Yeah, melee's running a bit behind the curve in the present build. Part of this is that the abilities offered for starting characters aren't all that great for melee, as it turns out. I think the new relics will address that somewhat. The stat issue is a good observation: I've been considering giving some melee-specific benefits to the stats they often need (for example, giving Agility a boost to physical damage too... less than Strength, but enough where it isn't so much of a 'tax'.) The differences in physical vs elemental weaknesses is kind of on purpose (though that doesn't necessarily mean it's a good idea!): physical types tend to be resisted/weak less, never (so far) immune, and even when resistant or weak, the effect of these is muted compared to elemental types, which are basically useless even against resistance, but which get amazing results when aimed at weaknesses. The idea was for physical elements to be a bit more stable and consistent, with the elemental types (Fire/Ice/Elec, specifically) being ones that depend more on matching spell to opponent.
The new relic that features the "magic" side of Light as a primary will be getting Punish, rather than Exorcism, so hopefully that will work out better. Exorcism will still be on the secondary though: in the new system you can take the same element twice, having Exorcism would let you "first strike" weak to Light targets rather than having to wait for them to attack (Exorcism also tends to hit considerably harder than Punish, which helps.)
Support starts do tend to be pretty strong, but I think this is more because many of the other options weren't as strong within their category as support was. With the relics I have tried to bump the others up to where support was, which will make the early game a teeny bit easier, but I think there's room for that: the bottom 200 or so (out of 500 or so) entries of the scoreboard are largely Tower:1 deaths. Adding in just Tower:2 takes you to 300 or so. I want even the early floors to be tricky... but 60% of all games ending on the floors where encounters are specifically weighted to be "easy wins" suggests I have some room to back off on that.
The observation about player SP vs. demon SP is interesting. It's true that your demons, collectively, have 800-1000 (depending on Tireless) SP total while you only have 100-125, I don't think I've heard anyone else talk about it in terms like that before (thinking of the demons' total SP as one pool, I mean) and choosing strategy accordingly. It seems to have worked out though.
I think I see why you maybe weren't drawn to Refresh too: with Refresh, some of the demons are Refreshing you rather than attacking, so while you get to enjoy their SP pool, you're losing attacks per turn. Maybe I need to make Refresh a free action like Draw Wounds... that way you get SP, and they still can attack themselves. I'll have to think about this one some more, it is definitely a great observation though.
Thank you.
(Small aside: One of the few definite advantages melee builds do have is they pretty much never have SP issues. Melee abilities are much cheaper to use SP-wise, and melee also has better Stamina recovery options ability-wise.)
Magic/Cunning builds do seem to have a lot to recommend them... It's possible I'm not assigning the right "cost" to "ignore hit roll" in terms of my balance equations: with a lot of this sort of thing I do the "proper" math as a starting place, then have to adjust it constantly based on the things I couldn't calculate for or didn't account for properly. Even though the trip from Tower:1 to Tower:17 is supposed to be the "early game", I don't really want there to be easy/boring parts that people feel like they're mindlessly repeating after a death. It sounds like my problem exists somewhat on both ends: Magic/Cunning might be a bit too nice, melee might be a bit not nice enough. I've had thoughts on this direction before now, this is just another bit of feedback for that stack. I'm working on solutions, some of which I went into above in this post where I first mentioned melee, and I'll make a post about this specifically when I have some more concrete proposals to throw out.
I am happy that T:17 at least was finally able to turn up the heat again. T:17 has a history of that for some reason, even though it was not specifically designed to be ugly (after all, in terms of the final content, T:17 is just another floor, it only happens that right now it is also the end of the current implementation.)
In short, I probably need to bring melee/physical up a bit and ranged/magical down a bit. Status effects may also need some poking too.
Thanks again for typing this up for me.
Feedback is the most important thing at this stage, and I really appreciate you taking the time to write this out.
I hope to repay the favor by putting the information to good use.
Bonus (like this post wasn't long enough): I'm friends with chillblain, one of the other winners, from pre-Demon days and have talked to him a fair bit about his win. If you're curious, his strategy was Vitality-based. He loaded up mostly on Vitality, took every reaction/survivability ability he could find (Fiery Veil, Shocking Veil, Arctic Veil, etc.) and waded into enemies more or less daring them to hit him, which of course, they were happy to do, much to their sorrow. His demon stable was 4 healers, 2 ranged DPS, 2 melee most of the time: he'd usually start fights with the ranged DPS and 1 healer out, shifting around as necessary (usually bringing in more healers as the first one ran out of SP, or switching to the melees in long fights that'd run the ranged DPS guys out of SP.)