Angband Philosophy II: Magic

[Jungle_Boy]

I played in the most recent competition quite extensively to get a feel for possible other magic systems. Some things I liked.

Being able to choose your magic realm
Some classes get more than one realm
Several realms to choose from
Different specialties between realms
Lots of different magic using classes to choose from
Being able to learn a spell multiple times to increase efficacy
Only four books per realm

There were several other things I really enjoyed about PosChengband but those relate specifically to the magic system.

My only complaint was that there were too many spell choices when playing the sorcerer and too much overlap between some of the realms.

My suggestion for Vanilla would be to split the all spells up into 4-5 realms.

Offense
Defense
Utility
Healing
Status/Summoning? (This would require significant changes to make it worthwhile)

Pure casters would get to choose two realms and hybrids would get to choose one. Spell failure rates and minimum level could be used to differentiate the spells across the classes and there would be little to no overlap between realms. It would be possible to learn spells multiple times to increase spell power, perhaps three levels, and there would only be 3-4 books per realm.

[Philip]

The way I imagine magic, ideally, is a certain number of realms, with their specific specialties, and weaknesses. The number of realms can be expanded, though no realm should be too similar to any others, with mages getting two full-strength, or one as very strong, and half-casters getting one, probably watered down a bit.

Spells, to remove book-clutter, would be found in the dungeon, and you would only see spells you can learn. You would then inscribe them into a book, because I would like to introduce limited spell slots and preserve limited inventory slots as issues for characters to deal with.

1) Arcane, most similar to what mages now get out of Raal and Kelek, with lessresistable reasonably high damage with side-effects occurring. On the other hand, most other things would have to be handled with items. Arcane would have more single target spells. I imagine stuff like Chaos Strike, some form of shard damage, like maybe a Rocket (yes debo), something Roar, Rift, Plasma spike.

2) Natural, mostly just *element* bolt/ball type stuff, choice of spells would affect damage more than side-effects. Natural spells would run in beams and balls, cloudkill style. Perhaps even cumulative damage, the way NPP does it for stinking cloud. Still rather weak for utility, like above. Spells would include stinking cloud, * bolt/ball, meteor swarm, some stuff lifted from the Druid classes in NPP and O/FA.

3) Holy, which would have weak, alignment-based damage, but solid healing, and maybe okay buffs or something. Would need to get okay at melee, or use spells to chip away at monsters. Orb of Draining, Healing, perhaps some form of SP regen, stat recovery, XP recovery.

At this point, it gets tougher. I have some ideas, but I suspect I have one too many classes. One of these should be split among the others, but I don't know which. Necromancer strikes me as something that should be present, but I have no idea how to implement it practically.

4) Artificer, which would excel at using and improving items. Would have perfect reliable recharging, some way of making other magic items more effective, as well as equipment (enchantment, except good). Would have to manage inventory carefully and choose a damage-dealing method. Typical spells would be Recharge, a good version of Enchant thing, dunno, we'd have to invent a lot of stuff.

5) A Buffer. Spells are independent of inventory/equipment, have a certain duration, and would make the caster more powerful while active. Maybe half-caster level melee and ranged combat, with defenses being top-tier in all categories. Would be weak once those run out, and wouldn't ever be too great at dealing damage anyway. Typical spells are Resistance, Shield, Haste self, Heroism, Bless, maybe even a weaker form of Globe of Invulnerability.

6) Modifier, who knows what is happening, and is capable of getting where they want to be. Would have perfect detection, excellent teleportation of self and monsters, even stuff like Banish, Destruct and Door/Stair creation.

You can't rework spells without a full class rework, or at least a plan.

Current classes

a) Warrior, would get affected the least. Power needs to be scaled appropriately, which is probably doable.

b) Mage, the Mage/Priest distinction makes no sense in Tolkien, so I would remove it. Magic is part of the world, accessible to gods and to mortals, and those in-between, though mostly the gods. In any case, either simplify Int/Wis into Will, or keep them separate with different effects, but casters should all use the same stat scheme. A mage would be built by choosing a damage category and one of the second group of categories.

c) Priest, see above, a priest would be built by picking holy and buff magic.

d) Rogue, would remain similar, except they get to choose between categories 4 and 6, essentially. Rogue spells are shit now, so this would be quite a boost, but hopefully not one that couldn't be compensated.

e) Paladin, would get to choose between 3 and 5, otherwise would remain the same. Gets a bit weaker, potentially, but can be tweaked as necessary.

f) Ranger. In Tolkien, Rangers are confusing. They are adept archers, swordsmen, and apparently have some rather advanced survival skills. I'm not even sure they should use magic, but maybe something more like techniques. Techniques would be weak, but only have time cost. Spamming would be useless. So, forage for remedies, build camp, track monsters, that kind of stuff. I imagine these as incredibly versatile, but ultimately, not too specialized. Damage output might suffer, but defenses will be more solid.

New classes

a) Archer. Ranger simply cannot be the only class capable of ranged combat. It is on its own, and with my modifications, a far too complex class. So, Archers would have the advantage that they would have powerful archery and less need for armor, but would need to rely on items for mobility, detection, and general utility.

b) High mage. These guys can only use one category, but man, can they use it well. Cheap spells, increased effects, anything you could want. Probably going to only ever be applied to one of the primary 3, but would be hilarious applied to the others.

c) Thief, has no magic, but abnormally high stealth, better melee than rogue, around the same as ranger, and its own version of techniques, with stuff like detect objects, and perhaps some form of assassination.

This leaves us with 2 base damage classes, no frills, two technique classes, one with stealth, one with more all-around function, 2 half-casters, one with stealth, one with melee, and two spellcasting classes. A total of 8 classes, with a fairly simple division, and diverse play options. Half the classes would use spells, half would not.

Considering this requires an entirely new class system, and also a rework of stats, this would have to be one of the later changes to be made.

There's stuff from variants to consider stealing as well. Ironband has an excellent stat system, much of which could be lifted more or less directly. The branches that are probably easiest to steal from are the NPP branch and the O branch. and there are plenty of ideas to be lifted from the more exotic variants such as Z, even though they'd need to be reworked to fit properly.

Short version: Remove mage/priest distinction in name/stats, put spells into groups, allow mages to choose two groups, half-casters to choose one, rework class system for consistency, steal stuff from variants. Big changes.

[Estie]

High mage: damage at the cost of utility feels weird for a mage (and warrior too for that matter, but we are used to it). I think I would prefer an aproach like ToME2 sorceror, who gets damage + utility at the cost of survivability (hp).

The wizard with maxed con and level can survive any 1 attack, but that need not necessarily be fulfilled. It is possible (and exciting) to fight things that can 1-hit you by avoiding LoS, utilizing hockeystick or other (possibly new, possibly restricted to high mage) around-the-corner abilities.
The same style of play could be used for the archer; low defense to the point of having to fight in 1-hit situations, high offense and means to fight out of LoS.

Spell groups: If there are going to be many small ones of those, it might be better to give access to 2-3 instead of 1-2; more possibilities that way.

I still would love physical damage to be the primary threat, to the point where you just cant melee something like a troll without sufficient armor (say, a plate mail).
Red dragon, no fire resistance ? You die. Troll, no plate mail ? You should also die (if you melee).

Anyway, all of this might well be far from the realities of the changes that happen. I would love to hear some thoughts by the people who actually program all this

[Philip]

There really aren't that many groups I have envisioned, actually. You still get a third of all of the spells, and two of the categories are more or less the same except with a thematic difference. To play what used to be mage, you really can just choose categories 1 and 6 and play almost exactly the same as you used to. Or you can do something new and interesting.

I think the discussion of melee damage and combat in general belongs in its own thread, honestly. I would just like to note that this would probably hurt warriors a lot less than mages, and possibly make priests entirely unviable. Priests make do by bashing until they need to heal, and if they need to heal every other turn or so (with moderate AC), they will no longer be able to use melee.

Any proposition which involves being weak enough to die instantly is pointless. No-one will play if they have a certain chance of dying every turn, since that means they will probably die pretty soon. If a Mage could be killed by even a tenth or twentieth of what he saw, he wouldn't have any safe places to hide.

[Estie]

Any proposition which involves being weak enough to die instantly is pointless. No-one will play if they have a certain chance of dying every turn, since that means they will probably die pretty soon. If a Mage could be killed by even a tenth or twentieth of what he saw, he wouldn't have any safe places to hide.

This is very very wrong. Diving is exactly what you describe: you have a chance to die every turn, way more than just 1/10th of monsters can instakill you and there is no safe place to hide. Yet people play that way (and some love it).
The ToME2 sorceror dies from a sneeze of any monster at depths, but his utility spells are so powerfull as to allow him to get by.

The diver suffers high risk high reward scenario for a while, at the end of which he either dies or gets out of 1-hit land. The proposition can thus be rephrased: a high mage is like a diver who can never reach 1-hit safety; he has to kill Morgoth at risk of death if he lets him get LoS.

[Philip]

Yeah, but you don't get into fights when outclassed while diving, you pick off the stuff that won't kill you. If something is awake and can kill you you don't fight, even if you can win, because you can also dies. A diver who doesn't reach 1-hit safety will die during the final grind, or will be woefully unprepared for the final fight. Or just get killed by Morgoth.

[Estie]

Yeah, but you don't get into fights when outclassed while diving, you pick off the stuff that won't kill you. If something is awake and can kill you you don't fight, even if you can win, because you can also dies. A diver who doesn't reach 1-hit safety will die during the final grind, or will be woefully unprepared for the final fight. Or just get killed by Morgoth.

This is simply untrue. The concept exists (tome2 sorceror), its proven, and fun enough that people played it. Wether something like that shall exist in vanilla is an open question, but what you say here isnt a reason not to do it.

[Derakon]

I think Vanilla should be moving away from instadeaths, and therefore do not support redesigns that make them more common. I recognize that instadeaths are a core part of modern Vanilla balance, but Vanilla should not have classes that are never free from the risk of instadeath. You should only be at such a risk if you're doing something seriously unwise like being at dlvl90/clvl30. During normal play, death should occur over the course of at least a few turns, during which time the player has a chance to recognize that things are going horribly wrong.

The problem with Vanilla as it stands of course is that the player's defensive options are so strong that giving them even one turn of survival is tantamount to saying that the player can never die with even minimally intelligent play. So we need to find ways to rectify that.

[Estie]

I think Vanilla should be moving away from instadeaths, and therefore do not support redesigns that make them more common. I recognize that instadeaths are a core part of modern Vanilla balance, but Vanilla should not have classes that are never free from the risk of instadeath. You should only be at such a risk if you're doing something seriously unwise like being at dlvl90/clvl30. During normal play, death should occur over the course of at least a few turns, during which time the player has a chance to recognize that things are going horribly wrong.

The problem with Vanilla as it stands of course is that the player's defensive options are so strong that giving them even one turn of survival is tantamount to saying that the player can never die with even minimally intelligent play. So we need to find ways to rectify that.

Well, how do you want to accomplish that ?

You can limit the options for disengagement once a fight started; for example, removing phase door scrolls might already be enough (next competition idea: vanilla warrior, phase door scrolls removed).

The thing is, then you commit to a fight, realize things are going horribly wrong while you lose hps during a few rounds, and you still die for lack of options. I think this isnt what you have in mind, is it ?

You can limit the players ability to pick their preferred difficulty, like Sil or ToME 4 do ("we only allow ironman mode now"). This is similar to the above case; while in theory you might have escapes, you are reluctant to use them as doing so would close doors.

You can limit the divination options (detect monsters). Have the player bumbling about in the dark, with the option to retreat if a dragon appears and takes away 1/2 his life in 1 breath. This sounds like the best option to me, sofar. But is it better than the current detect + risk 1-shot death ?

My point is that while I agree that ideally it should take more than 1 turn from "everything is fine" to "tombstone", it doesnt matter if you lose hit points continually during that time, or all at once on the last turn of events.

Running into a gmm because you didnt detect and getting KOed is better than starting a fight with a giant, realizing that you are going to lose, but lacking options to retreat.

[Philip]

Tome4 and Sil both in fact do this incredibly well. In Tome4, once you start running out of options, you might be able to disengage safely, you might have to tough out a temporary effect, and how likely you are to survive a nasty situation depends on prior actions, but not entirely. Additionally, it teaches you to recognize that something is not going well before you get confused and silenced. In Sil, it takes a while to die, but if you don't pay attention, you can get into a situation where retreat is either blocked or impractical because of speed issues. How characters deal with these situations is what is interesting.

EDIT: They don't do it by limiting ability to pick difficulty, since the game is designed from scratch, there isn't a difficulty that's blocked off or anything. There are a couple difficulty options in Sil, and some races are easier. And Tome4 does actually have varying difficulty *and* permadeath options, ranging from easy and lots of lives to absurdly hard, one life only.

[Derakon]

Well, how do you want to accomplish that ?

Ahh, that's the tricky bit, isn't it? I'm just an ideas man, I'll leave the implementation details to the rest of you.

More seriously, one simple potential solution would be to make all escapes be strictly limited in number. No casting Phase Door/Teleport/Teleport Other/Destruction/etc; you must use consumables, and they're in limited supply (wands/staves wouldn't be rechargeable, and rods wouldn't exist). Likewise, Phase Door wouldn't be sold in shops, except the Black Market. Every item you use now is an item you won't be able to use later. So you have the resources to escape some fights, but poor play will run your resources out and then you'll die.

[Estie]

@Philip: Vanilla has all that, too, but you can pick difficulty during game by picking depth without penalty. This is what I am missing in Sil and T4, I dislike this aspect of both games, so I guess we can just agree to disagree (technically, T4 has an endless dungeon mode, however it is lacking a winning condition. Why ?)

@Derakon: Global limit on escapes or cooldown like in T4 ? The latter removes "burst" escape, the former doesnt. Otoh, cd system is independant of game length.

[Philip]

Infinite Dungeon is just that, infinite (though you can only do each depth once). That's the point. You win if you die further than anyone has died before you. Or you win once you get to a certain dungeon level you choose ahead of time. Or you win by finding all the special lore. The clock in Sil isn't too brutally restrictive, though yeah, I get what you're saying.

[Derakon]

@Derakon: Global limit on escapes or cooldown like in T4 ? The latter removes "burst" escape, the former doesnt. Otoh, cd system is independant of game length.

Global limit -- you can use what you find. Same as with big healing potions in the current game, basically. If you want to save them all up for the late game, that's your call, but it may mean sticking it through in some dangerous situations early on. Similarly if you use your escapes at the first hint of trouble, then you may not have them when you really need them...

And of course if you want to grind for them, then again you're free to do so. Ideally the allocation rate should be such that this isn't generally necessary. We've managed it pretty well with the consumables people use in the final fights, so I don't think this is impossible to set up.

[debo]

Global limit -- you can use what you find. Same as with big healing potions in the current game, basically. If you want to save them all up for the late game, that's your call, but it may mean sticking it through in some dangerous situations early on. Similarly if you use your escapes at the first hint of trouble, then you may not have them when you really need them...

And of course if you want to grind for them, then again you're free to do so. Ideally the allocation rate should be such that this isn't generally necessary. We've managed it pretty well with the consumables people use in the final fights, so I don't think this is impossible to set up.

I think Halls of Mist has something marginally similar to this (limited escapes refreshed per floor), if anyone wants to try it.