Currently existing ai-tools are not AI.
Just complex algorithms.
I not understand difference between photoshop tools / midjourney / Grammarly / word grammar checker / chatgpt - it is all tools.
I prefer spell-less ranger as base variant because:
i can have 1/3 spellcaster ranger
i can have 1/2 spellcaster ranger
i can have full martial ranger
And i think that having different options will be good for ranger
Also - I know that WoTC creates average stats hoping that pcs will hit in 65%, but i not understand what hit probability for monsters was in their design documents. i think - ~40% (I hope to rebalance system to make monsters more similar to pc in sense of number of hit-points and average dmg/to...
for similar reasons i began with most simple decision - deadly encounter for me is probability of losing equal 50%
But it is not very helpful in describing other levels of challenges ....
Also, i think about using encounter building system. I create encounters usually by world requirements. If...
I make mant npc as pc and then convert them to creature statblock, so i calculate cr anyway.
i get fighter battlemaster for example at level 20 equal cr13
so if began to finde cr equal to level i would began with simple warrior - fighter champion and use him as baseline
I still trying to made something similar (and delete dependancy of challenge rating and proficiency bonus).
My starting point was to calculate CR of standard human fighter champion from 1st to 20 lvl.
Second, I set different challenge difficulty framework - hard is fight vs equal enemies (when...