Take a game type you like. Then give it some odd twist that makes it distinct from others in its category.
Twists can be game mechanics like - adding some aspect of another genre into a genre that normally wouldn't do that. Sometimes that means digging into areas other devs won't dare tread. When the Myst devs at Cyan made a Myst title (Uru) that embraced physics, realtime 3d, multiplayer social interaction, etc... basically challenging their own franchise and its norms in nearly every way, and they failed painfully to find a large enough number of players who embraced that, it killed the idea of multiplayer puzzle games [of any kind] for nearly a full decade afterward. Until, circa 2012, there was Journey. Which made a short but meaningful and emotive puzzle / platformer and didn't actually tell anyone it was multiplayer. They just sort of let players unexpectedly run into each other in the course of playing. This, quite frankly, was rather delightful. The art style was beautiful, the game was short but refreshing, a journey about the course of life and eventual death and trying to make sense of it, and kind of concluding that even if our lives end it's what we leave for the people along with us and who follow after us that matters... condensed into three hours that made 1/3 of players cry and which some described as the closest any game had ever come to being 'profound' or a spiritual experience.
A good example, one of my fave games in recent memory, is Outer Wilds [NOT 'The Outer Worlds', an RPG which came out with a very similar name independently by sheer accident at almost the same time} it is a game that absolutely is hard to categorize. Everyone tries not to spoil it and you really shouldn't go in knowing much of anything, but basically the reason for that is the bit by bit understanding of its world and the internal logic of it is pretty much the entire core of it. You're putting together clues and making sense of a weird mystery in a universe that isn't ours and has some odd physics logic in specific ways that aren't true in our universe. But as you start getting the various secrets and truths of it, that understanding allows more and more progress. Such that the game takes most players almost 20 hrs. to solve without help but if they knew everything they'll realize by the end and chain that understanding together then it'd be possible to hit the ending sequence within just 20 mins. with that endgame understanding. So in theory, it's a puzzle game, even a really clever real time 3d Myst like in some sense, insofar as it has puzzles, exploration and figuring things out as central aspects. But it's also very much a 3d platformer and spaceflight sim at the same time using flight and navigation / jumping around in various gravity levels on the different planets of its bizarre star system. You are an alien astronaut in an alien star system and a fairly fragile one, you can die in Outer Wilds and in fact you definitely will at some point. Every time you do you reset Groundhog day style, and figuring out WHY / HOW your character keeps resetting to the start of the day is one of the various core mysteries here. You need to figure out what is actually going on. And the weight of the realizations and twists that follow by the end are, emotionally and logically impactful in the implications. The game is good not because the graphics are great but because it has imagination and tells an interesting story. The audio design and music carry a lot of it too. But it's also good because it innovates on what a puzzle game can be in terms of the range of mechanics. It's a puzzle platformer that embraces zero-G and spaceflight simulation and the seat-of-the-pants risk-taking feeling of venturing into space and the uncharted and unknown as fully developed elements.
But the fact that they did the unexpected a bit, in a number of aspects... that's a thing indies can and should be doing.
A lot of the big hits in gaming history were doing something weird. When Will Wright was making the OG SimCity, he was talking about zoning and urban planning and had books on urban planning on his desk and everyone he talked to was like, why? How is this a game? It sounds super boring. He was nonetheless certain it could be an engaging premise. It launched Maxis, it sold a million-plus copies. It pretty much was the origin of its branch of the sim genre, and his title SimAnt soon after, while we don't think of it as such often, was arguably an experimental simple design prototype for Command and Conquer, Starcraft, Age of Empires, the entire RTS genre. No, it didn't have all the different unit types those games had but it was nonetheless a war of sorts playing out on a map in realtime between the player and an AI. And that was at the very start of the '90s. His last stroke of genius may have ben the Sims, which was inspired by his daughter and a dollhouse. He intuited the idea of a 'virtual dollhouse' and managing a family and their jobs, relationships and home decor would have a wide appeal. He was right - by 2002 The Sims had sold over 14 million copies. The point: Don't be afraid to hybridize genres, experiment with unexpected elements that aren't normally seen in a category, or take inspiration from outside the existing gaming world entirely. And don't be scared to run with genres that appear dead. The people at Valve knew they had something unique with 'Portal' (2007) but were also nervous as platformers, puzzle games, and the mix of the two, all seemed to be aging, retro categories on their way out. And so they released Portal as an almost... afterthought in the Orange Box. Hoping the combo of other things would sell it and Portal would simply be a nice little bonus. They were shocked when it became the most memed, most discussed part of the entire package. And when Portal went from afterthought to flat-out hit and pop culture phenomenon, it got a sequel and numerous other mind-bendy platformers would follow soon after with various other clever twists on the 'dead' puzzle platformer idea, from Braid to Fez to the likes of more recent titles that messed with players' heads in crazy ways such as Antichamber, Superliminal and last year's Viewfinder.
Point is, if you've got a cool idea and you're like, yeah, but isn't this genre dead? Well, if you come up with a new take on it, go for it anyway. You might be the one who brings a category roaring back into the forefront of pop culture. Like... I'm thinking about 'Star Wars' and how 20th century Fox tried to shut its creation down weeks before it was done. Because it was running over the planned $11 million budget. And they were scared it would lose all that money that they had spent on it and would rather axe it, kill it entirely than allow the 'bloated' movie to cost another million dollars or so. The movie barely made it to completion, simply by rushing like crazy to be 'done' a day before the project was slated to be axed. Which left George Lucas yelling at people until on the final day he completely lost his voice, and his wife completing the final edit of the film down to the last minute filling in VFX shots as they were completed in an epic string of all-nighters. This movie barely made it to theaters, and only managed to get there by sheer force of will by Lucas and his wife, the only two people who really believed it could be good. Star Wars was expected to fail catastrophically at the theaters to a largely now forgotten Burt Reynolds flick that came out the same week. It ended up breaking the all time prior box office record for the science fiction film genre - 2001: A Space Odyssey - by over fourfold.
Point is: Nobody knows what the hell will succeed or not. If people say the idea is dumb and you still believe it has potential, stick with it and nail down why you believe in it and do it anyway. That's the power indies have over the publisher based system. Indies can try things nobody else will dare try. And those things, if they genuinely have merit to enough people, can explode to become huge phenomena.
Don't forget that Minecraft is on of the biggest gaming properties of all time and its first release was very much an experimental little indie thing. Don't forget that the single most wishlisted game on Steam RIGHT NOW is "Manor Lords' - wishlisted by millions of people right now - and it's a solo indie gamedev project by just one talented guy who is absolutely working his ass off on it because he believed in it even before any of the wishlisters knew it was a thing in development.
I may not be pushing for these massively brilliant twists. I'm more of an... aesthetic twist person. I take existing genre types, tried and true ones, and try to do it well, and make it just a little unique via distinctive aesthetics. Because that's a tip you've got to consider too, which is innovate in areas you know you can innovate in. I come from an art background so I lean on that most, the coding side of things is where I sort of struggle and fiddle with things a lot until something works to complete each feature I want, and it's usually not optimally efficient in that side of stuff. I can write and direct video, and do VFX and 3d art, and have a soft spot for analog old school aesthetics sometimes. So when I approach a project like Miniature Multiverse, a 360-degree Mystlike (think Scratches, Schizm, Myst 3/4 interface but with large scale very realistic looking miniature models as the basis of the graphics from start to finish) which is still ongoing, it's basically an idea from 2010 that built to a small prototype test in 2014, just to see if it was even possible to do, and that ended with a failed Kickstarter that got a rare 'Projects we Love' designation from the KS staff but not enough backers to make it. And then circa 2016 I shifted it to Unity and revised my test system a bit and then production began again a little bit at a time in fits and starts. 750+ hours of work and $600+ spent later, I still feel it's only 2/3 there, but then that's one of my failings isn't it? Because I've got so many things to do here and I'm rotating between them all. And if you look at my itch profile, or my Etsy shop, or better yet both, you'll see what I mean.
There are always too many ideas and too many projects for me. It's my biggest problem.