What is good in roguelike gameplay?

[Bear]

Re: message and item descriptions:  That's more about immersiveness than depth, but I'll agree that they're important to the quality of the game. 

Interface considerations like modal door-opening and door-closing are an interface problem that  trades off muscle memory vs. situational awareness and doesn't affect depth-of-play as such.  Players whose primary means of mastering an interface is situational awareness and observation hate non-modal door-opening and door-closing passionately because, "This is STUPID! Why does this dumb game ask this question when there's only one possible answer?"  Players whose primary means of mastering an interface is muscle memory will hate the modal version just as passionately  because "AAARGH!  My hands can't learn to do it without taking into account whether there are other doors, or chests, or whatever crap that my attention isn't focused on!"  Either problem disrupts gameplay and flowstate for players who are strongly dependent on one of these methods rather than the other.  So, yeah, it does need to be an interface setting if you don't want to exclude either group.  But then there's a fight about which way should be the default and whether the players whose style doesn't match the default will stick around long enough to learn that they can change the setting. 

Re: 8-directional targeting vs. omnitargeting, I've thought about this one.  You're right that 8-directional targeting imposes constraints that can be developed into depth-of-play whereas omnitargeting does not, but I still object to it on two grounds:  First, 8-directional targeting without omnitargeting is hard on my suspension-of-disbelief, and thereby breaks immersion.  Second, I think of targeting more as 'interface' than 'game' and as a general principle I prefer it when only the 'game' introduces problems and the 'interface' only facilitates solving them.  When overcoming an interface problem becomes an important subgame, I think we're getting too far away from the spirit of the game.  For example, there's an implicit subgame (introduced by game elements such as monster movement)  of lining up trick shots, which plays monster movement rules against rules about ranges and effect-traversal paths, and can make a huge difference in play. Brilliantly effective use of trick shots can lead to quick victories in difficult situations.  And monsters and effects with different movement capabilities and traversals can make this subgame very interesting, without making the simple direct shot across the room impossible due to interface constraints.

I agree about identification with lots of sources of partial information contributing depth.  If you're going to run an interesting ID subgame, I think the characters should have access to things like "detect magic", and it's interesting to have "experiments" that divide things into classes like engraving does in Nethack or Item Feelings do in Angband or detect-curse does elsewhere, and so on.  But it's actually very hard to do an interesting ID subgame that will stay interesting in the presence of spoilers.  Once a player has memorized all the engrave messages and prices, etc, the ID game in Nethack is just a walk, or, more pejoratively, a grind.  If you have any ideas for an ID subgame that isn't blatantly unfair in favor of spoiled players, I'd like to hear 'em. 

[Ancient]

Re: Interface. Thank you for explaining this to me. I had thought those who scream to remove this stupid keystroke just do not want to be efficient and have smooth experience. How shallow.

You could try modifying NetHack identification to become more spoiler-proof. Lets take wands again. Items have KNOW_B/U/C and KNOW_PVAL (charges here) flags. Lets add KNOW_VALUE which would be set when seeing item in a shop. Then appraisal scrolls would reveal gold value. Also, have set of uniform descriptors for wands:

\oDeath: {lethal} {targeted} {stops bugs}
\oSleep: {debilitating} {targeted} {stops bugs}
\oConfusion: {debilitating} {targeted}
\oFire: {lethal} {targeted} {elemental}
!oBlindness: {unhealthy}
!oConfusion: {unhealthy}
!oHallucination: {unhealthy}
!oAcid: {unhealthy} {turns-to-fruit-juice when treated with u-horn}
...

Upon gaining knowledge of a descriptor check whether known set of descriptors and secondary qualities are unique to this item. If yes, auto-identify this item. Dipping an unicorn horn into a potion of blindness, confusion or hallucination would inform player that this potion has {harmful} descriptor or identify outright if acid.
"You infer that lethal wand stopping bugs must be a wand of death. (+250 xp)"
"You infer that wand of stopping bugs that costs 200 gold pieces must be a wand of sleep. (+50 xp)"
"You deduce by elimination that this wand costing 300 gold pieces must be a wand of fire. (+125 xp)"

Identification almost invulnerable to spoilers is technological gadget figuring out in Alphaman. But it is also boring. You just hit 'f'igure out until item is known or damaged beyond repair by your actions. Unless you have (or hope to) some means of raising intellect.

Berries of Alphaman still were interesting despite being totally spoiled. Ripe berry could be thrown to cause good damage if harmful but if it is stat modifying berry you just wasted a large boost or small permanent increase. Unripe berry could be eaten to identify with weak harmful effects or thrown on monsters for a little damage. There is also possibility of waiting for them to ripen.

[Skeletor]

I have to say you started a very interesting thread, Bear. Thank you!

My personal point of view:

   Items: complex = better. This adds tactics and roguelikeness.

   Enemies: I like it when different kind of enemies show different kinds of behaviors; anyway, there shouldn't be too many intelligent ones IMO. Managing to kill hordes of stupid predictable enemies using the player intelligence is still very satisfying!

   Interface: I think a good compromise between simplicity and flexibility can be done, but if I had to choose one then I'd go for the complex one.. implying the game experience it's worth it. But I think that item/world complexity don't necessarly need a difficult interface.

   Equipment upgrade path: inventory/equipment management is a very tactic and fun aspect in roguelikes, so it shouldn't be undervalued. Lots of stuff found, little inventory/equipment space, good stuff not so early but neither too later: random quarterstaffs of devastation in SMC1 (ADOM) never were a problem.
   
   Power Curve: I'd really like a roguelike where low-level characters have a chance to win the game, I just love when the player manages to surpass very strong / out of depht enemies using his intelligence and tactic. It's not only realistic, but also very rewarding for the player to know that with his knowledge of the game aspects and his tactical and strategical abilities he now manages to proceed through the game even skipping exp/gold rewards: this is ANTI-FARMING, one of the best game experience one human being can have IMHO. But I fully understand that this can be very difficult to obtain in a regular turn based roguelike (without losing the satisfaction factor for levelling and stuff).. and this is also why lately I'm very interested in action roguelikes (and projecting one).

[Ex]

This is an awesome thread and I'm not sure why I haven't commented on it sooner. Anyway, here are the things I like:

- Exploration: It's fun to go around in a dungeon that you know nothing about.
- Loot: I always love getting awesome stuff, and I like the lottery-like unidentified item system. It's great to randomly find a huge pile of gold in a roguelike, even if you can't use it for anything.
- Interesting Classes: I like classes that allow interesting and unique styles of game play. Things like the archeologist in nethack, the knight of the order in CVRL, the half titan in ZAngband, etc.
- Interesting advancement: I like getting new abilities that change gameplay when I level. I love the way DoomRL does it's advancement, and anything that has a fallout 2-like perks system.
- Killing: I love killing monsters of all shapes, sizes and kinds. And I love encountering new monsters.
- Scrolls, potions, and rings: I absolutely love those items with effects that you don't know about before you try them. Unless you're lucky enough to get an identification scroll.
- Upgrading equipment: I always enjoy searching for the best set of equipment and looking for that perfect set of enhancements on a weapon.
- Not dying: I hate dying in roguelikes. After all this time, it just annoys the hell out of me.

The great irony in the above list is that I've never implemented all of these things in my own roguelikes.

[jim]

I've got one. It's optional, heavily optional, and it's not always applicable, but I love it when it works.

The Base

This is the area where your character stashes all his hard-stolen loot and performs his alchemical transmutations, trains his henchmen, casts his ritual magic, breeds his horses, uses his telescope to augur the future, communes with his deity, forges his new sword, repairs his armor, researches new magic, writes his novels, seduces his romantic conquests, etc.

When not doing any of these things at the base, he is improving the structure of the base itself so as to better defend it from encroaching baddies.

[Vanguard]

Permadeath.

The risk adds so much tension and excitement.  Fear of the unknown gives roguelikes a better sense of mystery than you're likely to see in any other genre.   Being set against impossible odds and snatching victory from the jaws of defeat, not because you could abuse quicksave or anything like that, but because your tactics and preparation were just that good.  It's the best.

I've got one. It's optional, heavily optional, and it's not always applicable, but I love it when it works.

The Base

That's a good one!  I hadn't thought of that, but you're right.

[Ari Rahikkala]

Back when I used to play roguelikes more, my favourite moment in any game tended to be reading the first blessed scroll of identification in ADOM (usually in early CoC or late VD/PC). Yes, there is plenty of gameplay value in having complicated identification, but in ADOM you get so much loot it'd be difficult and tedious to identify all of it feature by feature. Usually while being next to an altar so you can bless all of your water too. You close the doors, read a scroll of ID, drink your stat increasing potions, maybe read some equipment-enhancing scrolls, sacrifice or drop the equipment you don't need... it's a moment of respite away from killing monsters, and you come out of it with less encumbrance, better stats, feeling refreshed and ready to take on challenges you never could handle before.

(it's funny how that works - sitting down and going through your equipment is far more of a reward in ADOM than gaining a level. I'm not sure roguelike developers really even see leveling as something to make the player look forward to... the only game I can think of offhand where I really anticipate gaining a level is DoomRL, whose traits are very powerful and directly linked to leveling)

[Z]

I don't like leveling. I think a smooth curve is better. That is, you still get stronger with experience, but you get e.g. a tiny bit stronger for every monster you kill, not a "level" stronger for every 50 monsters you kill.

[corremn]

I love leveling, it is what traditional RPGs are all about.  Striving for that next level, taking risks, hoarding equipment you cant use yet, and then finally you achieve it, and the sense of achievement hits you. Very rewarding. 
Roguelikes diminish this somewhat, but still its a major factor that I love about the genre. 

Players like to see noticeable differences not a smooth curve.  Its the same as replacing you mundane weapon with an awesome one, if you dont notice the difference between the weapons then why have them.  Players NEED accomplishments, without them the player will be quickly bored.

[Z]

I feel it is strange that a "role playing game" commonly means something like "a game about characters whose statistics are measured with numbers and who can raise these numbers by gaining experience".

When I play e.g. Crawl, I do not have any noticeable sense of accomplishment by getting a new character level (unless it is the highest level possible) or a new skill level. I get a sense of accomplishment when slaying a powerful monster (or overcoming another kind of challenge). There are games where the new level is celebrated e.g. by giving the player full health at this point, but I find it illogical.

The question of smooth curves versus sharp steps appears in many situations. Experience levels, stats (just weak/normal/strong? or 5-10-15 with all numbers in between? or 60-100-180?), resistance levels (there are lots of games which contain items which raise e.g. fire resistance by 5%, I don't think this is very fun; Crawl has 5 (6?) strongly distinguished levels and it feels better for tactics), speed (in ADOM you can have e.g. speed 101 instead of the usual 100, in another game you could have only 50%, 100% or 200%), ... I believe that, in most cases, a smooth curve feels realistic, but having only a few levels is better for tactical gameplay. (So, I have this Dexterity stat, which I can decide to raise from 15 to 16 at some point, it will raise my Evasion value using some formula which I do not understand, which will in turn determine the chance of avoiding the attack, by another formula which I do not understand... Good for simulation, but bad for tactics.)

[Fenrir]

We (well, maybe not me so much) throw around some interesting topics and ideas around here, but, when it gets down to experimenting, we have the "nDRLs". I think that a 7DRL event should be set up in a similar manner to the Experimental Gameplay Project. We choose a mechanical "theme" (leveling/item stats/dungeon generation/etc.) each time, then developers experiment with these themes to determine what works and what doesn't work.

[Bear]

I don't know about that. 

I think that the "NDRL" impulse is more about overcoming the tendency to lose focus and let the project back-burner forever than it is about experimenting.  It's a good medium for experimentation, but I don't think it's a good medium for trying out any new game mechanics except the ones that are very easy to code. 

Hmm.  Mixed opinions on leveling.  That's interesting.  I don't really have a strong opinion on that myself. 

I like the shorter power curves too (smaller difference in hits & damage between starting and winning characters) but they're harder to balance well. 

I think permadeath is a good thing; it's a game's way of demanding respect, and it leverages content so the player can keep getting accomplishment rushes off "further than I've been before" rather than just a linear playthrough based on backup files.

The Base-Building game is what's so great about Dwarf Fortress.  It supports as much depth of play as you want to put into it!  It's hard to work into a traditional "Quest" roguelike though. 

Re: "Anti-Farming" I totally agree it's an awesome subgame.   Encouraging it is the main reason why I plan to award score bonuses for smaller turncounts. 

The best I can think of for spoiler-proofing an identification subgame is this.  First have lots of item variability so individual items, even of the same type, have different range, damage, recharge rates, specials, and so on.  Second, have "rarities" - that is, types of "common" items that simply don't appear in all games.  If every game has an unknown number (say 5 to 15) "rarity" potion item types, you can't play spoiler-based process-of-elimination with any real certainty. And if those rarities are drawn from a large pool some of which are very rare, you'll occasionally be seeing potion types that you have not seen in any previous game. 

But an identification subgame needs to be over at some point.  I think before a character gets more than 50% of the way to a win, he should probably have a reasonably available and reliable means of fairly complete identification. 

[jim]

The Base-Building game is what's so great about Dwarf Fortress.  It supports as much depth of play as you want to put into it!  It's hard to work into a traditional "Quest" roguelike though. 

Totally. But I remember when I first successfully stormed the castle in Nethack and found myself with a base to end all bases. It was fun to store all my goodies in different treasure chambers and to line the entrances with Elbereth.

ADOM, meanwhile, is a game in serious need of a base. Chaos notwithstanding, the game has no real time limit, and the lengths people go to in order to "fake" a base (like leading the old barbarian back to the farmer's village) speak to the necessity of at least a safe storage closet.

Meanwhile, Omega has a base for sale at 50,000 gold, and Incursion (should it ever be finished) will have an upgradeable stronghold as a base of operations.

I think that a base can be a worthy addition to any (let's say) less linear roguelike, especially ones with an overworld or where maintenance / training / long terms planning comes into play.

[mariodonick]

Interesting thread I overlooked until now.

Setting and Story

Any game should have an interesting setting which gives the player character a reason for (whatever the ultimate goal of the game is). The setting should be consistent, it can have similarities to cliché settings, but should add its own flavor to these. Examples I like in this regard: Vanilla Angband (because of it's simple LoTR setting and it's deep depressing dungeons), Legerdemain (because it even has its own printed book). Examples I dislike in this regard: Nethack (too much inconsistency in monsters and items), Dungeon Crawl Stone Soup (often I forget why my character is in the dungeons at all).


Dungeons and World

I prefer non-persistant dungeons like in Angband and have the feeling that today most people disagree with this, because of some disadvantages, but it forces you to go deeper, deeper, deeper, if you don't want to explore all the paths to the exits of a level again. It's also good to have some kind of town level or even overworld in the first level. This strengthens the feeling of a consistent setting.

Dungeons should also have some things to play around with, for example oil on the ground that can be set on fire, or spider webs that can be destroyed by fire.


Monsters

Monster's should not be procedurally generated. Each kind of monster should have some speciality that distinguishes it from others, and the player should have to find ways to cope with this speciality. The preacher throws fire? Well, get fire resistant equipment, or resistance potions. There should also be some peaceful monsters and for some character classes, killing these should have a negative outcome (like no EXP). Monsters should also react on different dungeon states (e.g. getting hurt when standing inside fire).


Items

In general, items should be static, too, but identifying potions and rare and unique items should be required nevertheless, and there should be a SMALL amount of weapons with randomly chosen specials. It's good to have the possibility to combine items with other items, to create new items (such as combining bread and meat to a sandwich, or inserting a crystal into a socketed weapon). It's also good if enemies react on items thrown on them (such as getting peaceful, if food is given them), and if there are parts in the dungeons which react on items, too.


Inventory

Inventory size should be limited, because I simply like inventory micromanagement, and the decisions about item usage.


Quests

I like subquests, giving the possibility to get better equipment or money, and to get to know the setting and the story. I also like fed ex quests ("get me A from B") and simple kill quests ("kill X of type Y"), but many people don't.


Interface

As easy as possible. My ideal roguelike interface LOOKS like the interface of World of Warcraft or similar games and PLAYS like Diablo. As few keys as possible, and no huge boxes at the screen, except small status areas and an action bar. In console mode, of course, compromises have to be made. The interface should also let the player know what he can do in a given moment at the current position (not everything, but common things).

[Rya.Reisender]

I like roguelikes most when they are incredibly simple to play, but yet hard to master. I like it when the focus is on very short and simple decisions like "move up or up-left?" rather then complicated "I need to read a 20 pages guide book about it before deciding the whether I should join this religion or not" decisions.
I usually don't like item systems in roguelikes, because they usually ruin any fast-paced gameplay.
If the game is not fast-paced it has no replay value for me.