These sound like really cool ideas. I'm sorry they didn't quite work out though, that's always a bummer. What were these written in? Like pure TI-BASIC or hybrid?

Illusiat 2002: La Quête Ultime (The Ultimate Quest) (2002)

-pseudo 3D (with some actual basic 3D every now and then) RPG in the style of Star Ocean 2, Legend Of Dragoon and FFVII-IX in terms of map engine style, where battles were 3 chars vs 1 enemy. Story was a bit simple (it involved 4 elemental crystals, lol) but more developed than any Illusiat game prior Illusiat 9. Game was gonna be split in 5 chapters, have about 10 hours of gameplay and amazing animations for pure-BASIC. It was what was supposed to be Illusiat 5. Game was 79 KB and 2.5 chapters complete when Mirage corrupted my archive and forced me to reset the entire memory. I had no mean to backup at the time, except my bro's 83+, which he asked me to clean from the game backup made two weeks before as he needed the calc, so everything but the title screen were lost.

Team (2002)
Following the deletion of Illusiat 2002 in Feb 2002, my motivation was at its lowest. I tried making a dumbed down RPG split into chapters again, but it never turned out to be as good as Illusiat 2002. It was broken as well so I cancelled it. Following that, I stopped calc programming for a few months. Demo available for download.

Illusion of Mana/Secret Of Mana 83+ (2002)
First attempt at a Zelda-style game, using Illusiat 6 engine. Project died as I was not ready to work with enemy collision detection yet.

The Reign Of Legends 3-II: Theola Shire (2004)
An Omnicalc direct sequel to The Reign Of Legends 3, involving the only surviving playable character from the previous game as well as the mysterious guy appearing during final battle. 3rd char slot was just gonna be random characters that sometimes joined your party. Game was going to occur on the moon in 2235 AD, where humans took refuge following the destruction of almost everything on the Earth in 2200 AD. The game did not go far in planning stages and no line of code was ever written for it.

The Reign Of Legends 4ever (2004)
A RPG starting at the end of time where 4 groups of 2 warriors have each their own mission to accomplish. There are warriors of earth, wind, water and fire. Each need to go in the respective elemental worlds. Once each tasks are accomplished, the path to the final battle (and resurrecting the world as it was before its destruction) is opened. At this point you choose whichever character you want. In each world, if I remember, you usually had another person joining your party.

It was gonna be possible to swap between missions and return to visit the respective areas. Items from a mission could be used in the other, so some missions allowed you to complete others .

The project died because it was just too massive to take on. The overworld for the Earth world was only 1/4 finished and it was almost 100 maps. Also on UTI people started whining about the usage of ASCII graphics on the overworld and the usage of ASM libs instead of pure-BASIC, so it killed my motivation.

Reuben Quest III: Terror In The Sky (2005 I think)
5-level xLIB grayscale game. I then realized how hard such sprites were to create. Nitacku grayscale ditherer did not exist back then so yeah. Plus I had no idea for the game battles and to manage to fit everything in RAM (which gave me lots of trouble in the first two games). Also, almost everyone started hating large games with multiple files so I decided against starting this project.

Diortem (2005)
Omnicalc Metroid clone. The reason why it died is the same as the last one stated for Reuben III. However, this project was revived as Metroid II Evolution over one year later and ported to xLIB. Demo available for download. It features different enemies than Metroid II Evolution, like actual Metroids.

Project Omega: Race Ngine (2007)
Lotus Turbo Challenge style game using xLIB. It improved on Lotus in the way that when turning, the background scrolled left/right, but as the speed was already a bit low, I got worried the addition of more features would make the game too slow to be fun. In addition to that, I had some life/online issues at the time that hindered me. Demo available for download.

The Reign of Legends 4 (2007)
Another attempt at a ROL4, this time graphical again. You had to save the world from the approaching void. This world had 6 kingdoms representing an element: Earth, wind, water, fire, darkness and void. The story was a parralel with TI community events at the time as well as its state, although made into a fantasy RPG story. Even the 6 kingdom names had similarities with the active TI-83+ forums at the time, such as Omnitia, the kingdom of darkness, which is nearly deserted early in the game. The game featured a job system like Final Fantasy III for the NES and V for the SNES. Unfortunately, only the menu was finished, along with the job system and cure magic spells in it. Battles were started but not even close from being finished. As for the story, only the beginning and the end were coded. For some reasons I kinda lost interest after a few weeks of coding. The menu and some files are available for download and requires xLIB.

Illusiat 13: The Lost Chapter/The Last Chapter/The Final Chapter (2009)
Final game in the Illusiat serie, worked on over 6 years after the previous one that was meant to be the last came out. Repeated online issues with someone (added up to other stuff I had to deal with in a short amount of time), which I never really recovered from, teared apart my motivation and I cancelled the project during Summer 2009, despite the game being 90% complete (only the final dungeon/battle, ending and an optional dungeon  needed to be done, as well as a few enemies/bosses.). It was revived for a short while and renamed from I13 The Lost Chapter to I13 The Last Chapter, but only some bugfixes occured as well as cosmetical changes during saving/loading. It died afterward again. In 2010 it was revived again but no work was put on it except some planned story changes in my head. Finally I gave the project to another staff, although it is not a priority. Game is fully playable until the final dungeon and available for download.

Unnamed RPG (2007): It was going to be a full RPG with 3x5 text sprites. I got 10 maps and the store done (and a cheat ) and was working on the battle engine when I realized it would far surpass 24 KB when finished. Being hesitant to use xLIB, which I was afraid would add too much to the memory cost, I stopped dev.

Unnamed RPG 2 (2008): Now with a lot more experience with xLIB, I decided to try making an RPG again, but this time of the kind where you walk around with a gun (I actually made all the sprites which IMO were pretty decent ) and shoot enemies when you come to them. The AI stumped me, so I gave up.

Unnamed RPG 3 (2009): Yes, another one. I had an idea to make a pure-BASIC RPG with better compression than I knew of in 2007 and 5x5 sprites. Then I had a better idea for a puzzle platformer, so this got replaced by Absolute Insanity (which I actually finished ). And so I have yet to make a working top-down RPG...

There were many others where I got a bit of work done (or in the case of some, 90% or more done), then gave up for one reason or another (often due to boredom).

Fuel: Old contest entry, but it was super slow.  I lost the source code, but I found it again in a group on my calc last night ^_^

Ballz:  A remake of this game in C for the TI-89.  I think I still have the source somewhere, but I doubt I will ever release it.  I had it complete, with the exception of (1) the stars in the levels and (2) I only implemented about 1/3 of the levels.  Also, the physics were pathetic.  One good thing, the speed was 9000x better than that BASIC version.

DJ, I will probably continue it, even though it is slow.

Yes.  You can run it at 6, but God FSM help you.