I feel as though the people who say Dishonored is a spiritual successor to Theif either never played or don't understand what made the game what it was. Every game that isn't pure twitch reflexes and/or memorization is, all things considered, a puzzle game. Theif's puzzle is about navigating a space while managing visibility and sound. In Dishonored you can teleport so the puzzle of navigating space is very different. You don't have to earn every inch. Two spaces of shadow broken up by three meters of lit area isn't a problem. You don't have to ask the question "If I go down there, how am I going to get out?".
Then the way you look at guards is different. In Dishonored you can easily take out any guard and their friends in one on one combat. You see a guard and you lament "I could kill them but I'm going for the good ending". In thief you see a guard and think "I don't want to expend resources taking this guy out, and I'll never win in one on one combat". Which is better at getting the player to think like the character? In thief combat is dangerous and killing begets more problems. You shoot someone with a broad-head arrow they're no longer a problem, but there's blood on the ground that you have to expend resources, namely your water arrows, to clean up. You buy your gear so every arrow sent from your bow is one you have to replace with the money you steal or get from your fence. You're a better thief if you don't fire any arrows and can save the cash.
Thief's stealth likely never will be surpassed, at least by a Triple A studio. I really like Thief and I wanted to make a spiritual successor so I did as much research as I could into how it ticks. Thief is a software rendered game, meaning lighting information is accessible in other areas of software. The light gem is updated based on how lit the area actually is, not by a guess from a raycast measuring distance to nearby lights. This isn't the case with GPU rendered games (nearly all modern games) where there's a bottleneck between the CPU and GPU. People simply aren't thinking in terms of using the actual lighting information as something the AI could use. Even if they were, the project manager would go "Are you crazy? We have the player rendering these PBR materials in 4k at 60 FPS with more particles in a frame than there are people in this city, and you want to render the scene again from the perspective of the guards? I don't care if it's lower res, 3 FPS, and only uses the depth map, a stencil map, and the lighting information, we don't have the resource budget for it."
ual Strike where you just build the big unit and win.
