I put serious thought into selecting the right game for the long Textmind demo, but still ended up getting it wrong. Too much emphasis on finding a low-competition deep game to transform. That turned out to not be useful unless I wanted to actually fork DCSS, because I couldn't prove my improvements correct. Time to justification was too long without pvp. So I defer further DCSS work.
This time I selected for a game with sufficient depth, but also modularity and pvp-driven objectivity. As a result, I'm not delivering sublime genius, just straightforward strategic improvements. It doesn't show how Textmind can deeply augment to achieve lifetime genius. But that doesn't matter. The goal here is just to demo the method so that people can understand and try it.
Modularity is such a relief. It is vastly easier to write a guide to one scenario of a campaign, than to write about a procedurally-generated roguelike.
I still have yet to beat After the Amulet on Impossible in the fully vindicating manner I seek. I am quite close, however. I just reviewed two other elite playthroughs, critiquing one and learning from the other. Both of them feared to hold the Rampart turtle against the Necro hordes. I'll be rectifying that, doing the impossible, demonstrating that Rampart is superior when played according to designer intent, and that fielding complementary Might and Magic primary heroes is optimal.
This transformation is plenty profound enough for demo hype, even if it isn't awesomely deep like my DCSS work, which probably no one can appreciate until it's made into a fork.
[2021-08-08 Sun 12:50]