Class penalties/restrictions that make sense

[wobbly]

You could make the preference in the weapon instead. Blessed is already +wisdom. Stack the odds towards blessed blunts. That'd also preference blunt on a druid

[Djabanete]

TL;DR Suggestions 2, 3, 5, 6, 8, 10, and 11 below might be worth thinking about. Also can someone tell me if rogues and rangers feel distinct in current Angband?

I agree with bunnies that you should begin with how you want the different classes to behave. Also, you should keep in mind that in many cases it is more fun for the player if "correct" behavior is encouraged rather than "incorrect" behavior discouraged. Bonuses can shape strategy just as much as penalties, or more: Rangers are superlative archers because they're given an extra shot, and that's fun for people who want to win with ranged combat, but priests aren't necessarily worse at melee because of the pointy penalty. They just have to wait around longer until they find the right weapon, which in my view isn't more fun.

So how do we want the different classes to behave?

Warriors are the simplest, especially because vanilla Angband seems to specifically wish them not to have special abilities (combat-themed spells that could be accessed with the "m" command). Warriors are great at melee and ranged weapons. They should play in a straightforward style that prioritizes killing monsters through combat. They shouldn't be locked into a berserker style --- it's okay to be tactical and use ranged combat or terrain if the situation calls for it --- but they need a lot of hit points because they'll engage in a lot of hand-to-hand fighting. The archetype for the warrior would be someone like Boromir or, better yet, Miyamoto Musashi, who advocates proficiency in all weapons and warns that stylistic preferences are a weakness. As a special bonus, the warrior's maximum number of melee blows is increased by one. I would argue that warriors are perfectly designed already.

Warrior:
Melee: 5/5* (gains extra blow; see also Suggestion 2)
Ranged: 5/5
Hit Points: 5/5
Magic: 0/5
Stealth: 1/5

Paladins are a lot like warriors, but they're based on a much narrower archetype: that of the chivalrous Medieval knight. The real-world archetype would be crusaders and Knights Templar, and the fictional archetype would be a highly perfected ideal version of this (probably rare in real life), such as Ivanhoe. In addition, the in-game paladin wields magical powers. So how does the paladin differ from the warrior? First, historical knights wouldn't necessarily be bad at ranged combat, but their main advantage would be the ability equip themselves with horses, swords, and metal armor, whereas it's relatively cheap to equip any footman with a bow. So the knight's modus operandi would be to enter melee combat at high speed with the best weaponry and armor available. Second, idealized chivalrous knights would put a lot of emphasis on dueling, jousting, and throwing down the gauntlet for a worthy cause (in Ivanhoe, the title character and the Templar antagonist are always challenging each other to single combat). So it makes sense to distinguish the warrior as an indiscriminate combatant and the paladin as a duelist. Of course, in Angband, paladins are also expected to support their melee combat with buff spells and healing. Suggestion 1 (minor): Add a paladin-only Charge spell that moves a short distance and attacks in melee. Suggestion 2: My understanding is that, in recent versions of Angband, if you kill an enemy in n < m blows, where m is your total number of blows, you save energy (time). So if you have four blows and it takes two to kill an orc, you can kill two orcs in one move and two keypresses. I recommend making this apply only to warriors so that they become the crowd-fighting experts. Other classes can fight the way they used to in older versions.

Paladin:
Melee: 5/5
Ranged: 3/5
Hit Points: 4/5
Holy Magic: 3/5
Stealth: 1/5

Priests are based on a variety of archetypes, including more warlike roles such as Friar Tuck (in Ivanhoe) as well as more peaceful ones such as ... is it OK to say Elrond? He's a master healer who has fought in battles but doesn't consider weapons his greatest strength. Of course, he doesn't explicitly follow a religion, so he's not the best example. In any case, Priests shouldn't be precluded from melee combat and they should also, of course, have the very best healing. In my view, the pointy penalty doesn't make for good gameplay (as mentioned above), nor does it make sense from a flavor perspective. I think the AD&D 2nd Edition Player's Handbook justifies it by saying that "most religions take a dim view of shedding blood," but that's silly of course; a Mace of Disruption would leave most monsters a gory mess. Perhaps the AD&D people were thinking more along the lines of quarterstaves instead of Maces of Disruption; in real life, you can defend yourself with a staff in a nonlethal manner. In Angband, though, the point is to kill loads of monsters. Suggestion 3: Scrap the pointy penalty.
If priests have passable melee combat and great healing, how do they differ from paladins? Clearly, we expect them to be more reliant on magic. They need to have worse combat skills and worse hit points to compensate for their superior magic. This leaves their ranged combat pretty crummy, which is sad because Friar Tuck was a good archer in Ivanhoe; but maybe they could have some strategic flexibility in the form of a combined See Invisible + ranged accuracy spell that could be called Blessed Sight. Suggestion 4 (minor): Upgrade the priest spell See Invisible (that's a spell, right?) to Blessed Sight.
It still feels like priests need some tactical differences from paladins, if only because bad melee plus great healing is a painfully slow way to go through the game. Of course, priests will lean on Orb of Draining, but it feels like they need another angle. I would suggest exploring Glyphs of Warding as a tactical consideration. For example, if priests received a melee and ranged accuracy bonus for each adjacent glyph, that would create tension between fighting in an anti-summoning corridor and fighting in the open (surrounded by glyphs). Second, priests could get spells that use glyphs in some new way, for example, Incandescent Glyphs: Create a ball of (lots of) fire and light damage around every glyph in your line of sight, then each of those glyphs has a 50% chance to disappear. Suggestion 5: Devise ways to use Glyphs of Warding as a new tactical angle for priests.
Do priests actually need weapon and armor limitations if their melee and ranged skills are correctly balanced?

Priest:
Melee: 3/5
Ranged: 2/5
Hit Points: 3/5
Holy Magic: 5/5* (0% fail rate healing; in this rating system, I'm treating fully powered casting as better than any other 5/5)
Stealth: 1/5

Mages are expected to lean heavily on their magic. They should be resourceful and well informed (through detection spells). Opinions differ on whether they should be expected to deal damage primarily through spells or instead their spells should be support for standard combat and archery. Personally, I think they become more distinct as a class from rangers and rogues if they concentrate on killing monsters using damaging spells. Suggestion 6: Push mages in the direction of relying on damage spells to kill monsters, leaving the hybrid-class behavior to rangers and rogues. One or two truly powerful buff spells that improve strength, damage, and/or weapon accuracy could leave room for mid-game or late-game mages to use weapons for damage, if they find the right weapon for that strategy.

Mage:
Melee: 2/5
Ranged: 2/5
Hit Points: 2/5
Arcane Magic: 5/5* (in this rating system, I'm treating fully powered casting as better than any other 5/5)
Stealth: 1/5

Rangers are based on the Aragorn archetype, but with two differences: an emphasis on archery and an ability to cast mage spells. Rangers are decent at a lot of different things. Angband lacks a pure archer, so it makes sense for rangers to excel at archery (and it's now traditional), but other than that they should have a balanced style of play, using ranged combat, melee, and support spells; also, they're the only class besides the rogue that, flavor-wise, should have appreciable innate stealth. Suggestion 7 (very vague): A ranger's mana pool should be quite small, since thematically they can't be too lousy at melee and their archery is better than warriors'. Thematically they are also hardy and should have lots of hit points. Big-damage spells should not be in the ranger's repertoire. For balance and thematically, it would make sense if rangers' magic were limited to casting a couple of healing spells in a pinch, or a couple of detection spells or buff spells before engaging; no more.

Ranger:
Melee: 3--4/5
Ranged: 5/5* (gains extra shot)
Hit Points: 4/5
Arcane Magic: 2/5
Stealth: 3/5

Rogues are hard to distinguish from rangers. They, too, are generalists who want to cast arcane spells as well as fight both in melee and at a distance. Does anyone find that in practice they play similarly? I would suggest pushing rogues as far as possible in the direction of ninjas in Hengband, who rely on superior positioning as well as stealth. Suggestion 8: Give rogues as many abilities as possible that involve controlling battlefield positioning, either as innate abilities or as rogue-only spells in mage books. An example of an innate ability would be flanking strike: Moving from one position adjacent to a monster to another position adjacent to the same monster would cause an automatic melee attack against the monster. Examples of rogue-only spells (which could go in Mordenkainen's Escapes or in early books) would be Exchange Places (swap places with a monster, possibly with an attack as well) or Hit and Away (attack a monster and immediately teleport a short distance). Suggestion 9: Give rogues a highly increased chance of scoring critical hits with daggers, short swords, main gauches, rapiers, etc.

Rogue:
Melee: 3/5
Ranged: 4/5
Hit Points: 3/5
Arcane Magic: 2/5* (Detect Objects and some other rogue-only spells)
Stealth: 5/5

Thoughts on stealth and movement spells: It makes sense for there to be two classes that intrinsically gain appreciable stealth: ranger and rogue. The other classes might as well be equally bad at it. Suggestion 10: Wearing a medium weight of armor should trigger the message: "Wearing this much weight of armor reduces your class's innate stealth and hinders your movement spells" (Ranger: 3/5 -> 2/5. Rogue: 5/5 -> 3/5.) Wearing a heavy weight of body armor should trigger the message: "Wearing this much weight of armor negates your class's innate stealth and greatly hinders your movement spells." (Ranger: 3/5 -> 1/5. Rogue: 5/5 -> 1/5). (This goes for all classes with innate stealth and/or spells; adjust the message appropriately.) Movement spells would include the aforementioned Exchange Places and Hit and Away as well as staples such as Haste Self and Teleport Self. The penalty on movement spells may be small enough that fully powered casters (mage and priest), but not partial casters, could still bring fail rates to 0% for the more basic movement spells. This system gives non-warrior classes reasons to think about armor weight without necessarily ruling out heavy armor altogether. (A mage with maxed intelligence might be able to teleport at 0% fail rate with a medium weight of armor, for example; or a rogue who has found the One Ring might not care about stealth anymore.)

Thoughts on class differences: Rather than limiting equipment choice too heavy-handedly, I think that making smart differences in spell repertoire might be the way to go. Suggestion 11: Tailor the strategic options of the fully powered and partially powered casters by editing fail rates and mana costs of spells, as well as what spells are available, in a nonuniform way. For example, rogues don't need to be slightly less good at all mage spells than mages are; they could have a majority of spells entirely off-limits, be somewhat less good than mages at Haste Self and Teleport Self, and be extremely good at certain rogue-only spells that mages don't have access to. Rogues could be better at Haste Self than rangers, who could be better at Resistance than rogues.

[Sky]

Jeez, we started talking about a miserable -2 hit penalty, and now we want to nerf everything. Why not just build a new game from the ground up?

[Antoine]

Djabanete's ideas are pretty good

A.

[Frenchminstaer]

Djabanete's ideas are pretty good

A.

I agree, especially the suggestions for paladin and warrior.

[geoff_tewierik]

While reading Djabanete's bit about Rogues, that they're ninja like and should be able to move around in combat, I wondered whether a targetted Phase Self would work? I.e. move the rogue in a chess knights move or equivalent, where the user could target where they wish to land. Also the Exchange Places could work with any monster in LoS, rather than the one adjacent?

[Derakon]

The talk about Priests and glyphs gave me a thought, which is that priests could maybe sanctify tiles with different blessings, granting buffs to the player so long as they stood on the tile. They wouldn't necessarily have to be Glyph of Warding, since that's a high-level and (rightfully) expensive spell, but e.g. instead of a Bless buff, you could get a Glyph of Courage that grants Bless + Heroism so long as you stand on it. Some other possible glyphs:

* Glyph of Righteousness: your melee attacks are branded with Slay Evil / Undead / Demon
* Glyph of Zeal: spell failure rates drop to 0%, SP costs increase based on how much failure rate decreased
* Glyph of Recovery: doubles HP/SP regeneration, stacks with Regeneration
* Glyph of Succor: healing spells take no time (or maybe 25% normal time) to cast so long as they succeed
* Glyph of Wrath: adjacent monsters take damage each turn (placement can be targeted)
* Glyph of Protection: incoming damage is cut by 20% after resistances/AC/etc. are applied

This would work best if you could only ever have 1 glyph (of any type) present on the level at a time. Cast a new one and the old one goes away. And of course, monsters can attempt to destroy these glyphs just as they would the Glyph of Warding, plus they can move on them normally.

This feels like a class gimmick that has some legs and would stand to strongly differentiate priests from the other classes. They'd have a wide variety of different buffs they could employ, but only one at a time, and only so long as they're willing to stay put. But they also have the resources to be able to stay put for a long time.

[Philip]

Derakon, I would agree that only Glyph of Warding should prevent monster passage entirely. The others should give monsters debuffs when stood on. I would make either Courage or Righteousness try to fear a monster (giving Priests a potential early way to avoid combat with, say, tigers), Zeal could reduce monster casting frequency, Succor could block monsters from using healing spells, Protection could try to put a monster in stasis. I would also make a player standing on Wrath hit all adjacent monsters, with (normal blows+adjacent monsters-1) blows split between the adjacent monsters. Dunno about the other glyphs. I would also make glyphs a player is standing on indestructible, but make Warding give, say, 50 AC, Protection from Evil, and Resistance as effects when stood on, and block monsters entirely (with maybe 5 turns minimum durability) when left alone.

[Ingwe Ingweron]

While reading Djabanete's bit about Rogues, that they're ninja like ....

I have to disagree, although admittedly that is what they've developed a bit like in the past. The feature branch is starting to, in my opinion, correct that trajectory. If Angband is supposed to be Tolkien at its roots, I am hard pressed to think of the burglar Bilbo Baggins as a "ninja".

[Huqhox]

Agree entirely that the pure casters should be restricted to 4 melee attacks. And pure melee classes should get 6. The whole priests not using edged weapons is yet another hangover from D&D (where it was pretty arbitrary too) and I'd be happy to see that removed.

Not sure about bless giving +1 attacks. This seems a little overpowered. Some sort of extra damage vs evil would perhaps be both useful and thematic

[fph]

I like the glyph idea. Another thematic name for them would be "altar".

We should make sure that they are expensive enough to cast that priests don't end up spamming them too often. I still remember the ridiculously unbalanced "sea of glyphs" of Tome 2.3 rune magic...

[Derakon]

The whole priests not using edged weapons is yet another hangover from D&D

I like this phrase. I'm sure D&D has been the cause of many a hangover.

I like the glyph idea. Another thematic name for them would be "altar".

We should make sure that they are expensive enough to cast that priests don't end up spamming them too often. I still remember the ridiculously unbalanced "sea of glyphs" of Tome 2.3 rune magic...

That's why I suggested only being able to have one glyph active at a time. Glyph of Warding is only not spammable because it's expensive and a late-game spell; you can't really have an "expensive" spell that's still usable in the early game, without it becoming cheap by endgame*. So some limitation other than cost would need to be used to make early-game glyphs not spammable.

* I guess you could make the spell cost a percentage of max SP. But what's the point of increasing max SP then?

[fizzix]

It's not clear to me why we need to replace the pointy penalty with something else. Just remove it and figure out what the priest class wants to be interesting, and work towards it. You can probably do a lot with just manipulating the spells.

[Nick]

This is brilliant, thanks everybody. I'll give a brief summary of my current thinking; I'm currently intending to do individual threads for each new (or altered) class and how I think they should be further changed over the next while.

Class penalties, and class buffs, should be used in line with how classes are designed to be played, and they should play a role in encouraging the player to adopt this playstyle. To this effect, we should start with the class design first and foremost, before jumping into thematic class restrictions.

That's a nice summary, and almost exactly where I'm at. I tend to think of it in terms of what they represent rather than how they are to be played, but in practice it comes down to more or less the same thing.

So I think the pointy penalty and the glove penalty can just go. They're more annoyances than a meaningful balance mechanisms, and they don't really contribute to the ethos of any of the classes they affect.

On the other hand, I think the light penalty for necromancers (which has the same effect as the pointy penalty if the player is on a lit square) does work, emphasising the necro's affinity for darkness and fear/hatred of light.

A couple of other comments:
Djbanete - thanks for all those ideas. Note that this is change from the class/magic branch, so rogues and rangers are already looking more distinct.

Derakon - re the glyphs, FAangband already has this Mages in FA have a final book which is runes, which can be placed for various player effects; your suggestions are kind of a combination of this idea with monster traps. I think it's an idea worth keeping in mind and developing (note that new rangers now get a decoy spell, which is kind of a glyph), but in hindsight the runes weren't really a big win in FA.

[Derakon]

Derakon - re the glyphs, FAangband already has this Mages in FA have a final book which is runes, which can be placed for various player effects; your suggestions are kind of a combination of this idea with monster traps. I think it's an idea worth keeping in mind and developing (note that new rangers now get a decoy spell, which is kind of a glyph), but in hindsight the runes weren't really a big win in FA.

Naturally. By "final book" it sounds like you mean this was the deepest/rarest book, though? Which would make it less of a defining class feature since you wouldn't have access to its spells for much of the game. I mean, most people probably wouldn't say that being able to cast Banish is a defining class feature for the current vanilla mage, even though it is a very handy and powerful ability once you get it.

I guess I envisioned a glyph-based priest as getting new glyphs periodically throughout the game, having 1-2 glyphs per spellbook that would gradually get more powerful / less-niche effects and enable the priest to take on more dangerous enemies. It's more or less a more flavorful way to have a class that's bad at almost everything until they cast a spell that makes them great at one thing.

You could also go the Sil route (or the ToME 2 bard route, etc.) and have prayers be continually-cast spells that gradually deplete the player's SP in exchange for effects.

At any rate, I'm glad these discussions are stimulating.