Transmutations Reform
DCSS Transmutations are both a spell school, a background, and a playstyle. The primary focus of all three is supporting unarmed combat via various 'forms', which meld armour in exchange.
There's quite a few things about transmutations that players enjoy:
- Form-based gameplay is evocative (you're a shapeshifter!) and distinctive.
- There isn't a lot of other support for unarmed combat, so players enjoy the opportunity transmutations provide to use it.
- In the late game, transmutations can be very powerful, and players like powerful things.
- Playing a transmuter requires mindlessly re-casting the same form spell many, many times throughout a game. It's generally not an interesting tactical decision whether or not to be in a form, or which one to be in. Makework.
- To make this somewhat worse, good play sometimes involves switching to +int/+wiz/enhancer items, casting the form spell, and then switching back... every time you cast your form.
- It's quite hard to pivot either into or out of Transmutations based on found items, spells and gods. A fire elementalist can transition into another school; an axe-wielder can switch to maces. Transmuters, on the other hand, are generally forced to give up all of their skill training as a waste if they switch to weapons or conjurations. Likewise, the heavy skill investments in UC and Transmutations required to use forms make it hard to justify switching into Transmutation before the late game.
- The itemization game is less exciting. Most weapons and armour are irrelevant to transmuters, especially later on, since forms meld them. This means there are fewer moments of excitement at upgrading your kit, and more irrelevant scrolls of enchant/brand weapon and enchant armour. (It's unfortunate in this respect that the 'no weapon' playstyle is the same as the 'reduced/no armour' playstyle.)
There are a few commonly discussed options for addressing the issues above.
The most seemingly straightforward approach is to make form spells permanent once cast, unless dispelled by an enemy or left via ability. (This addresses, specifically, the 're-casting' issue.)
This makes the 'spell success' stat even less relevant than it is for form spells at present, since players can re-cast until they get a success and then keep that result indefinitely. To solve this, the usual suggestion is to 'reserve' mp while the form is active, reducing max MP by an amount related to the spell fail %.
This approach has two major issues. First, it redefines 'spell success' to mean something that has nothing to do with success and works differently from every other spell. This is complex and unintuitive, which isn't great for players.
Secondly, the MP reduction tends to be irrelevant for many characters. For much of the game, most melee-focused characters don't really care whether they have 5 MP or 30. This implies the need for a whole ecosystem of other effects to draw transmuters into caring about MP, like other 'permabuffs' that similarly reduce max MP, or other tactically useful spells like Manifold Assault or Vhi's Electric Charge. This is a large design ask.
For both of these reasons (especially the first), I'm disinclined to pursue this approach. It does seem like it could work in some version of Crawl, but it feels like a kludge rather than something that fits naturally into Crawl's spell system.
Crawl spells work best when they are tactically situational. If a spell is just always the right button to press, or if it's good to cast outside combat, then it becomes makework. If, on the other hand, the player finds themself weighing options ("well, Conjure Ball Lightning would be too loud here, so I'll cast Borgnjor's Vile Clutch instead...), then the game becomes dynamic and exciting.
One can imagine forms which work in that way. The player might choose one form for a fight against a group, another form for a fight in open space, or a third for a fight against spellcasting foes. They might even switch forms mid-fight. This would create dynamism - rather than simply casting the same form for every fight, each fight would instead present exciting decisions about the best form to use.
It's a very tough ask, though. We'd need a whole slew of 'tactical' forms, spread across a variety of spell levels and providing interesting tradeoffs with other forms of similar level. We currently have none, despite years of thinking about similar issues. Wereblood and Ozocubu's Armour might be usable as the base for two forms... but that's speculative, and a whole playstyle would need more forms than that.
Tactical forms do seem like they could work in some version of Crawl, but they're too daunting a design challenge for me to attempt.
At present, forms function more like equipment than spells. The player picks a form and sticks to it until ready to upgrade to a higher-level one; they put skill in and gets stats out. This is very broadly similar to how weapons work.
It seems reasonable to formalize that. Instead of forms being spells, they can be talismans, a new amulet-slot item type. When worn, these transmute a player who has sufficient skill into the corresponding form, melding armour and providing various bonuses. One could have a 'scale talisman {dragon}', a 'skull talisman {lich}', etc.
This is a radical change, but I think it is worth trying. The remainder of this document goes further into detail about this approach.
There should be some skill threshold before a talisman is usable. This helps save room for lower- and mid-level forms - even if you find a high-level form early, you won't have the skill to use it for some time. This threshold should be clear to the player.
Further, there should be a reason to keep skilling past the minimum skill requirement for a talisman. This allows more progression options for the player, and makes skilling 'softer' and 'fuzzier', matching the spell and armour skill games.
Each talisman should use Transmutations skill, plus optionally another school (e.g. Ice or Earth). The average of these skills will determine the talisman's power. (Intelligence, Spellcasting, wizardry, and armour/shield penalties are all irrelevant. This simplifies Transmuter statting and skilling, especially early on.)
There should be a minimum power for each talisman, below which it either reduces the player's max HP or simply does not work, and a maximum power, so that low-level talismans don't scale indefinitely. Power should primarily affect the damage bonus the associated form provides, and occasionally provide other incidental bonuses. This allows players to either focus on their melee skill (e.g. UC) for more damage + attack speed, or focus Transmutations for more damage + eventual access to later forms.
To allow easier transition between Transmutations and other playstyles based on found items, there should be a mix of UC-focused and weapon-agnostic forms. We should re-use existing forms that are exciting and evocative, and replace others.
Tentatively, we could use:
- Weapon-agnostic forms:
- 'Starter form' (melds auxen, perhaps themed as a 'beast form' that makes your extremities swole + distressingly furry & gives +slay)
- 'Dancing weapon form' (melds everything but weapon / shield / jewels) [perhaps 'hollow form' that only melds the body/cloak slot instead?]
- Statue form
- Lich form? (Reworked to be more effective in melee?)
- UC-only forms:
- Ice form (perhaps rethemed to be more exciting)
- Blade Hands
- Dragon form
- Storm form (optional)
Spider form isn't particularly exciting thematically or mechanically, so seems fine to cut it.
Transmuters will start with the new 'starter form' talisman. The goal is to have something which leans toward UC, but which still provides some value for a while if the player pivots into weapons.
They can also still start with Wereblood. Should add extra fun. :)
Beastly Appendage and the other form spells are gone.
Talismans should be moderately common. Transmuters should be likely (~90-95%?) to find at least one mid-level form by the time they reach Lair, and an end-game form by perhaps some time in S branches or Vaults.
It's plausible that, to accomplish this, we'll need to play with vaults or to make some monster types drop talismans. We can experiment - initial tests suggest that using about ~1% of all item weight is enough to accomplish this.
Minor: it should also be possible to gain talismans through acquirement.
Instead of having talismans take an equipment slot, they could be usable from your inventory, probably as an e(V)ocation (not taking Evo skill) but maybe as an (a)bility. We'd probably want this evocation to take a long time to use (5-10 turns?), to avoid inventory clutter from carrying around many swappable talismans. There would be no failure chance, and the form would still be permanent once set until the player cancelled it (probably with the same ability they have now) or until it's dispelled.
The main upside of this approach is not messing with Ashenzari scaling and not wasting character overview space on a slot that most characters won't use. It does make randart talismans trickier to implement, but maybe we could set up something special so that randart talismans' properties affect the player when evoked.
Using the amulet slot allow these to work for all species, and limit them to one form per character. (Hence eliminating the question of 'what happens if you put on multiple talismans?') They also make swapping slow, which seems probably good.
We could instead add a new item slot just for talismans. (Perhaps this would be a 'flesh' slot - jam it into your body to transform!) That would help itemization (since transmuters would still get to be excited about finding amulets). However, it might feel bad for non-transmuters, since they'd have a slot 'unused' all game. It would also have implications for Ashenzari (how would piety scale?) and for UI (an octopode would have a very long character overview... especially if they got the macabre finger necklace!). Hiding the slot except when it's not in use might fix some of these issues.
If talismans took an item slot, dispelling wouldn't affect them, and temporary forms from e.g. polymorph or lignification will override them. When their effects ended, the player would return to their talisman form.
Regardless, it should be possible to find randart talismans.
All form spells will be removed.
Non-form tmut spells will be unchanged. At time of writing, that's petrify, irradiate, yara's, sting, passwall, ignite poison, alistair's, wereblood, and bog. We could consider reschooling or removing these spells as well, since most of them have almost no synergy with the transmuter playstyle, but it doesn't seem necessary for at least the first draft of this approach.
- How should we distribute these talisman items, exactly?
- What slot should talismans use, if any?
- How frequently do we want non-transmuters to end up using a form?
- What kind of stats / damage bonuses should the new forms provide?
- How should we theme the new 'starter form'?
- Should we retheme ice form?
- Is this mean to megaziggers? (And do we care?)