Why is Angband fun?

[Tibarius]

total nonsense

Philip and Nick => I don't want to be insulting, but i consider practically all you wrote as total nonsense.

No surprise, you are ruining the game.

The way the game should be, is derived from the goals of the players.

For a give set of goals there is a perfect condition available that the game can reach (and then shouldn't be touched anymore, since any more modifications would reduce the perfect goal matching).

[Pete Mack]

Tibarius--I have no idea what you are talking about. There is no 'perfect condition', because players learn to exploit weaknesses in any game. Yet if you keep 'patching' those holes, you end up with a game that is grindish, and unfriendly to newbies. When is the last time you played Angband? What variant did you play--and are you still playing 3.0.x?

[t4nk]

Philip and Nick => I don't want to be insulting, but i consider practically all you wrote as total nonsense.

No surprise, you are ruining the game.

The way the game should be, is derived from the goals of the players.

For a give set of goals there is a perfect condition available that the game can reach (and then shouldn't be touched anymore, since any more modifications would reduce the perfect goal matching).

Wow, your rude rant really angered me.

First of all, you've already opened one shitty thread, so why are you shitting up this one now?

Second of all, the way the game should be is derived from the goals of the developers, who may or may not care about "the players"; and even if they care about some players, they don't necessarily care about others; especially upgrateful, stupid and rude people like you.

[wobbly]

The only perfect state is death and even the dead rot.
A game that changes creates new challenges, you can play it without playing the "same old game". It's harder to just roll out perfect winning plan "A". Your perfect game has a perfect solution you see? It achieves perfect boredom.

[Mondkalb]

Philip and Nick => I don't want to be insulting, but i consider practically all you wrote as total nonsense.

No surprise, you are ruining the game.

The way the game should be, is derived from the goals of the players.

For a give set of goals there is a perfect condition available that the game can reach (and then shouldn't be touched anymore, since any more modifications would reduce the perfect goal matching).

That was just rude.

[Philip]

Objection, yer honour. While I agree that player deaths are important for the metanarrative, I don't think "unfair" events should be.

A fair point. I think I may have confused my personal formative experiences with a general truth. I do think there needs to be some antagonism from the part of the game itself towards the player, but "unfair" probably isn't the only flavor of that.

You're taking a narrative-based approach to the game, and that's totally valid, but it's not how everyone plays. I take a much more gameist approach; I like the game because of the nature of the decisions it asks me to make and the way it rewards those decisions.

I think the metanarrative approach is very close to the gameist approach, to the point where I'm not entirely sure I can distinguish between them. The irrelevance of the nominal narrative and the creation of tense situations through game mechanics and the joy of finding powerful items are all core components of both approaches. The main point of difference, I suppose, would be the question of the extent to which prior games influence the emotional arc of later games. If it were strictly through the knowledge accumulated in those games, then there is a perceptible difference. On the other hand, if you hate Kavlax for all the times he killed you, rather than fearing him because of it, that's metanarrative.

It raises a great many questions which IMHO go well beyond Angband to other games, computer based and not - I found it particularly interesting to think about sport in this framework (Is cricket a game or a sport? Discuss.).

I hadn't considered sports as metanarrative. It does in fact raise a lot of questions in that context, yeah. I'm not sure if narrative and metanarrative are even distinguishable in team sports, because of the way fans, teams, and analysts talk about them.

[Pete Mack]

I am still wondering what an 'unfair' action is in angband. The only thing that comes up is insteadeath to an unseen drolem before full.detection is available. That is unfair in one way--at the time it was unavoidable. But it is fair in another: if you have light rods, and/or you are moving fast through the levels wherr drolems are common (or if you have found rPoison and more than 237 HP), such deaths are vanishigly rare. So long as such deaths are rare, unfairness isn't much of an issue. If you are losing every character to traps that you have no way to detect or disarm, that would be a real problem.

[Voovus]

I am still wondering what an 'unfair' action is in angband.

Probably worth considering what's "unfair" on a new player, rather than on an expert.

[PowerWyrm]

https://lparchive.org/Angband/

[Tibarius]

angband development

Tibarius--I have no idea what you are talking about. There is no 'perfect condition', because players learn to exploit weaknesses in any game. Yet if you keep 'patching' those holes, you end up with a game that is grindish, and unfriendly to newbies. When is the last time you played Angband? What variant did you play--and are you still playing 3.0.x?

I played almost all versions from 3.5 and later. Currently 4.1.3.

What i mean is, if i would be the only player around i would be able to define exact requirements how the game mechanics should work and the game should play.

Once that requirements would be coded the game would be in a perfect condition based upon the defined requirements.

Of course i am not the only player and obviously the requirements of the players are very different. Because i cannot loose any good word about the coming game mechanic changes Nick is writing about.

If Nick is re-coding Angband so that it is playable on future operating systems, like win10 or other unix/linux systems. That is great.

But the announced changes with new classes etc are ruining the game. It is no longer a moria follower but a complete new game, which i name Nickband.

So overall i do not welcome his work on the game mechanics, regardless of the fact that he does spend his free time for it. Actually i would encourage him to look for a new maintainer, who actually does what the name implies
=> to maintain the game.

If you want to develop a new game, develop a new variant.

That is my view on the current game development.

[Pete Mack]

Um. Angband has worked on many platforms since forever--Win10 of course is supported by the Win7 64bit update. It also works on OSX. I mostly play on an android phone these days. There was NO difference in gameplay between 3.5 and 4.0--the latter was entirely an engineering release. So what you are saying is you don't like 4.2, and have never played 3.0.x (which the really diehards consider the last 'really good' version. I have some limited sympathy for this view.) But 4.2 is more like 3.2 than anything else, and 3.2 was an intentionally broken release: it was unbalanced in ways that the developer.could not know without getting feedback from many players. 3.5 was a significant improvement.

So, rather than resorting to ad hominems: what don't you like about 4.2, beyond balance issues, which are expected in this release? By the way, the motivating reasons for 4.2 were two-fold:
1. Make the base classes more distinct from each other
2. Reduce necessity of teleport other. (I suspect this one is a pipe dream in anything that is recognizable as Angband, in the sense that Sil is not so recognizable, except as a roguelike that uses a derivative code base .)

[Tibarius]

base classes

Um. Angband has worked on many platforms since forever--Win10 of course is supported by the Win7 64bit update. It also works on OSX. I mostly play on an android phone these days. There was NO difference in gameplay between 3.5 and 4.0--the latter was entirely an engineering release. So what you are saying is you don't like 4.2, and have never played 3.0.x (which the really diehards consider the last 'really good' version. I have some limited sympathy for this view.) But 4.2 is more like 3.2 than anything else, and 3.2 was an intentionally broken release: it was unbalanced in ways that the developer.could not know without getting feedback from many players. 3.5 was a significant improvement.

So, rather than resorting to ad hominems: what don't you like about 4.2, beyond balance issues, which are expected in this release? By the way, the motivating reasons for 4.2 were two-fold:
1. Make the base classes more distinct from each other
2. Reduce necessity of teleport other. (I suspect this one is a pipe dream in anything that is recognizable as Angband, in the sense that Sil is not so recognizable, except as a roguelike that uses a derivative code base .)

The base classes were already distinct enough.
Warrior plays differently from Rogue from Mage from Ranger from Priest.
=> No need for new classes

The introduction of special rooms should be removed.
=> special rooms repeat themselves over time, i cannot see how they improve fun for players

=> the uniques are a mix of several different themes / topics
I personally would remove fixed uniques and replace them by random boss monsters with random powers, if that is not liked reduce at least to tolkien themed uniques

=> i would make random artefacts the standard game behavior, yet it looks like random artefacts are sometimes too strong in my eyes compared the the fixed artefacts

=> i think finding spell books add too much spells at once, i imagine it would be more fun if you could find spells one by one at times (on scrolls) which then can be written into a spellbook
(i admit tho, that this is also beyond maintaining and more like a major change)

But basically i already stated this things in the one or other post earlier. So it is very likely that this repeated posting does not add any value for anyone.

[Philip]

Well, since this post refrained from any personal attacks on me or anyone else, I will respond to it.

Classes were not even remotely distinct. You could do magic (the priest and mage spell lists both had everything), you could do bump attack, or you could do shooting. The distinction between a warrior (high hp, good to_hit, great fighting stats, but utility type stuff is unreliable) and a paladin (high hp, good to_hit, good fighting stats, but utility type stuff is unreliable) was minimal, especially after stat-gain.

Special rooms repeat themselves over time, indeed. You know what also repeats itself over time? Empty rectangular 8x12 rooms, plus or minus 4 on the dimensions. They improve fun by adding variety. The solution to special rooms repeating themselves over time is to make more special rooms (unless you have an idea for how to procedurally generate interesting dungeon architecture).

Random boss monsters with random powers is a terrible idea, since it just teaches players never to go near a random boss out of fear it has gravity breath as its only spell and will spam it forever. I don't know if you've noticed, but the tolkienization is actually in progress - Angels have been replaced with Ainur.

Feel free to suggest a better way to generate random artifacts.

As for finding spells one at a time, that is probably the smallest change of all the ones you suggest, as it happens. I rather like powerspikes, however. Some of the most fun I've ever had was either trivially handling stuff I had massive problems with just a couple thousand turns ago, or trying to deal with the depths without any of my dungeon books.

You seem to be convinced that "randomly generate it" is both a fun way to do literally everything, and also a trivial matter to design and code. This is despite having been able to observe maintainers try to make random artifacts work properly since 3.1 days at least, with a level of success you don't even seem to be entirely happy with.

Go play NPPMoria, I think it's on angband.live now.

[Pete Mack]

Warrior and Paladin have two differences, but they don't add up to much separation. The first kicks in early: res fire and cold. This makes killing Smaug and friends easy for one class but hard for the other. But not *that* big a difference, and warrior can carry other source of double resist if he wants to bother.
The other kicks in almost too late to matter: Heal300 spell turns paladin into unstoppable tank. At that point, the game is almost over (around CL 37, and paladin is just walking over everything short of an umdead pit.)

Ranger vs. Rogue is another. For most of the game, ranger is just a weak rogue. After cl 40--again too late to matter, Ranger becomes unstoppable killing machine with two glorious defensive spells, Glyph and Destruction. It isn't a class with good replayability.

[Derakon]

Special rooms repeat themselves over time, indeed. You know what also repeats itself over time? Empty rectangular 8x12 rooms, plus or minus 4 on the dimensions. They improve fun by adding variety. The solution to special rooms repeating themselves over time is to make more special rooms (unless you have an idea for how to procedurally generate interesting dungeon architecture).

Thank you for touching on this, because it's something I wanted to talk about actually and now I have an excuse!

Every procedurally-generated game has hand-built content, because the rules of generation count as hand-built content (this is a strong stance to take, but please bear with me). Whoever wrote the generator for Rogue (a.k.a. the first roguelike) had to decide on rules for where to place rooms, how to connect them with corridors, etc., and the net result is that you pretty quickly learn to recognize what a "Rogue level" looks like. That's despite the fact that Rogue has no hard-coded "put this here then put this right next to it" content of the type that Tibarius is so strenuously objecting to.

Similarly, strip away all vaults, special rooms, and pits from Angband and you're left with the basic underlying dungeon generator. It's a little more advanced than the Rogue generator is, since it has things like corridor intersections, labyrinth and cavern levels, rooms that are composed of two rectangles overlapping each other, etc. but you still pretty quickly learn what an "Angband level" can look like and what kinds of challenges to expect from it.

What matters is not whether you've seen the exact dungeon configuration more than once; what matters is the variability in patterns of play. If you have a very simple generator that nonetheless consistently creates interesting gameplay situations, then great! You don't need to iterate on it. There are roguelikes that achieve this, but they tend to be small, puzzle-style roguelikes like Hoplite or Imbroglio.

But if you have a simple generator that's part of a larger, more complex roguelike like Angband, you may find that the generator's rules aren't able to continue to produce engaging content across many plays. Indeed the only roguelike I know of that comes anywhere close to achieving that goal is Unexplored, and it's a very recent game where the devs consciously decided to put the dungeon generator front-and-center. And it wouldn't surprise me if it still has a lot of what some people would call hand-generated content.

In contrast, Angband's basic generator (stripped of special rooms) doesn't create a high variability in patterns of play. In particular, it typically fails to create:

* Interesting mixes of monsters (especially when high density is called for)
* Terrain that creates complex tactical decisions
* Incentives for the player to not simply abandon the level

The many people that have worked on map gen over the course of the last ~30 years have decided that, rather than try to come up with generalized rules that will produce the above situations, it's far easier, and more effective, to create a wide variety of small units of hand-generated content to provide them, and then modify the generator to let it procedurally populate that hand-generated content so that even if some portion of the dungeon layout is something the player has seen before, the exact pattern of play is still novel.

Consider a vault in isolation. You can see the "same" vault many times in many different games; with experience you'll learn where the walls are, and where the traps, items, and monsters probably are. But those traps, items, and monsters will change from one play to the next, because the generator populates the vault differently. Moreover, the vault doesn't actually exist in isolation: it is affected by a huge number of factors that completely change how you engage with it:

* What dungeon level is it on? A vault with giant scorpions is very different from a vault with Istari.
* What else is nearby on the level? A vault in the middle of the map requires different engagement rules than one that's off in a corner; there might also be a nasty unique or a graveyard nearby that affects how you approach or avoid the vault.
* And most importantly, what kind of character do you have? What class are you playing, what items and equipment do you have, what does your character want and what are they afraid of? A vault that looks like easy pickings to one character might be Stay Away Or Die to another on the basis of a single difference of their equipment, even if everything else were identical.

By way of comparison, look at Spelunky. It is composed of nothing but hand-crafted content -- each level is broken into a grid, and each cell in the grid is filled in with a chosen piece of handcrafted terrain. But because the arrangements are dynamic, and in particular because the monsters, traps, and items in the map are dynamically placed, the game still feels fresh after a huge number of playthroughs. People don't look at level 1 and go "man, I've seen this particular arrangement of walls and platforms before", because this time there's a bat and a snake instead of a spider and that changes everything.

In summary, trying to strip hand-generated content from a dungeon generator is, depending on your definitions, either misguided or actually flat-out impossible. What you should actually seek to do is maximize variability across playthroughs, so that the content you do have does not get stale. You can do that by coding an incredibly complex dungeon generator, assuming you have a year or so of free time, strong knowledge of graph theory, and the willingness to build the rest of your game around the quirks of the resulting generator...or you can spend a few weeks making a bunch of handcrafted units and figuring out how to populate them in a dynamic fashion so they continue to feel fresh after the 20th iteration.