Solo or ‘GM-less’ play using random tables to generate the conflict is a very old idea, but not one I’ve seen applied to Mouse Guard, which is very much supposed to be (I think) a ‘story first’ game with little in the way of mechanical crunch or simulationist aspirations.
For example, there are a lot of things in the 1st edition Advanced Dungeons & Dragons game system that could conceivably support GM-less play. There is a random dungeon generation system, random treasure tables, random magic item tables, various sorts of random encounter tables, random NPC tables, and the 1e Monster Manual with its ideas of number appearing, percent in lair, and demographics information could reasonably be used as a hex crawl generation system.
But what would be the attractions of playing in such a system? Well, for one thing it would randomly generate tactical skirmish level combat, so essentially the group could be playing a wargame with a certain sort of continuity. But, by comparison Mouse Guard has very little interest in meaningful decision making during combat. Mouse Guard is really only trying to create a transcript of the scene and the only details of combat of combat decision making really have to do with narrative currency like ‘fate’ and ‘persona’ points. Mouse Guards roshambo mini-game is like roshambo itself designed mostly to be a random generator itself, meaning that there is a sharp distinction between the aesthetics of play in something like D&D and Mouse Guard. You don’t play Mouse Guard because you care much about tactical positioning or resource allocation.
One other thing that makes Hex Crawl play interesting in a game like AD&D is character advancement. Mouse Gaurd does have traditional character advancement, but nothing like AD&D’s ‘zero to hero’ bildungsroman or its ‘character advancement as minigame’ that has been adopted by its crunchier later editions. There aren’t a lot of CharOp oppurtunities in Mouse Gaurd, nor is the act of acquisition of treasure a part of CharOp in Mouse Gaurd the way it is in D&D. You don’t have 5 or 10 tiers of magic swords in Mouse Gaurd, so there just isn’t the same ‘loot box’ aspirations that you have from D&D and its trad RPG cousins (whether PnP or computer games).
In short, I think you have a beautiful start on a PDF there but what you haven’t sold me on is what the aesthetic of play in a GMless Mouse Gaurd will be? Nor have you suggested how you are going to resolve the principle problem I see with GMless Mouse Gaurd, which is the fact that conflicts in Mouse Gaurd have explicit stakes and I’m not sure how you go about creating a random stake generator. If you don’t have a random stake generator for the different conflicts you have that makes some sort of sense,then you run into the problem that a game isn’t compelling if the same person responsible for setting the stakes is also responsible for resolving the conflict.
So at the least, you are going to need to show me that you can better generate random stakes/motives for the NPCs in the game.
I also don’t think you can do story tracks without forks in the story. And here I think is another big problem. Story based game depends heavily on the GM wearing the hat of ‘secret keeper’. To do a story game you vastly more need the GM as ‘secret keeper’ than you need one as ‘referee’ or ‘adversary’. I think you could get by in Mouse Guard quite well without a referee or an adversary, but I just don’t see how you could get by without a secret keeper.