With this one, I had to get it up and working and actually play it for a while to determine how I wanted the AI to work for the one-player option. Once I could actually get a game going, I realized it needed tweaking or the game would be essentially one long stalemate, but as I've tried to add these small changes I've kept bumping into the ceiling of my available memory. Also, I'm unearthing bugs in code written early in the process that hasn't been fully put through its paces until now, when I can play multiple turns of the game.
I've realized that one element of the game - the R)eclaim option on the menu - might be superfluous enough to get deleted and save me a chunk of space. The idea is, if two of the roaming caravans collide when they move, the tile on which it happens turns black and becomes 'barren,' with a value of zero. Any future caravans that reach it will also be destroyed, and the tile can't be bought or attacked by either player. The Reclaim option, only available if a player rolls a 6 and has an acre adjacent to the barren tile, allows the space to be reclaimed and put back into play.
I still have yet to play a long enough game where multiple caravans have crashed into each other, or where either player's empire has sprawled to the point where swaths of barren land are a problem. It's not a huge piece of the code, and it may come up infrequently enough that it's better to just make barren acres a permanent thing, and force players to work around them.
I aesthetically like the idea of the reclaim (it's also a part of the Enemy Lines board game, although there your pieces are in more of the caravan role and you're trying to navigate across the playfield). But this is the "murder your darlings" portion of editing that I've never been great at, in writing or in code, and BASIC is forcing me to confront it. I've at least wised up enough to do my code editing in Notepad and save annotated revisions, so I can immortalized that chunk of code in my notes if it ever needs revisited. If cutting this out gives me the space to make the last edits I need and make the game playable, it's a worthy sacrifice, and it probably makes the game make more sense.
All this, and I still don't have the AI added! I've already realized that my intended approach -- having the game make decisions and then reusing the code from the players' turn to make it look like someone's sitting there typing with you -- will likely not fit. There will likely be a few broad strokes in the AI's process and then a quick update to the playfield, more similar in execution to the portion of the "day" when the NPC caravans move. I'll probably release two standalone versions of the game, similar to the old-school releases with a keyboard-input version on one side of the tape and the joystick version on the other. It seems weird to think of how few people will end up playing the two-player version I'm currently sweating over, but without it, a 1-player version would likely never exist.
Hopefully I can wrap this up quickly! I've thought I was a day away from finishing it half a dozen times so far, though, so I'm not gonna hold my breath. But I am gonna keep going, because this backlog of games isn't gonna finish itself.