This is freaking weird...reading CRPG Addict's blog I see that I commented about P3 nearly 11 years ago, and that I completed the Amiga version of P3. I had completely forgotten that; I thought I skipped it since I already played it back in the 1980s.
This is freaking weird...reading CRPG Addict's blog I see that I commented about P3 nearly 11 years ago, and that I completed the Amiga version of P3. I had completely forgotten that; I thought I skipped it since I already played it back in the 1980s.
Glad I'm not the only one with these kind of memory problems at 42...
What's more interesting is that when I need to re-research something (because I completely forgot I had already done it years ago), then later I find my old notes, usually I discover that I have reached the exact same conclusions and went down pretty much the same mental path again. Eerie! The universe is quite deterministic, after all.
Hmm...I suspect the first two areas being already explored is by design, like it's the party's own turf.
But quite impressive with a fan patch as recent as 2004.
The Apple II and C64 versions came first. Those 8-bit machines used the same MOS 6502 CPU, so probably the game logic was ported over from one to the other unaltered. So I'd say that's the "definitive" version.
It's been documented for the Amiga and Atari ST ports (which happen to use the same Motorola 68000 CPU) that whoever did the 8-bit to 16-bit conversions screwed up a few things with the calculations (check out my earlier posts in this thread for further info).
The PC is again a different 16-bit architecture; the game was likely ported to the PC by yet another team (hopefully from the original 8-bit versions, but who knows...), so either it has its own share of bugs, or the guys who ported it took some liberties with the rules perhaps?
The Apple II and C64 versions came first. Those 8-bit machines used the same MOS 6502 CPU, so probably the game logic was ported over from one to the other unaltered. So I'd say that's the "definitive" version.
It's been documented for the Amiga and Atari ST ports (which happen to use the same Motorola 68000 CPU) that whoever did the 8-bit to 16-bit conversions screwed up a few things with the calculations (check out my earlier posts in this thread for further info).
The PC is again a different 16-bit architecture; the game was likely ported to the PC by yet another team (hopefully from the original 8-bit versions, but who knows...), so either it has its own share of bugs, or the guys who ported it took some liberties with the rules perhaps?
Do the other 16-bit platforms have the introduced changes/bugs from the PC versions that the Apple ][ doesn't have?
The PC version can also occasionally throw a high-level monster at you in a low-level outdoors area. I had thought this was a feature, but it's clearly actually a bug, so I'm planning to fix that.
- Rolling for stats is highly biased towards very low and very high values. This wouldn't be hard to fix -- it's clearly deliberate, and also clearly stupid -- the PC version just rolls 1d20 for stats that are normally 3 to 18. (The Apple ][ has a clever 3d6 mechanism where the dice are loaded in your favor.) Very low-stats characters are hilariously unplayable, very high-stats characters wreck balance, and they're both much easier to roll in the PC version. I want to write a bugfix for this even though this change was deliberate.
- Random encounters in dungeons are far more common than they ought to be in the PC version. This is clearly a bug.
- Random encounters in dungeons are far easier than they ought to be in the PC version. You are supposed to get thematic monsters similar to the fixed encounters. This is also clearly a bug.
I'm also planning to fix the PC version RNG -- it's crude and slow even by mid-1980s standards. (If you're playing at near authentic speed, that's actually what's causing a lot of the slowdown. I can't fix large monster animation, that will be slow no matter what I do.)
The PC version is in compiled C -- it's not written directly in assembly. I'm not sure what compiler, it's a strange looking one even by 1990s standards, never mind 2021 standards. SSI clearly actually wrote their own RNG and write-pixel-to-screen functions, too. (Well, the former was probably actually a good idea in C before about the late 1990s -- the standard C library RNG was notoriously bad even by the standards of awful RNGs. It just wasn't done that well here -- it looks like SSI grabbed some 1970s FORTRAN book.)
Do people think I should revert the PC damage adjustment too, even though that's also clearly deliberate?
I can post the relevant assembly code for any game mechanic if people are interested.
I'd have to learn an entirely new architecture for that -- that's far more time-consuming. This is much less work.
Wow, how many hours have you sunk into this? It's a pity you're most familiar with the PC architecture; it would be great to create an "ultimate" Atari ST version (because that has all 3 parts of the trilogy). I've skimmed through the list and it seems like the 8-bit originals have also their own fair share of bugs, then there are the porting bugs on top of those, plus the "questionable" changes.
I still hope for a Phantasie series facelift in a single engine for 1-4 and a construction set. I am still partial to the c64 P1-2 though.Chris Koźmik
I agree with what everyone above has said regarding dungeons. The only thing I didn't like about LoA were the simplistic dungeons. Simply adding some puzzles here and there, or even some unique text-based encounters would be nice.
Phantasie I has some really great dungeons with unique themes and fun interactions, while still having comparatively simple layouts. I'd definitely recommend checking out that game (or at least a walkthrough) for inspiration.