Perfect-Play Databases for Dodgem on 3×3, 4×4, and 5×5 Boards (MongoDB Dump)

This repository contains a complete MongoDB dump of perfect-play evaluation databases for the combinatorial board game Dodgem on 3×3, 4×4, and 5×5 boards. The databases were generated by exhaustive search using the open-source Python package dodgem: GitHub: https://github.com/sekika/dodgem-py Python Package: https://pypi.org/project/dodgem-game/ Online Dodgem page: https://sekika.github.io/dodgem/ Content The archive dodgem-mongodump.tar.gz (≈2.8 GB) is a mongodump of the MongoDB database dodgem_db, which stores exact game-theoretic evaluations for all reachable positions for board sizes n = 3, 4, 5. For each board size n, the database contains: eval_n collection Documents keyed by position key string: _id = "[[a...],[b...],turn]" Fields: value: evaluation score ≥ +100: the side to move has a forced win ≤ -100: the side to move has a forced loss 0: forced draw depth: depth bucket (decreases with deeper search) remain: heuristic “distance to exit” (pieces) depth_n collection Index/bucket documents grouping candidate positions by depth and remain. Keys: _id = "d{depth}r{remain}" with field key = [list of position keys] for small buckets For large buckets (> 300,000 positions): parent document: _id = "d{depth}r{remain}", { "large": 1 } sharded sub-documents: _id = "d{depth}r{remain}i{index}" fields: { "dr": "d{depth}r{remain}", "index": index, "key": [...] } The database sizes are approximately: n = 3: 1,963 positions n = 4: 393,900 positions n = 5: 164,308,067 positions Total disk usage (uncompressed MongoDB) is about 14 GB. Usage To restore the database into a local MongoDB instance: tar xfvz dodgem-mongodump.tar.gzmongorestore --drop mongodump After restoration, the dodgem Python package can connect to the MongoDB server (database name dodgem_db) and: play Dodgem against a perfect opponent on 3×3, 4×4, and 5×5 boards; traverse the game tree from any position; perform further analysis using the Python API. References E. R. Berlekamp, J. H. Conway, R. K. Guy (2003), Winning Ways for your Mathematical Plays, Vol. 3 (2nd ed.), A.K. Peters, pp. 749–750. Historical discussion and early win–loss tables (4×4 and 5×5): https://ics.uci.edu/~eppstein/cgt/dodgem.html Wikipedia article: https://en.wikipedia.org/wiki/Dodgem