Chess Maps

From Wiki.q.c

Revision as of 07:25, 23 February 2010 by Quodlibetor (Talk | contribs)
(diff) ←Older revision | Current revision (diff) | Newer revision→ (diff)
Jump to: navigation, search

Chess Maps can be thought of as either topological or topographical maps created by a game of chess, as generative art limited by the rules of chess, as an algorithm for visualizing a game of chess, and, of course, as a technique for generating not-unpleasant images.

Maps are generated by the movement of pieces across the board. As a piece moves it leaves behind a `trail' of color, with white and black having different colors. As pieces overlap their colors support their own side's coloration, but they change the trails left by the other side.

This is a project created by Brandon Maister, Artem Chernenko, Pedro Espinoza, and Kenneth Cordero done for our expressive computing class at hunter college.

Contents

Implementations of the Trail

In our processing implementation the map starts off as an 8x8 grid with background set to (0,0,0). As the pieces move they add some color to the squares that they move over. Which means that as you click on the board, some color is added to that square. So in for example our kasparov vs. deep blue game images (tarball) every time black moves, 60 is added to the red of the squares it passes over and lands on. White gets +30 blue and +25 green, just because that color looks nicer to me than just +60 blue.

We have a variety of other color schemes and trail techniques. For example, in a couple of our games each piece is assigned a unique color and it completely overwrites any previous color in that square when it goes over it, as in the topmost shot on this page. We also experimented with having only the squares that were landed on colored.

A completely Random Board

Observations

Aside from the aesthetic value of the finished and in-progress boards, they provide an interesting view of games that have been played. A surprisingly good feel for the game that has been played can be had from a completed chess map, assuming you know what you're looking at, otherwise there are just a bunch of colors on a grid.

An interesting question is whether just completely randomly coloring the board (as at left) leads to a noticeably different board than a board that has been colored by random moves or an actual chess game. (Below right and left, respectively)

Unfortunately, the random move generator was extremely difficult and unpleasant to use and basically begged for human error, which made it impractical to perform a large number of random games. The chess map visualizer was not so unpleasant to use, but there was no way to recover from accidental or incorrect clicks, and it is too easy to make mistakes or get lost in a game.

Basically, this would be far more effective with a parser.

However, there is an obvious difference between games colored by the rules of chess and completely random colorations, and the image below suggests that there might be a noticeable difference between random-move generated games and real games by the very experienced. As with the propensity for error though, it would take a very significant investment of time to see if these effects are more than just illusory, although this would be mitigated, if not eliminated, if we had a parser.

So, in the long run, it is pretty apparent that this project could have used lots more computer love.

Two maps side-by-side

A whole new way to play

The visualization of chess in this way opens up the idea of games of chess being played not primarily to defeat the other player, but to create new or novel images. Of course, it would be nice if there was an implementation that didn't require clicking on a square every time someone moved.

a game of chess mapped in oil

Ideas for further exploration

Visualization

  • Both sides set to the same color instead of warm / cold colors
  • Different colors for each piece
  • Play with different games or rulesets

Implementation

  • A chess program implementing these ideas might be cool (at parties)
  • a parser would really open up the possibilities
    • real generative art
    • collaborative games

Exploration of the effect of rules

  • Do the rules of chess affect the eventual appearance of a completely random-generated art?
    • Would random games look significantly different to randomly colored squares?

links

images of games

implementations and details