![]() |
| The Assembled uComputox Kit |
The kit uses an ATTINY84 to simulate the behaviour of the Computox. The state of the board is shown on nine, bi-colour LEDs, with player input captured using 12mm contact switches. The uComputox uses the same algorithm to play single-player noughts-and-crosses and incorporates both the secret/beatable and random modes of play found in the original. This makes it ideal for children to tease siblings or friends with their inside knowledge.
In addition to tic-tac-toe, the uComputox also plays the classic machine learning game Hexapawn. Created by Martin Gardner in 1962, Hexapawn demonstrates reinforcement learning through a cut down version of chess. When switched on, the uComputox is a noob, making moves at random. However, the uComputox does not repeat its errors and as you beat it over 20 or so games it becomes unbeatable.
![]() |
| The Drof Computox Exhibit at Retrofest 2026 |


Comments
Post a Comment