Listing out some ideas I have for what I'm thinking of making. These are in no particular order right now.
1. Arena / Twin-stick shooter: Going with a typical player in the middle (either a man or a ship, ship is easier to animate) who can shoot in a full 360 degrees using either mouse aiming or the classic 2 sticks on a joypad. Enemies would attack in various patterns. I've prototyped this playstyle in C++ with OpenGL in the past (image of prototype)so using that as a reference along with using my existing C#/OpenTK library means this should be easy to get up and running and so would make a good first month project.
2. Single room platformer: This would consist of a simple platforming engine in which each level is a single room on the screen (no scrolling). The original point of this idea was to allow me to experiment and learn about level design. Exact specifics about each level are unclear right now, whether there are collectibles/puzzles/enemies/all of these and the number of levels depends on time constraints.
3. Scrolling jumping platformer: The two more well known examples of this genre are Canabalt and Robot Unicorn Attack. Adding multiple levels you drop down to with the higher ones being quicker might be an interesting twist to this. Building the levels randomly (and consistently) will be an interesting challenge.
4. Ship management: Something similar to FTL, but much simpler. This would be more a prototype and I think it's one of those genres that a lot of devs would like to experiment with.
5. Side-scrolling shooter: This is a genre I've made a couple of simple Flash games of and I experimented with it using music to generate enemies, but I've never put a lot of thought into a comprehensive game in this genre. Several levels with various weapon/ship upgrades and limited lives. Not bullet-hell, more casual.
6. Survival: Something similar but simpler than Don't Starve. Mouse controlled movement, simple crafting. Again more a prototype but something to experiment with.
7. Hoshi 2: I love this game but I feel it's missing something and I'm only sure of parts of that. I would like to update the UI and sound effects especially. Some sort of online score system would be nice. At a stretch I'd like to add multi-player but I'm unsure how easy that would be to do in a month. I want to add more game elements as well and update the AI. It's extremely simplistic atm and overwhelms in a horde-style instead of actively hunting you. I'll write more about what I want to add/change if I decide to pick this up as a game.
8. Fixing Whirligig: A lot of Whirligig's problems are a result of some questionable design decisions when I first started it. Sometimes player rotation goes out of whack and there are pauses as the new planet bitmaps are produced after every impact. Additionally the physics (especially in bouncing) are a little odd at times. The amount of refactoring and fixing required would mean this would take as much time as a new project.
9. RTS/TD hybrid: This would consist of a small group of 'units' that you move around a map, gathering resources, defending yourself and building more units whilst being attacked constantly. I'm unsure how well this will work and may require a lot of tweaking.
10. Procedurally generated exploration: Not exactly a game, but it would create a landscape you could explore procedurally (that is, on the fly, computationally and consistently). The reasoning is that the Survival and RTS/TD hybrid ideas above would require such code before implementing the gameplay specifics for those games. Whilst I doubt this would be a huge amount of fun, with enough thought behind the landscape generation it should create some interesting scenery. Separating it out like this means I'm more likely to spend time fixing issues with the generation before moving on to gameplay elements which should produce a more robust system I can then use for more games.
11. 3D puzzle style game: Something akin to Marble Madness although I'm thinking closer to Arcturus (an old RISC OS game) with varying obstacles, different surface types, slopes and enemies to navigate.
12. Fixing/finishing Five Point Five: Five Point Five was a game I started for a TIG competition regarding multi-player only. Technically it works, although there's no input set up for all 5 ships and it's local only (looks something like this as I left it). The base gameplay is to destroy the other players, however you can't damage them until all the asteroids are destroyed. You can however push them around with your bullets. The asteroids drop power-ups like rapid-fire, power-shots, faster shield recharge. Once the asteroids are destroyed you can damage each other and whoever is left at the end wins. Has plenty of particle effects and you can see the heavy influence from Spheres of Chaos. Fixing and finishing it would require adding net play (which would be the biggest challenge) and a proper menu with rebindable controls. If you plug in enough joysticks it will work fine locally but having a couple of computers for it would make more sense. The levels are set up internally in such a way to reduce lag issues - they're based on a single number and procedurally generated from that, with all the sub-asteroids based from that original number, meaning that clients only need a location and time for a split and not all the individual asteroid information. Moving the game on from the original spec to include AI players would also make things interesting.
So those are 12 ideas. I don't know whether I'll use all of these as doubtless I'll have more and some aren't hugely interesting. There's also various puzzle games like Tetris or Puzzle Bobble that I'd like to clone and would be pretty easy to do.