I'm sure there are tricks yet to be discovered but I doubt they would be significant enough to make any bigger impact on any productions. On MSX1, there has been a quite significant improvement in use of technology and also to some extent music and gfx. But as someone mentioned a way back there isn't really many at all productions yet that combines high use of technology, art, and good game design, so certainly there is still room to make an MSX game that will be better than anything previously made (I have a feeling MOAM could be that killer game 
)
On MSX2 I think there is a lot to discover / optimize / make better and even more on Turbo-R. Problem here is time, to make something really utilize the hardware resources to its maximum requires both a lot of skills and time.
The MSX configuration specs are very low (VRAM: 16 KB, 128Kb rom, etc), it's not very stimulating. The specs should be MSX2, no max RAM limit, etc. 
People,
There are already other competition for higher specs. But yet no entries anounced.
So feel free to enter that one.
But remember that the MSXdev is the most succesful one. And that is not because developers do not want to work with low specs.
But then again I'm a screen2 junkie 
@fixato: Please do.
Maybe I can make one wish for MSX2 game makers... 
Please convert more early eighties arcade games to MSX2, the ones which have lots of multicolour sprites etc, as with MSX2 it is possible to make graphically 100% arcade perfect conversions from those old arcade masterpieces and not so well known non-masterpieces too 
I think that sound/music may be only thing which may differ a bit from arcade original, but that´s minimal thing.
Hahaha incredible!!!!
too bad they are for MSX1 
The MSX configuration specs are very low (VRAM: 16 KB, 128Kb rom, etc), it's not very stimulating. The specs should be MSX2, no max RAM limit, etc.
I agree...the specs of msxdev'08 are too low...
even with maximum compression I know we'll exceed the 128kb boundary....
I can understand putting a limitation on the ram and vram being that not every one has a mapper and not everyone (might) have a msx2 but I really dont understand the need to limit the rom size??? even a romsize of 1024kb will run on a MSX1 so what's the big deal?
It's an endless debate. The Dev organizers generally argue that larger games lead to larger projects which lead to a bigger chance that a game is not finished in time or even worse: eventually canceled.
I on the other hand tend to argue that most of the bits 'n bytes of a large game are made out of artwork, which may -depending on the person who makes it- take little time to make. About 40% of MJTT is music data, just to name something. I was able to make a complete MJTT level tileset in an evening and there are countless more examples why artwork may consume tons of memory while not very labor intensive to make and while not being out of place in the actual game.
Map editing is another thing. With the proper map editor you can make huge maps relatively quick. In my opinion it's quicker to draw a new level tileset than to code a new semi-advanced game feature.
There may also be the argument of preventing 16KB minigames having to compete with oodle-size megaroms, but I'm not sure how relevant it should ever be. For dev'05 the two 48KB ROMS (maxed size back then) were also considerably larger in scope than the 16KB entries. So, 48KB or 128KB or 1MB, a larger game is larger than a small game, you can as well make the specs completely free.
Then again: their rules. For other rules: other contests..