This is the actual final boss of Bruce's Story.  Bit of a spoiler really, but I figure there will be three games in the saga, so it's not giving away too terribly much in the grand scheme of things. :-)  In the demo, all your skills are set at the highest level, as if you leveled them up fully in-game (skills level individually in Specs), and Scott is set at level 50, equal to the boss.  And in case you were curious, the one-on-one final battle is 100% canon.

I want to know how you pulled off a zombie/tonic drive, though!  The AI is set up to do everything in its power to make that strategy extremely difficult to pull off effectively--curing Zombie, healing himself, insta-killing you if you press the issue...  Not that it can't be done, but it's designed to be a very risky proposition.

Also, overall, what did you think of the AI?

I couldn't keep that tactic going, if you really wanted to know. What does Flare do? I think it has something to do with it.

Overall I think the AI is really quite good. I'm impressed! I like how you have the ability to scale the difficulty up and down, which is a good design strength. Tells me you made a lot of early on considerations which is nice. Pokemon's battle system to me felt ad hoc, whereas yours was carefully planned and it certainly gives a better impression on the player when you realize attack-attack-attack doesn't quite cut it.

Flare is just a standard fire-element magic attack.  Unlike in FF, where it's one of the most powerful magics there is.

So yeah, Robert's attack patterns were carefully balanced for fairness over several play-testing sessions, so I can be sure any difficulty concerns at this point are likely due to the numbers being off somewhere else.  For example, in my original, way-too-hard demo, Frostbite didn't do enough damage per hit to make it a viable tradeoff compared to Ignite, its opposite, and Defend/Guard left you wide open due to its rank being too high.  Merely adjusting these parameters was enough to bring the difficulty back in line without touching the boss's AI logic at all.

As for "attack-attack-attack", yes, Robert's attack patterns were purposely designed to punish that kind of strategy.  What the hell kind of final boss would he be if you could just blast him into oblivion with no strategy at all?  Not a very good one, that's for sure!   And believe it or not, you actually have it easy here--I gave you a free Scan in the demo.  In the final game you won't be able to see the final boss's MP and status afflictions like that, since Scott never learns the technique that reveals that info.  The battle is much more of a challenge without constant access to that information.

Wow....

I am at my parents with my Windows PC now, and I played your demo... Wow. It is awesome. I like the music and the gameplay, although I have no experience with it so I fail in battle xD (I know nothin of combos).

Keep up the good job!

Haha, yeah, that's the downside to releasing a demo of the game's final boss (well, other than the obvious spoiler potential of course! )--you're kind of throwing everyone who plays it to the wolves since it's meant to be the most difficult boss in the game.  But I figured hey, that's the best way to demonstrate the full potential of the battle system, plus I can show off with the AI , which is something I couldn't really do if I were demoing an earlier boss.

Just because I'm nice and I really want to see someone beat this boss, I'll give you a hint as to why you're probably losing: I don't know if you've ever played FFX, but the battle system here is essentially an enhanced version of it.  It's called CTB (Count Time Battle or Conditional Turn-based Battle, depending on who you ask) and is turn-based, but not in the traditional sense: turn order is variable and asymmetrical.  You can create more turns for yourself relative to the boss by using lower-ranked attacks, which is how Robert sets up those Quickstrike combos in the first couple phases (QS is Rank 1).  The rank is displayed next to the attack name--that's the little box with a number in it (the color of the box indicates the element, which barely matters in this case as Robert has no affinities).  You get reckless and start throwing around high-ranked attacks without thinking?  You will be killed, that I can guarantee.

Here's a link to the Readme file on GitHub that explains a little more about how the battle system works, and also explains the different status effects:
https://github.com/fatcerberus/spectacles-i#spectacles-bruces-story

Ive played a lot of games, that have debug rooms you can get to that were obviously used for battle testing?

Usually theres a way to toggle abilities and stats prior to a battle starting, then there are some custom enemies designed to use as many aspects of the engine as possible?

I wouldnt say "the final battle" is a requirement, just for testing (Hell FF7 had some giant pyramids).

But ill check it out when I get in

Well technically this is just the final battle of the first entry in the saga, so not terribly spoileriffic, just a little. :-)  But I picked this battle in particular to use for a demo because it was the best way to show off the powerful AI.

I've started working on this project again, now that miniSphere has become a more-than-worthy successor to Sphere 1.x

The first thing I did was to go through the codebase and update everything to take advantage of ES6 awesomeness wherever I could.  Prototypes became classes, string concatenation became templates, arrow functions are used where appropriate, you name it.  The end result is that the code is much, MUCH more readable and ultimately, maintainable.

Just to illustrate what ES6 can do for a codebase, compare this (BEFORE):
https://github.com/fatcerberus/spectacles-i/blob/aa9775a420004711e7b39c4807d77b5fd2034edb/scripts/Stat.js

To this (AFTER):
https://github.com/fatcerberus/spectacles-i/blob/master/src/battleEngine/stat.js

I don't even feel that I need the long-winded comment blocks describing the methods anymore, that's how expressive JS has become.  It's awesome.

Having finally gotten around to resuming work on my flagship project again, I would say that my goal of making the Sphere v2 transition as painless as possible has largely been realized in practice.  Things can be upgraded locally as you learn the new API without touching large swaths of the codebase at once, and the new stuff works seamlessly with the old.  Which is important, because the Sphere v2 API is such a paradigm shift compared to Vanilla's API (<-- my retroactive codename for the original engine) that forcing veterans to learn the whole thing upfront before migrating is just not practical.

The thing is, I don't know how many extant Sphere v1 projects there actually are at this point.  But rest assured that for whatever ones there are, the transition will be a smooth one.

Anyway, I'll post some more about the game soon.  In particular I'd like to lay out how the battle engine has shaped up, as it's ended up surprisingly complex of late.  There's a lot of different battle mechanics now; more, even, than I envisioned in my original ideas.