I'm not really talking about commercial failures, or something that one might deem an aggregate failure via metacritic. You don't have to ship failures.
The best kind of failure is just a prototype, a thing where minimal art assets were committed, minimum peripheral effort, before identifying a bad idea. "Bad" is obviously subjective but for me it's anything in the category of "this isn't that fun after all...".
On the flip side the best moments in game design for me are the "holy shit" moments of being suddenly invigorated and excited by some combination of elements that are just super fun to play with.
Even then, depending on how you define failure, there's no way to create guarantees of "success" (also depending how you define that). This is especially true in the realm of commercial-failure, where you can create a highly polished, excellent game with an excellent critical reception, and sales could still be abysmal.
Even cloning a successful Final Fantasy game, if your art doesn't match up to what the original game studio produced using thousands of hours of top-sourced professional industry talent, you're at a disadvantage in the inevitable comparison.
Like probably most typical game designers I'm afraid of failure too, but I create for the fun/love of it. I think personally my obsession with tinkering and building things would always outweigh those negatives.
I think you could probably modify your thread title (if that can be done?) or at least your OP to be more like "ways to avoid failure during game design/development" as a more concise way to get that advice.
It's a huge subject. A much bigger subject than whether to be original or to heavily appropriate.