Luring (was: DCSS has a power creep problem)

[crate]

If the problem is only retreating 30+ steps then there are lots of things that could add a cost to retreating that would make it clearly not a good idea in general. However I'm already not convinced that you want to retreat that far anyway in current crawl, and the "solutions" I can think of probably have more unintentional effects than intended effects.

Speed 11 monsters would obviously work here, and I'm pretty sure (about 80%) that it would improve the game for experienced players. The biggest problem is I'm not sure that it would improve crawl for new players at all, and might be a pretty sizable turn-off. Additionally, there would be some difficulty communicating that faster-than-the-player is the default monster speed via xv.

[sanka]

My personal opinion, altough I am not a very good player as some others in this thread:

- I am also not convinced that retreating 30+ steps is optimal in most situations in current crawl
- I think moving away from monsters being optimal in most situations is a good thing in crawl, which leads to more interesting and complex gameplay than if it was not true
- I would like if most monsters were speed 11 instead of 10, because sometimes retreating can be really annoyingly long otherwise. Same speed as character monsters are not as interesting, and slower ones are usually even worse. Of course there could be some variance but the marginally faster ones are the most interesting, therefore they should be the default. You are right tough that it may confuse newer players.

[kuniqs]

Let's give monsters a 120 dmg attack when they pursue you

[Lasty]

I don't think speed 11 monsters is a good solution. It means that (for most chars) all monsters are sticky. Once they reach melee range, they will never leave it unless you use a consumable. That leads to situations where your character is almost certainly going to die, but can take dozens or hundreds of turns before actually ending the game -- e.g. you get into melee range of a D:2 ogre w/ a char who has very little chance to survive it; you can probably run away from it for a very long time before it eventually gets the swing that kills you.

I would far prefer a solution where (example numbers) after 6 turns of moving towards the player, the monster has a (num_turns_moving / num_turns_moving + 10) chance of automatically getting swiftness applied to them. It still makes dragging monsters a long way painful, but it means that if a speed 10 monster doesn't kill you during the swiftness, you are guaranteed to be able to escape; either way, the encounter will definitely end.

[crate]

Once they [monsters] reach melee range, they will never leave it unless you use a consumable.

For what it's worth, I'm not sure how this is a problem. This is pretty much how melee combat in crawl is supposed to work, given that pillar-dancing is seen as undesirable and to-be-prevented. If the goal of energy randomization was to make combat un-sticky, okay I guess I'm wrong here (my understanding is this was not the reasoning behind energy randomization, and I don't recall it ever being brought up as a reason for energy randomization to continue to exist, but obviously I am not a dev).

I mean, you already have that sticky-monster situation with a monster that is actually likely to appear on d:2 (adders). No one playing a normal-speed race tries to flee from adders by running after they get next to you, you either use your consumables or you roll the dice and fight. And adders close distance twice as fast as speed-11-movement would. I don't think I've seen "adders are sticky" brought up as an actual design problem before. (Complaints about adders being too strong, sure, but not the sticky-monsters thing.)

I'm not saying that giving monsters swiftness is worse (and there are plenty of good arguments for avoiding speed 11 monsters even if you think "luring", whatever that means in the context of this topic, is a real problem), but I find your reasoning a bit strange.

[ydeve]

Once they [monsters] reach melee range, they will never leave it unless you use a consumable.

For what it's worth, I'm not sure how this is a problem. This is pretty much how melee combat in crawl is supposed to work, given that pillar-dancing is seen as undesirable and to-be-prevented. If the goal of energy randomization was to make combat un-sticky, okay I guess I'm wrong here (my understanding is this was not the reasoning behind energy randomization, and I don't recall it ever being brought up as a reason for energy randomization to continue to exist, but obviously I am not a dev).

Maybe this clears it up. Speed 11 monsters prevent fleeing from fights in almost all cases, not just for melee. If you turn a corner and see a pack of monsters you can't handle, you have to burn a consumable, because just walking away (before they reach melee range) is no longer an option.

Instead of reducing tedious behavior, this would increase it, as players would be encouraged to run back to the stairs more often. It replaces luring in general with luring+stairdancing (they're the same thing now - you can still break up packs using stairs, but it's more tedious than luring ever was). Except not engaging in the behavior is worse for you than not luring now.

In other words, crawl becomes Zot defense with the option of stairdancing.

[crate]

Maybe this clears it up. Speed 11 monsters prevent fleeing from fights in almost all cases, not just for melee. If you turn a corner and see a pack of monsters you can't handle, you have to burn a consumable, because just walking away (before they reach melee range) is no longer an option.

This is a different argument, unless Lasty was being very sloppy with his language (in re-reading, perhaps this is the case; if so, being precise is important!). Lasty is talking about the situation where monsters are already in melee range, and they are already sticky in that case (or, are supposed to be, but sort of aren't because of energy randomization, except they really are mostly sticky, because energy randomization doesn't really accomplish making things un-sticky).

I also don't think that speed 11 is as inescapable as you suggest here (they close distances really slowly. Did you know that ugly things are already basically speed-11-movement?), but I admit that this is a situation where you would need to play to really find out. It's also not really any different with respect to stairs, since you already need to use stairs to avoid a speed-10 monster eventually closing with you (you at some point need to do things that are not just moving, during which the monster will get adjacent to you). Just in some cases they will close on you before you reach stairs, so there is a resolution more quickly, because after they get adjacent you stop running.

[archaeo]

I also don't think that speed 11 is as inescapable as you suggest here (they close distances really slowly. Did you know that ugly things are already basically speed-11-movement?), but I admit that this is a situation where you would need to play to really find out.

Patches branches welcome!

[Lasty]

Once they [monsters] reach melee range, they will never leave it unless you use a consumable.

For what it's worth, I'm not sure how this is a problem. This is pretty much how melee combat in crawl is supposed to work, given that pillar-dancing is seen as undesirable and to-be-prevented. If the goal of energy randomization was to make combat un-sticky, okay I guess I'm wrong here (my understanding is this was not the reasoning behind energy randomization, and I don't recall it ever being brought up as a reason for energy randomization to continue to exist, but obviously I am not a dev).

I mean, you already have that sticky-monster situation with a monster that is actually likely to appear on d:2 (adders). No one playing a normal-speed race tries to flee from adders by running after they get next to you, you either use your consumables or you roll the dice and fight. And adders close distance twice as fast as speed-11-movement would. I don't think I've seen "adders are sticky" brought up as an actual design problem before. (Complaints about adders being too strong, sure, but not the sticky-monsters thing.)

It's true that adders are sticky, but in a better way: they're much faster than you. The period where you can't escape and are likely to die but haven't died yet is much, much smaller. Having all monsters be faster than you isn't the problem in my mind, having all monsters be just barely faster than you is. For the scenario I'm discussing rather have all monsters be speed 13 than speed 11, but that's obviously a much bigger balance swing. That's one reason I like the swiftness suggestion: the speed increase is larger (and can be any arbitrary value of larger), but it doesn't impact balance nearly as much.

As for the purpose of energy randomization, that predates my involvement with crawl, so I have no perspective to add on that front. I assumed the goal was to make running away eventually end, either with the player's death or escape. The only possible other purpose I can see it serving is to make it slightly harder to count AUT on an approaching monster, which to me seems like a smaller problem than the solution is.

[all before]

I could be wrong here, but it seems like speed-11 monsters would lead to more luring, not less. (By luring, I mean walking monsters back to stairs, or I guess back far into already cleared territory, though I think stairs is nearly always the best choice.) Take speed-11 hydras as an example. Currently characters who lack the ability to kill a lair hydra (or don't want to spend a consumable) have the option to simply retreat to the stairs and go back down another. In the case of a slightly faster hydra, retreating is still an option only if the hydra is encountered relatively near the stairs. Which means that a weak character really wants to avoid making any noise far from the stairs. So, when the character encounters popcorn monsters that it would normally just plow through, it should instead lead these monsters back to the stairs to avoid potentially attracting the attention of a hydra. Basically, speed-11 increases the incentive to lure non-threatening monsters.

[duvessa]

https://crawl.develz.org/wiki/doku.php?id=graveyard:speed_randomization
Energy randomization was introduced in an attempt to get rid of pillar dancing. (Here's my last argument about it.)

I am sure everyone arguing against randomized energy is aware it was introduced to combat pillar dancing. This is why, in my first post in this thread, I stated that it does nothing to combat pillar dancing, except in the single case of orc priests. Other monsters either:
1. are speed 10, do not have ranged attacks or a means of becoming not (effectively) speed 10. Can be "pillar danced" in almost exactly the same way with or without energy randomization, you just use 1 more wall tile as part of the "pillar".
2. are speed 10, have means of becoming effectively not speed 10 (haste, slow, confuse). Could never be pillar danced in the first place because they can haste/slow/confuse you after you attack.
3. are slower than speed 10, same result as category 1 except you don't even need walls.
4. are faster than speed 10, same as category 2.
5. do not appear in the first few dungeon levels, after which pillar dancing is useless against any monster.

Now if every monster in the early game were the same as an orc priest (speed 10, ranged attack that does damage and damage only, nothing else) then energy randomization would seem a lot more reasonable to me. But the vast majority of early-game monsters are in category 1, and pillar dancing/kiting them takes a much larger amount of real-life time with energy randomization. Furthermore, randomized energy introduces

in the cases where you do intend to fight a speed 10 monster, it is optimal to move away from it until it is next to you, because that minimizes the number of actions it gets; previously you could just wait.

which is really horribly incredibly awful and I consider it far, far more tedious than pillar dancing orc priests.

Lasty, do you think that 11 is a worse "default" movement speed than 10?

[Lasty]

duvessa: In the absence of any other systems that would make an encounter end, 11 is probably better than 10. 11 at least probably makes encounters end, where 10 doesn't.

[HardboiledGargoyle]

crate, I feel you're retreading a lot of our discussion with your cogitations. Much of this thread has been dedicated to ironing out the actual problem being discussed here. We've been breaking each other's backs over preciseness. Personally, I think that the worst kind of luring is having a creature notice you and give pursuit while its buddies just go to the location of the shout and start wandering. Which is exactly the thing that makes luring "trivialize" Crawl. Repositioning doesn't do that.

If you really want to make players charge in against monsters then the only real option I see is to implement a very harsh clock of some sort for each individual monster. (Because crawl regen is so slow, it must be per-monster; otherwise retreating to better terrain will prevent enough damage that a global clock actually only encourages luring.) But any implementation of such a clock (it must be very harsh to encourage any sort of forward movement) is sure to punish weak characters severely, which surely causes greater problems than it solves.

OK, I'm not certain what you mean by "harsh" and what you imagine the clock would do. But especially, I do not understand any of your reasoning beyond the first sentence in the above quote. However, I agree with the gist of "if you want players to charge, implement a short timer for individual monsters". And yet I disagree about it punishing weak characters severely: it is only a matter of equally valid perspectives whether this would be a debuff to charged monsters or a buff to lured monsters. An example might be: dazed status, and reduced monster awareness score, for several turns after waking up. Justification: you gain an advantage in taking the enemy by surprise. This is actually extremely intuitive and what newbies always do when they first begin playing Crawl, even if they quickly learn not to do it. Current Crawl monsters are like those burly guys in movies that blink wide awake and put a knife to your throat when you gently tap them to wake them up. Unlike probably all other posters in the thread, I think the clock should depend on time/turns only, and be absolutely independent of the monster taking steps toward the player, or the player taking steps away from the monster, which would be weird.

[Utis]

If the problem is only retreating 30+ steps then there are lots of things that could add a cost to retreating that would make it clearly not a good idea in general. However I'm already not convinced that you want to retreat that far anyway in current crawl, and the "solutions" I can think of probably have more unintentional effects than intended effects.

(Emphasis mine.)

Thank you! I was really beginning to doubt myself while reading this thread. My gut feeling is that situations are quite rare where one wants to retreat 20+ steps; and they're rather satisfying if they occur at all. With regard to speed 11: I don't mind much, but all_before's point seems plausible. In spider I'm retreating more than in other levels, since letting one or two opponents get a couple of free hits is much better than getting swarmed.

If retreating 30+ steps is really such a good thing in general that it needs to be discouraged, then I like the idea best that retreating so far should alert more monsters. Only, that I think it should occur in a way that is perceptible within the game world, not through some obscure clock. E.g. let monsters start to shout really loud if they pursue the player for more than 10 turns without attacking. That would be actually flavourful.

[dpeg]

There's been an experiment with movement speed 11 for all monsters that currently have speed 10. That sounds drastic, but I think it's a good idea. However, it does address pillar dancing rather than luring -- the movement speed difference between 10 and 11 is so small that you can lure more than one LOS radius away.

[all before]

dpeg: an experiment as in an experimental branch? is it online?

[ontoclasm]

Here: a patch that increases monster walk/swim speed and removes energy randomization. It's a tiny little change, so there are tons of edge cases that I blatantly ignored; e.g. fast species, boots of running, etc. etc. etc. But you can get the general idea of how it plays.

(One consequence of removing energy randomization is that you can predict exactly when a monster will get a double-move; it happens exactly every 10 steps. This is probably bad but I don't care enough to do anything about it.)

n.b.: I don't want this merged into master. I made it because it irritates me when people spend dozens of pages theorycrafting and making definitive statements about something they've never tried, especially when it took me -- an amateur programmer with only a vague understanding of Crawl's code -- about 5 minutes to put together this patch. Try it first, then spend dozens of pages arguing about it.

e: Just realized it might help to explain how to use a patch: 1. Get git and a copy of the source code. 2. Put the patch in your source folder (or wherever). 3. Run 'git am speed11.patch' to apply the commits. 4. Make Crawl.

[Rast]

e: Just realized it might help to explain how to use a patch:

You will get 100x as much people trying your patch if it's available for immediate play on a public server rather than only available as a patch.

[ontoclasm]

Well, yeah. But I don't have a server to put it on

[Siegurt]

"One of the best tactics is luring. it is cheap cost action so you should always do that when not turn attacks. This unfun tactics completely destroys the game balance. "
-- everyone except many Japanese agree.

I disagree and am not Japanese, I don't think the problem with luring is balance, the problem with Luring is that it gives you a small advantage for no cost other than a lot of boring effort.

Resting is very different in how the games expectations are laid out. I really feel like you consistently complain about the fundamental aspects of Crawl's design in such a way that I can't even feel like you *like* the game or even the game genre. You consistently propose remaking it into a completely different game at such a fundamental level that I feel like you should just go look for something very different that is at least a little closer to your ideal.

[radinms]

I really feel like you consistently complain about the fundamental aspects of Crawl's design in such a way that I can't even feel like you *like* the game or even the game genre. You consistently propose remaking it into a completely different game at such a fundamental level that I feel like you should just go look for something very different that is at least a little closer to your ideal.

I really feel like people consistently complain about luring, that is a core tactics of roguelike games, i believe, should create different game, not crawl.

[Siegurt]

I really feel like you consistently complain about the fundamental aspects of Crawl's design in such a way that I can't even feel like you *like* the game or even the game genre. You consistently propose remaking it into a completely different game at such a fundamental level that I feel like you should just go look for something very different that is at least a little closer to your ideal.

I really feel like people consistently complain about luring, that is a core tactics of roguelike games, i believe, should create different game, not crawl.

That's interesting, since you don't need to lure in games where there's no risk added by fighting in the same locale, a natural consequence of Crawl's noise system, and there's no additional risk posed by fighting in place with Rogue or Nethack (or Moria for that matter) note that "luring" in the crawl sense, is different from "tactical positioning" aka "stand in a hallway so only one thing can beat on you at a time" Luring is about changing the *general* location of a fight, not the position in which you're standing. If you mean that "stand in the nearest 1 tile wide corridor" is part of the core tactics of roguelike games, then you're right, but that' isn't what people complain about.

[radinms]

there's no risk added by fighting in the same locale, a natural consequence of Crawl's noise system,

it is impossible because you can't control noise *freely*.

"luring" in the crawl sense, is different from "tactical positioning" aka "stand in a hallway so only one thing can beat on you at a time" Luring is about changing the *general* location of a fight, not the position in which you're standing. If you mean that "stand in the nearest 1 tile wide corridor" is part of the core tactics of roguelike games, then you're right, but that' isn't what people complain about.

"stand in the nearest 1 tile wide corridor" is obviously part of luring.
to lure enemies into safe areas by using noise is obviously part of "tactical positioning".

[ydeve]

When I lure, I don't stand in the nearest corridor and fight. I'd still be near the black and other monsters could be attracted by the noise. (This is not called "luring," this is called "positioning"). Sure, I'd fight them one at a time, but I wouldn't be able to rest in between.

Instead, I'll walk 1.5-2.5 screens back and pummel them whether I'm in a corridor or not. No other monsters are likely to show up anyway. Plus I can rest up to full hp before going to lure the next monster.

Edit: To clarify, the first doesn't change the general location of the fight, the second does.
Also, the first does not make use of noise to catch the attention of monsters one at a time, instead it's lack of noise control has the possibility of making more monsters find you.

As an example of using noise for luring, you stand back from a room that you can't see into and throw rocks into it. Once a monster stumbles into view, you start running away. The other monster's wouldn't have been able to pinpoint where you are, so only the first monster follows you.

[HisMajestyBOB]

If the issue with luring is that the player is using noise in a way that increases tedium (by slowly bring one monster out at a time through shouting or combat, and avoiding situations where something unknown and nasty is attracted by noise), then why not remove the noise mechanic? Instead, change monster behavior so that:
1. All monsters in a room awaken when one awakes.
2. Monsters are tied to a room* and don't wander more than X distance away from the room.

This would prevent luring one monster away and avoiding group fights and also prevent stair-dancing. One drawback is that monsters become more easily avoidable, but I would argue that this is not a bad thing:
1. If the player avoids monsters, the player doesn't get XP and thus advances more slowly.
2. If the player avoids the monsters, the player doesn't clear the room and misses out on any treasures therein.**
3. This removes additional tedium that occurs when the player must go to great lengths to avoid a nasty monster wandering the level. If I find Grinder on D:2, I have to tediously avoid him whenever I see him by going up a level and down another staircase whenever he appears. If he's tied to a particular area of the level, I only have to avoid that part of the level.

* Doesn't necessarily mean a single room with 4 walls and a door. This could be an area of the level of a certain size and/or contiguous area.
** Item gen could be changed to emphasize placing items closer to monsters instead of spread throughout the level.

[ydeve]

This would also result in making escaping any encounter trivial (don't like that black mamba, just walk away), unless the "rooms" were big. But if the rooms were big, you just made luring easier to do.

[bananaken]

If the issue with luring is that the player is using noise in a way that increases tedium (by slowly bring one monster out at a time through shouting or combat, and avoiding situations where something unknown and nasty is attracted by noise), then why not remove the noise mechanic? Instead, change monster behavior so that:
1. All monsters in a room awaken when one awakes.
2. Monsters are tied to a room* and don't wander more than X distance away from the room.

Something similar to 2. was implemented for Elf:3 a while back:

  Code:

--------------------------------------------------------------------------------
38f3aaf | Steve Melenchuk | 2015-11-22 12:30:09 -0700

Mark the Elf:3 vault enemies as patrolling.
They're supposed to be guarding the loot, and having them be lured out
too easily trivialises the difficulty of the vaults.

Though the purpose was to nerf luring strategies, and all it did was make it so you can slowly kill enemies at the edge of their patrol distance. Instead of making noise and luring back to a killhole, you would make noise and lure enemies to the edge of their patrolling distance. It was often *easier* because you could fight only a few enemies at a time despite making tons of noise at your location.

Personally, abusing noise and monster luring in dangerous places like Elf:3 are one of the fun highlights for me during a Crawl run.

[HisMajestyBOB]

This would also result in making escaping any encounter trivial (don't like that black mamba, just walk away), unless the "rooms" were big. But if the rooms were big, you just made luring easier to do.

The idea is that it would make escaping easier, but at the same time you're losing out on XP and treasure and thus not advancing in the game. The size of the "rooms" would need some work, but it shouldn't make luring easier as all enemies immediately become alert once one in the room is alert. Effectively you're breaking each level down into set tactical battles that the player can choose to engage in or not.

Though the purpose was to nerf luring strategies, and all it did was make it so you can slowly kill enemies at the edge of their patrol distance. Instead of making noise and luring back to a killhole, you would make noise and lure enemies to the edge of their patrolling distance. It was often *easier* because you could fight only a few enemies at a time despite making tons of noise at your location.

Make enemies follow you if they're within LOS; otherwise, they stop pursuit and immediately return to their area.

It would be tricky to balance, but should eliminate the worst of luring while still preserving the tactical options of retreat/avoidance and maneuvering to fight one-on-one. Other options suggested, like making all monsters move faster than the player or forbidding the player from retreating in some way without burning consumables, eliminate tactical options.

[svendre]

Here's a solution for players that don't want the temptation of the notion that luring is always the optimal strategy:

Make a new race:
** Tree **
(bonus ac and resists)
cannot move
once in awhile determined by RNG, the Tree can random teleport or random blink

O.k. but, seriously... I really don't in fact believe that luring is always the optimal strategy. I'll give an example: You're a low level mummy trained for melee. You spot an orc priest, and you have no immediate way to get out of LOS. So, if you attempt to "lure" him back somewhere, along the way it may decide to smite you one or more times, possibly killing you as you try to back-pedal to a corner, stairs or whatever position could reduce the time he is looking at you from rage. On the other hand, if you charge it you 1) can begin applying damage to them sooner, possibly eliminating the threat completely and 2) you decrease their odds of using smite since being at melee range becomes a new choice for them to swing at you. That example is also like trying to lure a torment or hellfire monster.. good luck with that. If you're melee and you don't close the gap asap, it is likely going to hurt, badly.. to run him somewhere.

If, every single monster could be attracted to you from a position where it'll enter your LOS from maximum visibility distance *and* there was always an ideal tactical position to drag it to where they will be one-on-one in melee range immediately *and* there weren't characters which were melee only.....then I might agree that it could always be optimal to lure.

Oh yeah and I noticed the nifty double orc priest thing (not sure how new that feature is really though.) Your baby melee mummy rounds the corner and sees two orc priests several spaces away, behind four other orcs...... you better not lure, or run to the stairs... b/c those priests should surely get a speed boost.. and be able to smite and run at the same time! Charge!!!! Yeah! Luring problem solved for that guy... there won't be any luring, b/c he'll be dead If one believes any certain choice shouldn't always be obvious or it isn't interesting.. then how does one account for a situation like that, where if you don't head for a staircase real fast, you will almost certainly always have a slim chance to survive.

[sanka]

O.k. but, seriously... I really don't in fact believe that luring is always the optimal strategy. I'll give an example: You're a low level mummy trained for melee. You spot an orc priest, and you have no immediate way to get out of LOS. So, if you attempt to "lure" him back somewhere, along the way it may decide to smite you one or more times, possibly killing you as you try to back-pedal to a corner, stairs or whatever position could reduce the time he is looking at you from rage. On the other hand, if you charge it you 1) can begin applying damage to them sooner, possibly eliminating the threat completely and 2) you decrease their odds of using smite since being at melee range becomes a new choice for them to swing at you. That example is also like trying to lure a torment or hellfire monster.. good luck with that. If you're melee and you don't close the gap asap, it is likely going to hurt, badly.. to run him somewhere.

This is actually a very bad example. Note that if the orc priest smites you it will loose a turn chasing you, and there's a good chance that it will cast some cantrip, so sooner or later it will be at the edge of your LOS, and you can safely go back to the stairs and escape. If it's six squares away, than it can at most smite you two times.

On the other hand if you rush towards it, it will have multiple turns to smite you while you are running and when you are trying to kill him, possible missing with your attacks, etc - not talking about the second orc priest that you will find just behind the first one usually.

[Quazifuji]

I feel like this thread's gotten hard to follow, so I'm going to attempt to summarize my understanding of the approaches to dealing with luring that having been proposed. If I'm right, this post will be a reference for people who have gotten kind of lost, and maybe help focus discussion. If I'm wrong, someone can correct me and then I can learn from it.

It seems to me solutions to deal with luring fall into a number of main categories:

1. Make monsters do bad things while following you. Generally, this consists of either making monsters faster (or giving them a swiftness-type ability) so that they damage you while following you, or making them make lots of noise while following you.
   
2. Explicitly punish luring through something other than monster traits/behavior, e.g. wake or spawn monsters when a monster is chasing the player.
   
3. Increase reward for low turn count/punishment for high turncount in order to indirectly discourage luring.
   
4. Change monsters AI to make luring impractical/impossible (e.g. limit how far monsters will follow you, possibly including monsters setting up ambushes). Giving all monsters a swiftness ability they use while chasing can also somewhat fall into this category, since it limits how far monsters will chase you too, although not through AI means.

Personally, I vastly, vastly prefer something from categories 1 or 4. Unless I'm misunderstanding category 2, it would be one of the most spoilery mechanics in all of Crawl. Category 3 has the potential to be spoilery as well, but more importantly, I like Crawl's current state where turncount is mostly a matter of score and rarely relevant to winning. I also think enforcing faster turncounts will cause some undesirable balance problems, and affect resting much more than it would affect luring.

I think option 4 could be very effective, but is difficult to do well. As others have pointed out, simple solutions (e.g. giving monsters a "leash" to limit how far they'll follow you) are often very abusable, and complex solutions are... well, complex.

So personally, I'm in favor of category 1. My personal favorite solutions I've seen are either making monsters continue to shout occasionally while chasing the player, and/or giving all monsters a swiftness ability they use when chasing the player for long enough (for walking monsters, this could easily be flavored as sprinting, and I'm sure we could come up with a more generic flavor text representing a monster exerting itself for a burst of speed followed by slowness for other monsters). I don't know if these ideas will solve the problem (and they certainly do nothing to help stair dancing), but I think they'd help make luring feel a bit less safe. They also have two other very nice benefits, in my opinion. First, they don't punish small amounts of tactical repositioning at all, only longer chases. Second, I think both would be very intuitive and easy to convey to an unspoiled player through text, which I think is something extremely important that hasn't been discussed enough in this thread. It could also be possible to combine the two mechanics if both helped but neither one on its own punished luring enough.

I'd be very interested in an experimental branch with either or both of these mechanics to see how they affect things.

[zxc23]

Any swiftness-based 'solution' will massively increase unavoidable deaths in the early game due to the inability to pillar dance or even to reach staircases.

[ontoclasm]

Any swiftness-based 'solution' will massively increase unavoidable deaths in the early game due to the inability to pillar dance or even to reach staircases.

"It will change balance" is a terrible reason not to do something. Every meaningful change affects balance. If removing or changing a stupid thing nerfs (or buffs) players too much, it can be compensated for in non-stupid ways (e.g. more starting HP). You're essentially arguing that we shouldn't repair that car we have on cinder blocks out on our lawn because then we'll have to weed the garden it was in front of.

[zxc23]

Any swiftness-based 'solution' will massively increase unavoidable deaths in the early game due to the inability to pillar dance or even to reach staircases.

"It will change balance" is a terrible reason not to do something. Every meaningful change affects balance. If removing or changing a stupid thing nerfs (or buffs) players too much, it can be compensated for in non-stupid ways (e.g. more starting HP). You're essentially arguing that we shouldn't repair that car we have on cinder blocks out on our lawn because then we'll have to weed the garden it was in front of.

I haven't seen mention of compensating changes in this thread (by anyone advocating large changes like swiftness to chasing monsters). It looks as though balance is largely being disregarded, which is why I'm raising these points.

I don't really like the idea of swift ogres. Extra HP wouldn't matter in this scenario. If the player started with a few consumables, like fear, teleport, curing, then that could be workable.

[duvessa]

Balance is usually disregarded when crawl is changed. It doesn't really cause problems because crawl was never well-balanced in the first place.

[radinms]

zxc23,

i guess devs think you broke the game balance by using luring tactics and did 31-streaks

[bananaken]

"It will change balance" is a terrible reason not to do something. Every meaningful change affects balance. If removing or changing a stupid thing nerfs (or buffs) players too much, it can be compensated for in non-stupid ways (e.g. more starting HP).

I agree that "it will change balance" shouldn't be a point against it, but swift ogre goes well beyond that I think, it just sounds completely busted.

[daggaz]

As you get into the dungeon, load more and more of the consumables in monster hands rather than on the floor. If a monster notices you, it will use the consumable if it cannot engage you after X number of turns. This repeats for the same monster if you break away from the fight.

Or something like that. Now you have an incentive to quickly engage and kill opponents, and you might even get the number of consumable to go down a bit. Especially if monsters get bored and use their consumables if they dont see the player X^X number of turns after the floor has been entered.

[Axle F]

This discussion gets to the heart of the game and motivated me to make an account. Having read all the comments, it seems to me there are two mutually contradictory goals:

1) Discourage excessive luring

2) Avoid helping powerful characters while hurting weaker characters (for example, by having more timed portals)

I think there is a way to sidestep the contradiction. It seems what we'd really like to prevent is not the viability of luring at certain times, but the motivation toward it as a general strategy. The motivation mainly comes from it being a way to massively improve survivability in general through large stretches of the game. If the player *must* face situations where they cannot lure, such as unavoidable portals with small zig-like maps, they will die if they have been relying too much on luring. Moreover, if they can survive the portals most of the motivation to lure will be gone, because they are able to take on groups.

Now of course players can still burn consumables to get past the portals, and use luring to economize on consumables, but this at least seems to make a dent in the problem. Luring would be used more situationally and you'd have to be even more patient to make it worth bothering.

While this indeed hurts weaker characters, it does it in a more Crawl-like way, by killing off the unviable early rather than gradually starving the habitual lurer of loot like a timed-portal approach would do.

[ontoclasm]

This discussion gets to the heart of the game and motivated me to make an account. Having read all the comments, it seems to me there are two mutually contradictory goals:

1) Discourage excessive luring

2) Avoid helping powerful characters while hurting weaker characters (for example, by having more timed portals)

I think there is a way to sidestep the contradiction. It seems what we'd really like to prevent is not the viability of luring at certain times, but the motivation toward it as a general strategy. The motivation mainly comes from it being a way to massively improve survivability in general through large stretches of the game. If the player *must* face situations where they cannot lure, such as unavoidable portals with small zig-like maps, they will die if they have been relying too much on luring. Moreover, if they can survive the portals most of the motivation to lure will be gone, because they are able to take on groups.

Now of course players can still burn consumables to get past the portals, and use luring to economize on consumables, but this at least seems to make a dent in the problem. Luring would be used more situationally and you'd have to be even more patient to make it worth bothering.

While this indeed hurts weaker characters, it does it in a more Crawl-like way, by killing off the unviable early rather than gradually starving the habitual lurer of loot like a timed-portal approach would do.

I'm confused; what's your actual suggestion?

[HardboiledGargoyle]

I think the gist is: make the reward of timed portals play such a big role in your power level that without completing them, you’ll basically just die however you play (or, “you failed to reach the portal. you die…”) and with completing them, you’ll be so powerful that you can steamroll your way to further portals and not even care about luring.

[lethediver]

Idea: you get more exp if you defeat monsters in rapid succession.

[Axle F]

Actual proposal: player gets whisked into a small-map portal randomly when they enter a down-staircase on certain dungeon levels and must fight their way to the exit. Or maybe something like banishment: orc priests can invoke Beogh to "invite" you to a bailey.

Actually simpler than this would be to have more enemies capable of the vault warden's sealing rune ability, or perhaps trap doors that slam shut behind you after you pass through them and you must reach a switch on the other side of a group of monsters to leave the room.

I just played another game with luring in mind, as an OpMo. I realized I used luring (not excessive luring, but going up to one screen away) as a way to adjust the difficulty to where I wanted it. As an Octopode melee character the early game is quite hard and without at least some minor luring it becomes ridiculous. When I got a bit stronger I would lure less to up the challenge.

This seems like a fundamental conundrum for a game with randomly generated maps that aren't tailored to the class/race at all. There is no way for the design to ensure that difficulty is always in any kind of sweet spot like with a pre-planned map tailored to the character. The fun parts of the game are when the difficulty does find those sweet spots, but inevitably part of the responsibility for adjusting the difficulty falls on the player: when RNG gives you something insanely difficult other otherwise impossible, it might almost be expected for the player to use luring and other "cheap" tactics to smooth out the temporary difficulty spike. It's not so much of a problem for most players because the tedium of doing it all the time would turn most players off. Of course there will always be some for whom winning means so much that they are willing to endure any amount of tedium to keep alive an unviable character.

This can either be looked as "there can be no accounting for odd players" or if things like endless streaking are to be addressed, include unavoidable portals/trapdoors/sealing runes to make tedious tricks even less useful as a general strategy (rather than a situational tactic).

[Aethrus]

I want to add my two cents here and say that to me this conversation seems to be mixing up two very different concepts: luring and retreating.

Luring, to me, is when there are enemies that are grouped together, designed or placed by chance to be one encounter as a whole together, but the player separates them from each other and handles them independently, either by drawing one's attention and pulling back, or more commonly (or easily) by stairdancing. Finding an lone enemy and pulling it back to a nearby hallway or whatever is essentially the same thing as this strategy, just being performed preemptively and ritualistically. I'm fine with this being removed.

Retreating however is seeing an enemy that you can't fight, don't want to fight, or aren't currently prepared to fight / in a favorable position to fight, saying 'nope' and walking away. Not pulling part of its group upstairs with you, but totally withdrawing, perhaps coming back to deal with them later, perhaps putting a big old exclusion over where they were and leaving them be as long as possible.

Retreating is a fundamental tactic in Crawl and should not be punished in the same fashion as luring. When I first started to learn how to play Crawl, one of the first things I was told was that you don't have to kill every monster you see, that the most powerful way to defeat an enemy was, in a way, to simply walk away from them. Obviously, not all enemies can be simply walked away from, but the principle should apply. Universal swiftness mechanics, speed boosts or other penalties for any sort of 'tactical retreat' would undermine this facet of gameplay, not to mention how they'd make harder races and backgrounds even more difficult. A TrBe or a MiFi might be able to handle gameplay where they can't exit a fight without a consumable, but not a DEFE or what have you! Crawl is a game where if you die, you know why you died and how it was your fault. I like that element of the design, and punishing players for leaving fights runs contrary to that principle! If I see an Ogre that could one-shot me if I get into melee with it, run away, and it calls for friends / speeds up and kills me anyways, I didn't really have a chance, especially early on. I guess I should have played an Assassin and curare'd or been a Berserker for extra HP.

To me, the best solution would be to have groups of enemies spawned together stick together to some extent, changing AI so they can attack in groups, and perhaps most importantly, changing stair behavior so if an enemy follows you up stairs, its allies can follow you up later as well after a turn or two, making the most powerful luring tool less potent.

Might it still be most efficient to draw enemies back away from unexplored territory when encountering them? Sure, but I don't see how that's a problem. Making enemies walk towards you is roguelike tactics 101 and we shouldn't make it a goal to punish every possible practice considered to be optimal or 'tedious' by some. It's a turn-based game, of course there are going to be optimal strategies, and of course using them should be more effective. If the game was designed so a very experienced player using a powerful race/background/god combo still had trouble winning it would make weaker but interesting combos unplayable.

[partial]

I want to add my two cents here and say that to me this conversation seems to be mixing up two very different concepts: luring and retreating.

Retreating is a fundamental tactic in Crawl and should not be punished in the same fashion as luring. y

this

[HardboiledGargoyle]

I agree with most of Aethrus's post, but I would argue that having big groups of monsters follow you across stairs breaks the use of retreating as a fundamental tactic. You want the game to let you use stairs to retreat, but also to swarm you with enemies if you use stairs to lure (your terminology). Is the player ever allowed to disengage from a fight while it's happening, or must the decision always be made before the fight starts? It's as if the game is supposed to read minds - whether the player intends to lure or retreat - and respond accordingly. I would have glossed over this but you highlight it as "perhaps most importantly".

[Aethrus]

I agree with most of Aethrus's post, but I would argue that having big groups of monsters follow you across stairs breaks the use of retreating as a fundamental tactic. You want the game to let you use stairs to retreat, but also to swarm you with enemies if you use stairs to lure (your terminology). Is the player ever allowed to disengage from a fight while it's happening, or must the decision always be made before the fight starts? It's as if the game is supposed to read minds - whether the player intends to lure or retreat - and respond accordingly. I would have glossed over this but you highlight it as "perhaps most importantly".

Fair point, I'm not really putting forward any specific suggestion, rather, I don't have a particular idea in mind I want implemented. I just wanted to make my statement on what I felt should be the priority.

That said, the idea I had in mind would be that going up the stairs would be totally safe if no one came up with you immediately, the same as it is now. The difference would be that if you pulled adjacent enemies with you, others would follow. Still not perfect if you step downstairs next to a nasty group, but that's what I was suggesting.