: Testcode for hard AI for KSquares

Albert Astals Cid aacid at kde.org
Wed Aug 14 21:39:15 UTC 2013


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/111705/#review37799
-----------------------------------------------------------


Hi Tom, do you think you could turn this into some kind of auto test? So that we run 100 games and check that medium in general wins to easy, and that hard wins to medium?

- Albert Astals Cid


On July 25, 2013, 11:22 p.m., Tom Vincent Peters wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/111705/
> -----------------------------------------------------------
> 
> (Updated July 25, 2013, 11:22 p.m.)
> 
> 
> Review request for KDE Games and Matt Williams.
> 
> 
> Description
> -------
> 
> This is meant to test whether the new (hard) AI code has changed the medium AI strength in KSquares as a side effect. 
> Hard AI code: https://git.reviewboard.kde.org/r/111662/
> 
> This is not supposed to be comitted, it's just for testing purposes.
> 
> Use by calling KSquares from command line in demo mode 'ksquares --demo'.
> 
> Changes:
> - old and new ai code side by side
> - changed demo mode to two player mode
> - speed up of demo mode
> - most debug messages commented out
> - overall game score for statistics
> 
> Changes not included:
> - changed playing position of old and new AI
> 
> Result:
> - the new code for the hard AI doesn't seem to change the strength of the medium AI.
> - AI playing in second position has a strong advantage
> 
> 
> Diffs
> -----
> 
>   src/aicontroller.h beaafa8 
>   src/aicontroller.cpp 0bff5e6 
>   src/gameboardscene.cpp 0fa669a 
>   src/ksquaresdemowindow.cpp eea7e5a 
>   src/ksquaresgame.h ca4fe8b 
>   src/ksquaresgame.cpp 74a06d3 
>   src/ksquareswindow.cpp 092c795 
> 
> Diff: http://git.reviewboard.kde.org/r/111705/diff/
> 
> 
> Testing
> -------
> 
> Seems like the second player has a serious advantage:
> 
> ksquares(18337) KSquaresDemoWindow::gameOver: Game Over
> ksquares(18337) KSquaresGame::playerSquareComplete: Player 1 won
> ksquares(18337) KSquaresGame::playerSquareComplete: Player 0 ( oldAI ) won 391 times
> ksquares(18337) KSquaresGame::playerSquareComplete: Player 1 ( newAI ) won 1314 times
> 
> ksquares(17513) KSquaresDemoWindow::gameOver: Game Over
> ksquares(17513) KSquaresGame::playerSquareComplete: Player 1 won
> ksquares(17513) KSquaresGame::playerSquareComplete: Player 0 ( newAI ) won 288 times
> ksquares(17513) KSquaresGame::playerSquareComplete: Player 1 ( oldAI ) won 1008 times
> 
> 
> Thanks,
> 
> Tom Vincent Peters
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-games-devel/attachments/20130814/58906558/attachment.html>


More information about the kde-games-devel mailing list