Mipey gets productive!

So, yeah. I decided to lead this as a devlog, because this is what outlines my current project. Any constructive input is welcome.

Anyway, after a period of counter-productivity, I have decided to sit down and get started on a project of choice. However, I am terrible when it comes to choices. I am easily distracted... oh hey, a butterfly! Ahem, as I was saying, I am easily distracted, while I am working on a project, I tend to get unrelated ideas which beg me to work on a wholly different project.

This ends here. I have got to prove myself that I can be productive (and contributive), however I've got to pick the topic for next project out of haystack of ideas.

That is where I am abusing you lot. Hey, where are you going? Get back here! I just need to cast your vote, opinion, whatever, so that in the end I can disregard all that and do something unrelated instead.

Ahem. So, before I list the options, let me state that I am a novice yet prolific writer, I don't easily run out of ideas. So it is no wonder that I want my projects to be full of content (story wise). Since I am a novice game developer as well, I should be doing something relatively simple. I am familiar with various game play elements, however, so feel free to try me.

Anyway. Here goes.

Space Opera 1 - The humble beginnings; the idea revolves around mankind's first serious steps into exploitation of the space.
The player would supervise a foundation and development of asteroid mining expedition, which would eventually lead to foundation of colonies and finally the first conflict. Lots of moral situations. Emphasis on simulation of fragile space economics, however that may be too difficult, so it would most likely just be an adventure.
Gameplay: real time top down 2D (think Solar Winds), preferably turn based (think Space Rangers 2).
Low chance of completion (simulation is hella difficult to develop).
Space Opera 2 - Part of a grander setting, this one tells the story of a lone pilot sent beyond the solar system.
Voyager (the probe) like, only with real flesh and a huge pile of database (think Internet saved onto one location) with a conforming AI to keep him entertained on the lengthy journey. In an attempt to make the AI a more interesting conversational partner, the pilot's reprogramming attempts leave him with severely damaged database and a brand new, yet dumber AI. Which turns out to be a genuinely learning AI with the mentality of a baby. The adventure takes them through boring void, an uninspiring extrasolar system and... uh, now that is an interesting part of the story. An adventure focused on the story telling.
Gameplay: Undetermined (adventure with the engine from above project or visual novel-like)
Medicore chance of completion (depending on the implementation of game mode; if it plays like a linear visual novel, it is easy but boring)
Space Opera 3 - More of a space shooter.
This game would focus on the grizzled war veteran, the last rebel to be defeated, who now tries to get himself killed in popular arena battles, observed by billions of bored civilians. On his last unsuccessful attempt to get himself blown up - all those daring youths are just clueless - the guy is approached by a strange group of civilians pleading for the practical use of his skills. Basically as the uber-skilled pilot riding an outdated space fighter, you escort a group of civilians fleeing the oppressive regime to settle down in unclaimed space. A bunch of missions within context. Oh, and a thickening plot, leading to the player fighting against a whole out-of-control government, while trying to rack points up with the audience via TV show. Literally. For style, taunting, humor and simply explosions. Think Truman's Show, only in space, with intense action and, oh, the guy very willingly participating in it rather than escaping the attention. As long as he gets to kick some butt and enjoy doing so.
Gameplay: Engine from the first project (top down tbs/rt 2d shooter)
Fair chance of completion. (Fairly simple to create a 2D space shooter)
Tower defense 1 - yet another of those tower defence games, except that I want to try a twist.
Instead of getting points, money or such, the player has to salvage the defeated enemies and assemble his defenses from the salvage. High level of customization, low supply. Make too powerful weapons, the enemies get blasted so hard they yield little to no scraps. Balance between power trip (the build-meanest-weapon-ever aka humongous-sword-spikey-hair syndrome) and the need for resources.
Gameplay: Tower Defense (it is pretty much a given)
Medicore chances of completion (balancing issues).
Tower defense 2 - More like a RPG game with TD elements.
The player only has a limited number of characters, which he can develop based on their performance in the battle. Classic classes, etc., acquisition of extra manpower between battles, inter-character relationships, scenes, multiple endings. Yeah, a typical RPG blended with typical TD. Probably an ideal project to try my custom character development system ideas out.
Gameplay: Mix of TD/RTS (the characters would roam within their area they are assigned to, etc.)
Fair chance of completion (sprites, balancing)
RPG 1 - Much in the spirit of above project, the game is focused on character development.
Except this is more about team turn based combat game (in spirit of X-Com) with indirect control of characters (through your main character's leadership). Story is basically like this: war veteran stumbles upon a hapless band of crappy would-be-mercenaries, takes pity on them and educates them how not to hold the sword by its edge etc., until during the climax the player can only watch those same AI-driven companions use what they have learned to rescue him. The game is won if the team manages to rescue the player. Basically the player trains the AI companions; the success of game is determined by this.
Gameplay: Overhead/projected side, pretty much standard tiled RPG heavily slanted on dialogs
Low chance of completion (difficulty of AI programming, sprites)
RPG 2 - Old school adventure RPG.
Focused on one to five characters, hack'n'slash, quests, puzzles etc., rich and engaging plot. Various plot ideas, may share the RPG 1 plot. Nothing inspiring gameplay wise, other than a custom battle system which relies on character's skills rather than player's motoric ability to bash the button.
Gameplay: Old school RPG?
Kinda low chance of completion, due to the overwhelming content (sprites, balancing).
Puzzle - A simple puzzle game. Perhaps Sudoku.
Gameplay: point and click. So easy a monkey could beat you at the game.
Highest chance of completion. (Fallback option in case everything else proves to be too difficult for my abilities.)

To help with suggesting the most appropriate topic, I am listing my abilities as best as I can tell:

Programming: Fairly basic (fair understanding of game development, underlying elements etc., but no skills to speak of; luckily Construct is the fool's tool, ideal for fools like myself)
Art: Programmer's art only (and yes, it is ugly)
Math: Below average (I have a fair understanding of maths, but for the life of myself I can't wrap my mind around more complex equations, functions etc.; splines cause my headache, though I can eventually - through trial and error - manage them.)
Game design: So-so (I tend to know which ideas are feasible, what can be implemented and what is just a feature bloat or nightmare to implement)
Writing: Fairly good (in my own opinion, at least. Even if grammar and spelling may be flawed, I make up for it with the content.)
Commitment: Rather low (I tend to get distracted or drop projects when my spirit is low)
Attractiveness: Hot damn sexy (Too hot for my shirt)
Previous work: Grid Defense, a simple TD prototype not worth mentioning. Just a proof that I can make something that is more or less playable. (It may require August 2008 DX9 optional libraries).

And yes, it is a single man project, as I don't want to waste anyone's time until I feel more confident about my abilities. Feel free to use any of my ideas; they are worth nothing without any sort of implementation, at least. But it would be appreciated if you would nudge me into the right direction and whip me until I spit a half-decent game out. I do need an inspiration and encouragement.

Now then, now that I have bombarded you with all this text, do please tell me: which project should I stick with?