So let's talk about accidental vs intentional design.

Most of Deepest Dark is firmly in the latter category. I go in with a specific goal, think about how I might accomplish that goal, and then execute the plan with surgical precision. Sometimes I'm dead on target, sometimes I'm a bit off and the idea needs refinement, sometimes I'm way off target the first time and that's what pre release testing is for. I generally don't believe in accidental design, because it involves flailing in the proverbial dark and hoping you accomplish something and hoping whatever it is you do makes the game better.

That said, a few brilliant things have come about by pure accident. The behavior of certain enemies is one example. By improving the AI of select enemies purely with the intention of making the game harder, I also added several new mechanics that didn't exist before but that let you deal with that difficulty while retaining enough dynamicness that they do not become rote and static.

A better example though, and the reason I made this thread was that there is a certain area not very far in the game. I believed that the encounters there were isolated and did not appear anywhere else, and also were fine as they were. I also noticed all the encounters in a much later region were much easier than they should have been and buffed them. Turns out they actually did share an encounter, and by pure accident I added an optional boss in the game. There's already dozens of those there intentionally, and I didn't believe that was really the enemy present until I saw video evidence. But that was the enemy present. And my first thought was that it shouldn't be there, but then I watched closely. Turns out it was actually a very difficult but winnable fight. You could even make it a much stronger win with better planning and strategy. This mistake, this oversight on my part actually improved the game. This thing that should not have even been there made the game better.

Now usually when something goes unexpectedly in game design it's a very bad thing. The vast majority of the time it is, and I certainly won't start the whole bang on keyboard and hope for the best approach. I did think it was a cool story, and also a funny one given I've seen a large number of deliberate and thought out design decisions elsewhere have worse results than me derping out.