I've been thinking about roguelikes and roguelites of late, and I'm (genuinely) curious about something:

In my (admittedly limited) experience and to the best of my knowledge, highly-traditional roguelikes tend to use combat based heavily around random numbers: you bump into (or click on, or whatever) an enemy, one or more "die-rolls" are performed to determine whether you hit, and if so, another is performed to determine your damage. If the enemy survives, the turn passes to it. Barring a few tactics like kiting the foe into a trap or taking advantage of mechanics like attacks of opportunity, there's little apparent space for player skill in the game

So, in particular to fans of traditional roguelikes, do you enjoy this sort of combat, and if so, what is it that you find enjoyable about it?

These questions came about (at least in part) as a result of the fact that I've recently been playing both Delver and Pixel Dungeon. Delver I've been very much enjoying, while Pixel Dungeon I'm finding somewhat frustrating, and I've concluded that the main difference is in the combat: I don't feel that I have much real agency in Pixel Dungeon. When a foe kills me in Delver, I generally (and barring one or two annoying enemies) feel that it's because of a mistake that I made: I foolishly blundered into a teleportation trap and found myself surrounded by foes and disoriented, or didn't manage to avoid its attacks, or kept missing with my wands, or some such thing. In Pixel Dungeon, I feel that my skill (or lack thereof) has little to do with anything: I lose because the metaphorical "dice" say I lose, and that I find frustrating.

(To some degree I also dislike having to identify items, but that's primarily as a result of the general paucity in methods of doing so: I seem to end up either not using potentially-valuable items, or taking the reckless (but in some cases worthwhile) approach of experimentation. That, however, is more a niggle than a major annoyance, and I might enjoy the mechanic if there were simply more resources for identification, perhaps at various costs or tradeoffs. It might be worth noting that Delver seems to have fairly light costs for experimentation, making it a fairly viable course of action in that game.)

On examination, I feel that I enjoy most of the elements of traditional roguelikes, save the combat. I like randomly-generated dungeons, random (and better yet, randomly-generated) loot, and random enemy placement. I just don't like randomly-resolved combat.

However, it does seem that there are those who enjoy such gameplay, and thus I'm here looking for illumination on that perspective.

(One quick note: I don't want to get into the morass that is the debate about what does or does not qualify as a "roguelike", and what falls under the label of "roguelite". That's another argument for another thread! ^^; )

In traditional styles the 'skill' is in decisions and thinking several moves ahead. Yes, it is highly random, and you can get sessions where you get a series of really bad rolls which lead to situations where you can't see a way out, but that is part of the game. Often if you die it was because you over extended yourself, moved too deep and let yourself get caught fighting too much too close together and couldn't make use of healing methods. It Is a risk and gamble when you move to 'the next square', but you're not really meant to just keep moving forward blindly. Careful thinking with regards to the mechanics is where the skill lies in many of these games, and making decisions like taking several steps back before moving forward to make sure the combat happens when and where it is most favourable. (Usually very applicable to ones that implement some kind of character collision detection, and forcing combat into bottle necks where you aren't as easily flanked for example.)

For what it's worth, I do employ at least some of those tactics that you describe (bottlenecking in particular)--but even so, a single low-level rat can still kill my character if the dice say so, simply by giving me a string of bad rolls and it a string of good ones.

In traditional styles the 'skill' is in decisions and thinking several moves ahead.
...
Careful thinking with regards to the mechanics is where the skill lies in many of these games ...

Hmm... Okay, that does make sense and I do see the appeal of that, I do believe--but isn't that appeal undercut when all one's careful strategy and tactics can be undone by the whim of the "dice"?

I found the following points on the Pixel Dungeon wiki:

Even with ideal play, your results will be heavily based on luck and you will only rarely have a “successful” playthrough.

(Reference)

And from a different section:

You may expect 1 successful game for 7 to 10 tries for warrior,1 for 12 to 15 for mage, and 1 for 20 to 25 rogue, but these numbers are just an estimation.

(I presume that those numbers are given presuming ideal play.)
(Reference)

This seems to indicate that even if one makes no mistakes at all, one's skill can--and is very likely to be--undone entirely by a set of bad "rolls".

Naturally, the numbers above are specific to Pixel Dungeon and may not reflect other traditional roguelikes; however, my impression--and I do stand to be corrected--is that this pattern, on one scale or another, is fairly typical of the genre.

Yes, it is highly random, and you can get sessions where you get a series of really bad rolls which lead to situations where you can't see a way out, but that is part of the game.

Fair enough.

This prompts a new question: How would you feel about a game that was just the same as a traditional roguelike in every way--cruelly difficult, sparing with resources, turn-based, and requiring clever and careful strategic play in order to win--but in which combat was based entirely on player-skill (presuming that the skill involved wasn't an action-game skill, and was one that you enjoyed using--perhaps some sort of microcosm of the base gameplay), such that a hypothetical "perfect" player could expect to win every or nearly-every game by virtue of skill? Would that improve the game for you, lessen it, or have no effect?

One thing that might be worth mentioning regarding my own perspective: I don't enjoy casino-style gambling--slot machines, roulette, etc. From my perspective there's little real attraction to them.

Dice based computer games really do not do too well with "modern" players ... this guy actually explains it a lot better than I can ...

Dont look at pixel dungeon beyond a time waster. Theres no strategy beyond constantly rerolling a game till you find a decent weapon on the first 2 floors.

Never played delver before. If youre on a computer though you should check out dungeon crawl stone soup. Out of all the roguelikes if played it has the best tactics in my oppinion.

It can get tedious sometimes do to its length. But at any point in the game anything cankill you if youre being careless. 99% of everything has a use. Amd dieing with anything in your inventory usually means you couldve survived if you used what was in your inventory.

Roguelike games, such as Nethack and family, are tactics games.

The games you mentioned are crappy games with random outcomes. These are NOT tactics games.

Tactics games are based on stats and probabilities.

For example, let's assume you are playing nethack. You see a creature you aren't familiar with, you can look it up in the monster manual. You're fairly high level, and the creature, an Archon, looks tough. Possibly five attacks per turn: gaze attack that stuns+blinds for 2d6 turns, two projectile weapons at 2d4, claw attack of 1d8, and an occasional spell that at your current level is 9d6. Resistant to fire, cold, sleep, electricity, and poison. Speed 16, can fly/float, can see invisible With all that information you can decide if you want to run away, or if you want to risk attacking it. As it deals damage you can track how much damage you did to it, and tactically decide what attacks you want.

Dont look at pixel dungeon beyond a time waster.

The games you mentioned are <omitted>y games with random outcomes. These are NOT tactics games.

Heh, it's quite possible that I've simply been primarily exposed to poor examples of the genre.

Thankfully Pixel Dungeon was at least free.

For reference, the roguelikes (excluding roguelites and similar) that offhand I recall playing are:

• Rogue
  • Which is understandably a primitive form of the genre.
• Pixel Dungeon
  • Which I now gather is a poor implementation of the genre.
• Delver
  • Its inclusion as a "roguelike" is probably debatable--but again, I don't really want to get into the argument regarding definitions in this thread. ^^; Suffice it to say that, for me, personally and intuitively, it feels more like a roguelike than a roguelite. It's also rather fun. (It's not really a tactical game, however.)
• A brief foray into Nethack
  • This was some years ago; I think that at the time I was scared away by the sheer number of possibilities. ^^;
• A few others, I think, but which I only vaguely recall.
  • Again, I believe that I tried these many years ago, now.

When I consider games outside of the roguelike genre there have, I believe, been other cases in which randomly-determined combat results were a frustration to me. In Battle for Wesnoth I was an inveterate save-scummer (for which I feel no shame), because having my troops keep missing due to unlucky rolls just felt frustrating--especially in a game in which holding onto individual units gave significant benefits. Morrowind suffered from being in the first-person, thus making the random determination of hits feel somewhat unfair.

That said, I will admit that, as far as I recall, I didn't mind the process as (I believe) implemented in Neverwinter Nights or the Baldur's Gate games--and a large part of this, I think, is that the effect of the random die-rolls seems to have been far lower than that of actual tactics. Poor tactics could lose a battle, while poor die-rolls might cost one or two extra spells or potions, or prompt a change in tactics. As you say, they were tactical games.

I think that I'd still prefer that the random die-rolls weren't present in them, but in games like the RPGs just listed they at least didn't feel as prevalent as in the roguelikes that I'm familiar with.

With all that information you can decide if you want to run away, or if you want to risk attacking it.

Heh, my experience thus far--in Rogue and Pixel Dungeon, at least--has been that running away has tended to be counter-productive: the monsters just follow, leaving me to either turn and face them or run a high risk of ending up sandwiched between the pursuer and some new foe ahead, sealing my fate. (I note here that random spawning of monsters means that even previously-visited sections of the dungeon may not be safe to run to.)

I take it, then, that other games make running away a more viable tactic?

If youre on a computer though you should check out dungeon crawl stone soup.

Hmm... Having looked that up (I'm playing on an Android 'phone at the moment, but there does seem to be an Android version), it looks as though Stone Soup might be quite good. There are some elements (primarily in the setting and writing) that put me off a little as a matter of personal preference, but I intend to give it further consideration--thank you for the recommendation.

I take it, then, that other games make running away a more viable tactic?

Yes, in most of the big roguelike games running -- at least in the short term -- is a very good tactic. Sometimes it means locking a door behind you and staying in the same level, sometimes it means fleeing to another level, sometimes it means more dramatic measures. However, some creatures are smart enough to open or smash down doors, some creatures are smart enough to follow between levels.

Nethack and family (Angband, etc) are very numbers-heavy tactical games. By looking over your character's stats and knowing the dungeon depth you can quickly see if the player is in over their heads, or if they have a moderately good chance of survival.

Your odds of survival are quite slim, most nethack players die the first few hundred times they play. It took over 3 years before my first time finishing the game. If you are not used to tactics games, to going slow, you will likely lose. At the first sign of trouble list your inventory, save, and quit. Give it some time, figure out your options, and then carefully react. I've had times where in my rush to play I died while carrying blessed scrolls of genocide -- I could have removed the creatures from existence -- or carrying scrolls of taming -- I could have made them into pets or at least peaceful -- and amulets of lifesaving. Never be in a rush with a turn-based tactical game.

For what it's worth, I do employ at least some of those tactics that you describe (bottlenecking in particular)--but even so, a single low-level rat can still kill my character if the dice say so, simply by giving me a string of bad rolls and it a string of good ones.

One way to alleviate that is to not use "pure random", but to use a more controlled form of fake randomness.

Imagine a player has 70% accuracy. Instead of rolling "if(random() < 0.7)" for every attack, which yes, could make the player miss two hundred times in a row, an alternative method is to generate the player's next N hits, and then shuffle them.

For example, with 70% accuracy, you might populate a stack of the next 100 attacks, fill it with 70 "hits" and 30 "misses", and then shuffle the stack. For each attack, you pop off the top of the stack, and when the stack reaches empty, you regenerate it.

This guarantees that the player will never happen to get completely crappy roles (and it also guarantees that the player will miss some of the time).

You can apply additional rules to it: Such as making the misses evenly distributed instead of clustered together. You can also mix in additional effects besides just missing and hitting, like criticals, enemy ripostes, and so on.

And you can still add in some randomness. If the player's accuracy is 70%, then generate a stack of 100 attacks where 70% plus or minus 5% is a hit. Tweak it to your satisfaction. But control it, as a designer.

The same applies to all forms of randomization. You don't want every chest to randomly be trash, and you don't want every chest to randomly be 'epic gear of unbalanced power'. You don't want every door to be trapped, and you don't want zero doors to be trapped.

Yes, in most of the big roguelike games running -- at least in the short term -- is a very good tactic. Sometimes it means locking a door behind you and staying in the same level, sometimes it means fleeing to another level, sometimes it means more dramatic measures. However, some creatures are smart enough to open or smash down doors, some creatures are smart enough to follow between levels.

Interesting--and perhaps another point in favour of the theory that I've simply not been playing the right games.

If you are not used to tactics games, to going slow, you will likely lose.

I've played other tactical games (I mentioned the Baldur's Gate games, for example; the specific style of combat may be different, but it's far from a mindless hack-and-slash, as I recall it). I'm probably not accustomed to tactics games that are quite as numbers-heavy as you're describing the Nethack family to be, in all fairness.

One way to alleviate that is to not use "pure random", but to use a more controlled form of fake randomness.

...

That actually looks like a pretty decent and quite clever implementation, actually; I could see myself enjoying a game based on such. ^_^

... and you don't want every chest to randomly be 'epic gear of unbalanced power'.

A guy can dream, can't he? :P

So, based on what I've learned thus far in this thread. I think that if I were to look for a roguelike now, I'd probably look for something less intensive than the Nethack family, but with more depth and better gameplay than Rogue or Pixel Dungeon. I'd still prefer that random numbers weren't involved in the resolution of combat--I'd be enjoying the combat in spite of the use of random numbers rather than because of it, I feel, but would at least be enjoying said combat--but it does seem that there are likely implementations that I'd be happy-enough with. (I don't believe that random numbers as the basis of combat resolution are a requirement for all forms of tactical gameplay.)