Recently I found out that Limbo was available for free this month for those who had a PlayStation Plus subscription so I took advantage of it and played through it. I was not impressed. It's not that the puzzles were badly designed. While rather basic, with the majority revolving around switches and boxes, they often made you consider all of the resources that were available to you. If a puzzle game had good puzzles, logically it should be a good game, but I didn't enjoy it. Why?
My first thought was that it was due to the bad story. I'm okay with ambiguity, but Limbo doesn't give enough to pique interest. In the early portions of the game when they introduce that mind control slug, I thought perhaps you were some kind of high functioning zombie but instead we get something...that might be a personal hell? A bad story could kill a game, but I'd prefer to play those escape the room games over replaying Limbo so that probably isn't it.
If the problem wasn't with the story, then the only element remaining for this game is the puzzles, which I've already said were good, but there was a great element of frustration with solving them. I died a lot in this game. Deaths are always a source of annoyance but they only become truly frustrating from an unavoidable circumstance or from making the same mistake twice.
The art style resulted in the first. Deaths occurred regularly from things that you didn't realize were dangerous or from not realizing certain actions were dangerous until after you triggered them. However, you can't blame the art style for the one that sticks in my mind; a pair of crushers with buttons under them, the first triggered by not stepping on the button and the second by doing the opposite. This kind of trial and error gameplay conflicts with the expectations of a puzzle game and therefore is likely a reason why I considered it a bad game.
Making the same mistake over and over again is a frustration many experience, but more with platformers than puzzle games. There are lots of puzzle-platformers out there (eg. Portal), but the first half of the game doesn't include the same physics based puzzles as the latter half and as such when they turn up suddenly you're in the wrong mindset and therefore are more susceptible to annoyances.
In summary, it seems that my main problem with Limbo came from it not living up to the expectations I had developed for it. Trial and error gameplay conflicts with the assumption that a puzzle game will be set on logic which was only compounded as logic gave way to precision platforming as the game went on. The take away from this experience for me was that when designing a game it's important to know your audience and their expectations.