Are required resistances fun?

[Hi]

Runics are something brogue has
They are not really things like fire resistance.  They are things like 'when you are attacked, share the damage with all adjacent enemies." or "sometimes friendly temporary clones of you enemy appear when you get hit.".  Each runic is new toy that you have to adjust your tactics to fit.  Even the runic that makes you immune to gasses has more effect than just meaning you can ignore traps, it means it is beneficial to seek traps and trigger them so that the monsters will be gasses while you attack them.

[Vanguard]

It's also absolutely possible to beat Brogue without any of them.  They're nice effects that allow for different tactics without making your success excessively dependent on the item generator.  That's the way it should be.

[Gr3yling]

It's also absolutely possible to beat Brogue without any of them.  They're nice effects that allow for different tactics without making your success excessively dependent on the item generator.  That's the way it should be.

Just out of curiosity, what about luck in other aspects of play?  Isn't success in most roguelikes heavily luck dependent?  I mean, "hardcore" players (not me) seem to like the idea that you can lose a new character to a fireball trap.  Isn't it pretty widely accepted that you are a the mercy of luck in many ways when you play a roguelike?

[Vanguard]

I think the ideal roguelike should feature heavy randomness that influences both the player's long and short term decisions but at the same time be deep enough that good tactics and preparation can always overcome bad luck.

[Quendus]

I think the ideal roguelike should feature heavy randomness that influences both the player's long and short term decisions but at the same time be deep enough that good tactics and preparation can always overcome bad luck.

That's a good ideal, but I don't think it's possible to reach it without either compromising difficulty or introducing a dependence on spoilers. In my interpretation, Heavy randomness in a difficult game will always make the early stages deadly, and making good tactics and preparation able to overcome that in the early game would necessarily entail either a shallow game or a need for spoilers. Deep mechanics are not good friends with words like "always", especially in a stochastic environment.

I tried to think of a counterexample to this, but all that came to mind was my own game Encircled. It has less than ten spoilers (everything else comes from simple but rather abstract rules, which produce relatively deep positioning tactics that vary in each game), the weapons given to enemies and made available to the player have big effects on tactics (not so much on long-term strategy); it's difficult, but good tactics and a bit of strategy can mostly overcome bad luck (except sometimes the transition to ranged weapons in the midgame).

I think it fails the ideal because there isn't wide variation between games in the set of weapons generated, so similar weapons get generated on similar levels very often. I aim to improve that in future development, but development has been stalled for a while (and if it continues, the first priority will be to make the overly abstract rules more palatable to people who haven't learnt group theory).

[Vanguard]

Even if 100% fairness is an unattainable goal, it's good to come as close as you can without ruining your game's good qualities.

Like, if you start the player out with a few get-out-of-jail-free consumables, bad luck in the early game isn't as lethal.

[Darren Grey]

I think the ideal roguelike should feature heavy randomness that influences both the player's long and short term decisions but at the same time be deep enough that good tactics and preparation can always overcome bad luck.

I think randomness is a bad term for what roguelikes really do. The purpose of the randomness is to introduce the unknown, to make the game spoiler-proof and constantly tense, as you don't know if your plans will always succeed. But this hidden information doesn't need to come in the form of the dice roll - it can be based on complex AI interactions, or limited vision in a procedural environment, or restricted information on item effects. Over-reliance on dice rolls to achieve hidden information can just lead to a frustrating game, or at least a non-tactical one.

Someone mentioned the Tower of Eternal Flames. I think it's a great example of a thematic challenge in a game involving resistances, which can be overcome in a variety of ways. Yes, it's somewhat spoiler-reliant, but it doesn't insta-kill you and there are in-game hints to help you figure out the right ways to do it. And in a replayable game it remains a fresh and interesting challenge. Possibly the best part of all of ADOM.

As much as we love randomisation between games it's still the scripted challenges that we talk about most. Every character ends up a little bit different in their approach, but the central scripted challenge forms a good talking point in communities and a memorable milestone across every game.

[Vanguard]

I think randomness is a bad term for what roguelikes really do. The purpose of the randomness is to introduce the unknown, to make the game spoiler-proof and constantly tense, as you don't know if your plans will always succeed. But this hidden information doesn't need to come in the form of the dice roll - it can be based on complex AI interactions, or limited vision in a procedural environment, or restricted information on item effects. Over-reliance on dice rolls to achieve hidden information can just lead to a frustrating game, or at least a non-tactical one.

Agreed.

[Quendus]

You object to to word "random" in a lot of contexts (like map generation) as a lead-in to some important game-design considerations. I tend to agree with them, but not so much with the objection to "random" as a term to describe or motivate game mechanics.

It's just a technical term to describe stochastic systems in general, as well as a common function programming languages use to choose numbers from a uniform distribution. It will still apply to any method of procedural generation or game mechanics with a random component or any kind of hidden or external information.

I think "randomness" is not a bad word at all, and that the main point is that when designing a game one should always put thought into how content is generated, how the result of an action is determined, and how this will interact with the rest of the game, rather than adding a call to random() and forgetting about it. Complaining about a word whenever it's used is just a distraction.

"Randomness" is in fact pretty important to the concept of a roguelike. They're about beating a system that uses randomness to confound strategies based on foreknowledge.

[Vanguard]

Right but some roguelike developers and players worship randomness and want it used everywhere regardless of whether it makes sense.

Like, randomized combat is a good thing in small doses.  Unpredictability is fun and it can test the player's ability to react to unexpected situations.  But when you've got a significant chance to fail casting your spell and a significant chance to miss your target and the spell does something like 1d100 damage you're going to get wildly different results from the same set of actions.  The player is no longer fully in control of their own success.  That kind of system makes failure feel unfair even when it really was your own fault.

[Darren Grey]

Randomness is good and wonderful, and I like it when it's used appropriately. My point here is that it can be used for different things - hidden information is one, but variety can be another, and random can be entertaining in its own right. As a term it gets associated with the latter two, but in roguelike design one of its most important uses is the former, and there are other ways to achieve or enhance hidden information.

But yeah, I probably do rant about it too much

[NON]

Like, randomized combat is a good thing in small doses.  Unpredictability is fun and it can test the player's ability to react to unexpected situations.  But when you've got a significant chance to fail casting your spell [...]

I totally agree.

In the latest version of Infra Arcana, I erased a lot of senseless randomization for the reasons you mention. In particular, it used to be that spells had a certain percent chance of success (I tried slightly different approaches to this in many versions) - and damn it was frustrating to use! Now it always works, but drains spirit (mana), feels much more satisfying and tactically interesting.

Another similar issue was the resistance system - there were resistance vs physical effects and resistance vs mental effects, measured in percent. Various traits gave bonuses to these. When something tried to blind you for example, there was a check against your physical resistance - if it was 40%, you had a 40% chance to avoid the effect. It just felt... really random, in bad and confusing way. Even if you had very high resistance (like 80%), you could still be afflicted many times in a row, making you wonder if those traits really did any good at all.

Something like a shuffle bag could possibly have solved it to some degree. But I chose to rip out the system and change it to something deterministic. This feels more fun to me. If something blinds you, then you're blind - unless you are resistant to blindness, then you're guaranteed to be immune¹. It's certainly more tactical and challenging than just rolling the dice and hoping for the best.

¹IIRC Nethack uses this system, and have often been criticized for it - but mostly because (some) resistances are too easy to attain.

[Darren Grey]

You can have the resistance be a level that reduces the duration of the blindness. Someone hits you with a weak blind and you have a strong resist then you reduce it to 0, and there are various grades in between. This gives you something inbetween the chance system and the on/off system.

[Vanguard]

You can have the resistance be a level that reduces the duration of the blindness. Someone hits you with a weak blind and you have a strong resist then you reduce it to 0, and there are various grades in between. This gives you something inbetween the chance system and the on/off system.

Yeah that's a lot better than complete immunity or a % chance to resist.

I also really like the way Dark Souls handles resistances.  If you're 60% resistant to poison or whatever and something tries to poison you, it'll fill up 40% of your poison meter.  If they do it again right after it goes up to 80%.  The third attempt will put it past 100% and only then do you actually get poisoned.  Your condition meters drain over time so if you win a fight or run away after being partially poisoned you don't suffer any consequence.

Another idea I came up with was making resistances delay conditions instead of preventing them.  So something casts blind on you and you have 5 points of blindness resistance, nothing happens immediately but you go blind five turns later.  If they cast it a second time the duration gets shorter.  There should also be a way of treating conditions for both NPCs and the player, something simple like spending X turns doing to clear all upcoming conditions.  That way it isn't viable to cast something really nasty and hide until it goes off, and if you win or escape from a fight before your own conditions go off, you don't have to deal with them.

Both of these systems would really powerful conditions to become viable and tactically interesting without being cheap.  You could even allow instant death spells to affect the player and every enemy in the game without turning it into a luckfest.

[miki151]

I also really like the way Dark Souls handles resistances.  If you're 60% resistant to poison or whatever and something tries to poison you, it'll fill up 40% of your poison meter.  If they do it again right after it goes up to 80%.  The third attempt will put it past 100% and only then do you actually get poisoned.  Your condition meters drain over time so if you win a fight or run away after being partially poisoned you don't suffer any consequence.

Does the game indicate the current value of the meter in the UI? If not then it's pretty much like randomness to me, just through a hidden variable than RNG.