The algorithm used to solve the compulsory problem suffers from two deficiencies: it makes no attempt to explore the map efficiently (the technique used to remember the last exit used from a tile and to cycle through them is guaranteed to explore the entire course eventually, but does not attempt to find an efficient strategy), and it fails to draw inferences regarding tiles based on nearby tiles: if a tile is explored that is connected to a previously-explored tile, it does not mark the connection as explored even though that would be possible. This latter change, in particular, would make a large difference in the time required to explore a course, at the expense of increasing the complexity of the ruleset. For purposes of illustration, it was felt that the simpler though less efficient ruleset would be more appropriate.
This ruleset includes the four reflexive rules previously mentioned, for elementary navigation. There are 27 tactical rules used in identifying tile types and 35 to decide strategy for tiles which are traversed more than once. 49 strategic rules handle intersections, including five which also have a role in tile type idenfitication. There are 95 tile states (of which 40 result from the tile rotation implementation, and appear in no rules) and 22 robot states.
Executing the compulsory problem on the course shown in Figure 3 gives the result shown in Figure 12.