mimics

[XuaXua]

The "optimal" argument gets used so much. Not every single person is interested in playing an "optimal" game.

This.

If someone is has bad enough OCD that they're going to throw rocks at everything (you know exactly the sort of player I'm talking about), THEN LET THEM.
Not everyone has to do it, but honestly, much like 'v' was evokable on ranged weaponry at distances, the same sort of key could be used to force an attack.
That's about as best a suggestion as I can make to resolve it.
Actually, "revert the patch that stops us from attacking mimics before we know they're mimics" is probably the solution there.

When you start getting more and more OCD players, then it's... well, that's Crawl.

EDIT OMG thanked minmay.

[danharaj]

Hey can we stop trivializing a serious mental disorder by using it to describe meticulous styles of playing video games?

[dpeg]

Where danharaj is absolutely right: Choose your words (abbreviations rather) more carefully.

Where minmay is not right: Walking up to features is actually better than shooting at features. (1) Fewer different keys to be pressed. (2) The mimic can attack. (3) The mimic exposure walk needs longer (yes, food does not matter enough) but also exposes the player to more squares (i.e. potential monsters).

Where minmay is right: walking up every feature (silly as it is) encourages to fully rest all the time, so as to minimise the mimic risk. However, it is unfair to put all the blame for that on the poor mimics -- that problem is rooted much deeper in the game.

Where everyone but galehar is not right (all IMO etc.): We have seen players do incredible things to "maximise" their performance (often with activities detrimental to the objective). If it can be done, it will be done. And not necessarily by marginal numbers of players. Do we have to repeat the shop selling stories from Nethack? Yes, it's not a very clever tactice. Yes, newbies fell for it, again and again. And yes, there are cases where it is optimal to sell dungeon rubbish. Don't assume that Crawl players are not prone to this.

I like it when dungeon features flee from my @ -- thanks Xom! I love it when dungeon features bite me -- thanks galehar! It's a fun little piece of Crawl humour, with occasional tactical implications, and the current installment is quite fine, in my opinion.

[Blade]

Mimics were a "fun little piece of Crawl humour" when they were absurdly bad at imitating items and not overly threatening.

Unfortunately, that stage has passed. Now, they're an incredibly obnoxious, often unexpectedly dangerous enemy type that has the potential to turn things nasty when the player should be safe.

As mimics are currently implemented, they are bad and absolutely not humorous (unless you find it funny to be running away from a dangerous enemy only to be constricted by a statue in explored territory).

[eeviac]

Maybe mimics wouldn't be so infuriating if they didn't encourage fully resting after every fight AND discourage autoexploration AND weren't guaranteed the first attack AND didn't have high damage attack/poison/constriction AND didn't resist nearly every spell school.

[Deimos]

Why don't we tie in mimics with Traps & Doors?

Treat them like hard-to-detect traps. Mimics cannot be damaged until they are detected, and they cannot be detected by simply waiting next to every feature: when they enter LoS, they have a single roll that decides whether they will be detected when in range of T&D. The chance to fail this roll scales down with T&D. T&D now will detect them at a range that scales with your skill level.

If you do happen to detect a mimic, it is because you had enough of the appropriate skill level. You are rewarded with the ability to kill it from afar. I think this has two good sides: fixing most of the annoyance with mimics (except for their extreme combat prowess) and giving another reason to level up T&D, which is lacking right now.

[KoboldLord]

Mimics were a "fun little piece of Crawl humour" when they were absurdly bad at imitating items and not overly threatening.

Unfortunately, that stage has passed. Now, they're an incredibly obnoxious, often unexpectedly dangerous enemy type that has the potential to turn things nasty when the player should be safe.

As mimics are currently implemented, they are bad and absolutely not humorous (unless you find it funny to be running away from a dangerous enemy only to be constricted by a statue in explored territory).

I'm not sure what you intended with this statement, but the thing I'm hearing is "Mimics are now totally awesome features in the game, and the revamp was brilliantly successful." Getting ambushed while running from a dangerous enemy is Fun.

[pratamawirya]

Yeah, I kind of like the new mimics.

Why don't we tie in mimics with Traps & Doors?

Treat them like hard-to-detect traps. Mimics cannot be damaged until they are detected, and they cannot be detected by simply waiting next to every feature: when they enter LoS, they have a single roll that decides whether they will be detected when in range of T&D. The chance to fail this roll scales down with T&D. T&D now will detect them at a range that scales with your skill level.

If you do happen to detect a mimic, it is because you had enough of the appropriate skill level. You are rewarded with the ability to kill it from afar. I think this has two good sides: fixing most of the annoyance with mimics (except for their extreme combat prowess) and giving another reason to level up T&D, which is lacking right now.

This is a great idea.

[eeviac]

How about we go the other direction instead - remove their hiding completely. They just attack on sight. I'll admit there's a certain charm to getting chased by doors and gold piles; it's their ambush mechanic that makes this all so aggravating. I'm a proponent for getting all damage traps removed from the game, and right now mimics are much worse than any blade trap.

[pratamawirya]

How about we go the other direction instead - remove their hiding completely. They just attack on sight.

What will be the purpose of them being mimics, then?

[crate]

Mimics were a "fun little piece of Crawl humour" when they were absurdly bad at imitating items and not overly threatening.

Unfortunately, that stage has passed. Now, they're an incredibly obnoxious, often unexpectedly dangerous enemy type that has the potential to turn things nasty when the player should be safe.

As mimics are currently implemented, they are bad and absolutely not humorous (unless you find it funny to be running away from a dangerous enemy only to be constricted by a statue in explored territory).

I'm not sure what you intended with this statement, but the thing I'm hearing is "Mimics are now totally awesome features in the game, and the revamp was brilliantly successful." Getting ambushed while running from a dangerous enemy is Fun.

I'm not sure why you're running from a dangerous enemy into unexplored territory. Or if you get ambushed by a mimic while running into already-explored territory, then you just died to not being thorough enough in your mimic-checking!!

[njvack]

As a slight aside: How hard would it be to write a lua script that automatically placed 3-tile exclusions on mimic-able features and items? (Yes, I realize this would be unbearably tedious and need serious UI tweaking in some areas. Consider it a thought experiment.)

[One-Eyed Jack]

You could have it so that feature mimics only turn evil when your tension level is high enough (or when them turning would raise it to a certain level.) Otherwise they just act normally (you can walk through and drink fountain mimics, open/close door mimics, etc.) So say you are trying to lure a yak pack back through the door but THE DOOR IS NOW A MIMIC

Pro: Increase mimic lethality (and "fun")
Encourage paranoia
Approach/throw tactics don't work

Maybe with items too: You finish putting on the glowing gold dragon armour. A kobold comes into view. The glowing gold dragon armour was actually a mimic! The mimic constricts you. The mimic hits you. You feel poisoned.

[eeviac]

How about we go the other direction instead - remove their hiding completely. They just attack on sight.

What will be the purpose of them being mimics, then?

If they attack on sight, it's still some stairs or a door trying to kill you, i.e. funny. It also promotes the 'everything in this dungeon wants to murder you' aesthetic. This way we could also add wall mimics!

As a slight aside: How hard would it be to write a lua script that automatically placed 3-tile exclusions on mimic-able features and items? (Yes, I realize this would be unbearably tedious and need serious UI tweaking in some areas. Consider it a thought experiment.)

This already happens when worshipping ash and you spot a mimic. I imagine it could be extended to all dungeon features.

[Blade]

Mimics were a "fun little piece of Crawl humour" when they were absurdly bad at imitating items and not overly threatening.

Unfortunately, that stage has passed. Now, they're an incredibly obnoxious, often unexpectedly dangerous enemy type that has the potential to turn things nasty when the player should be safe.

As mimics are currently implemented, they are bad and absolutely not humorous (unless you find it funny to be running away from a dangerous enemy only to be constricted by a statue in explored territory).

I'm not sure what you intended with this statement, but the thing I'm hearing is "Mimics are now totally awesome features in the game, and the revamp was brilliantly successful." Getting ambushed while running from a dangerous enemy is Fun.

Did you miss the part about "explored territory?" I'm not interested in going around every single statue in a place with 20 of them to make sure they -all- aren't mimics, and if I find out that one of them is a mimic after going back to a place that looks safe for all intents and purposes, I will not think of it as fun. I will think of it as the game punishing me for not obsessively checking everything, and that is a Very Bad Thing.

[njvack]

Alternately, would it be possible to script such that autoexplore would visit mimic candidates?

[bobross419]

How about we remove the stupid thing instead of changing everything else to accomodate it. You could make autoexplore sell your useless items to shops but that doesn't make it any better of an idea.

This or T&D is better than my idea.

[njvack]

How about we remove the stupid thing instead of changing everything else to accomodate it.

I meant "is it possible" not in the sense of "would the devs please do this" but rather "are there technical reasons why I could not do this" -- literally, is it possible?

[XuaXua]

That Traps and Doors thing is GREAT.

I think a mimic detected with Traps and Doors should appear onscreen as a "sleeping" mimic (has the zzz status in tiles), which then qualifies it for being attacked (and wake it only if the player moves adjacent or it is otherwise interacted with in some fashion, per standard "sleeping" rules).

[XuaXua]

Did you miss the part about "explored territory?" I'm not interested in going around every single statue in a place with 20 of them to make sure they -all- aren't mimics, and if I find out that one of them is a mimic after going back to a place that looks safe for all intents and purposes, I will not think of it as fun. I will think of it as the game punishing me for not obsessively checking everything, and that is a Very Bad Thing.

Just explored all of Vaults:4 and made my way to the statue-surrounded crypt entrance.

One statue was a mimic! Run!!!!! (and kill).

That actually was fun and a good use of a mimic. Traps & Doors detection? Excellent excellent excellent idea; Deimos gets all my stars.

[galehar]

I admit that I overdid it when I introduced depth scaling. The scaling will be changed by normal tiers, and they'll probably be nerfed a bit along the way. inept and normal mimics will have a hold attack (damage-less constriction), ravenous ones a weak constriction (about half damage). Inept don't have poison attack neither any resist.

Autoexplore could check features, but then why not items?

I also like the T&D detection idea very much, but only if it's a single roll. We don't want players waiting near every feature. Maybe we could use the same system for traps btw, however it wouldn't work well with secret doors.

[Galefury]

I will repeat what was said on IRC: just remove secret doors. There are very few interesting uses of them.

[njvack]

Autoexplore could check features, but then why not items?

Yeah -- if I get around to lua scripting this, that's probably what it'd do -- visit all features and items that are potentially mimics, probably inscribe items as {checked} and maybe drop an inscribed stone next to features; I don't think there's another way to mark dungeon features?

[pratamawirya]

Okay, for some reason, some players don't like mimics because they reward thorough-checking of every feature, and yet some other players (me included) find it very fun to get ambushed by mimics for not checking thoroughly every feature.

You can't blame us for liking the current mimics. You can't blame us for thinking that they are fun. So, now thorough-checking is rewarded greatly, huh? I don't care. It is FUN. But... you care and you think it's not fun at all. Okay.

How could the players get divided like this? lol

[danharaj]

I also like the T&D detection idea very much, but only if it's a single roll. We don't want players waiting near every feature. Maybe we could use the same system for traps btw, however it wouldn't work well with secret doors.

Detected mimics will probably be just as interesting if not more interesting than undetected mimics. A detected mimic controls up to 8 squares. Crawl is a tactical game so that control can be very significant. The player must make a decision on when to kill the mimic, if they can at all. Part of the problem with current mimics is that there is no interesting decision to be made. The only decision, to check every feature, is terribly tedious and it doesn't ever really feel rewarding to kill a mimic anyway.

As for the T&D detection idea, I think it would be reasonable to just auto-detect mimics with HD less than or equal to your T&D skill.

[njvack]

I'm actually kinda in the "being ambushed by a door is fun" camp.

But it's even more fun to think "OK, if I was going to really exhaustively check for mimics within Crawl's mechanics, how could I code it?" It's the most meta of meta-games -- the hack is its own reward.

And it'd be funny to have a game in which I'd picked up and inscribed every single item.

[Arkhan]

I've never actually died to a mimic in .10. I've come close by holding a movement button down to grab a heal wounds pot on a caster, but I've never actually died to one. While them becoming a threat is a good thing, I'm not a fan of what sounds like an incredibly obnoxious mechanic. I mean I wouldn't mind if a statue grabbed me in explored territory when I wasn't in danger, but not as a game-ender in a place that should be safe. I like the T+D idea, and I'm not entirely sure why this isn't the case already.

I think we should just focus on the fun aspects of them. There's a few ways to have more fun with mimics- like why not have them imitate features and objects as planned, but make them visually different from the same item and activatable from a distance. Like if you get a Heal Wounds potion mimic, make it a different color than what Heal Wounds is in your current game. That way you're crawling through the dungeon and you see a heal wounds potion that shouldn't be oran- heeeey. Or maybe an up staircase with just two of the steps melded into a ramp shape instead, subtle changes like that would encourage the idea that the mimics aren't perfect masters of disguise but will still catch inattentive players, while not harming good ones (because they'll notice things like an elephant statue without a trunk.)

I mean, I realize this would require a tileset for everything it could apply to (except scrolls and potions) but it could be an incremental side project over time or just limited to those easy ones like potions to begin with.

Edit:
If you want a particularly powerful mimic, make it produce a warning line or something once you're within Field of Vision + 1 range of it or something. The player doesn't have to be careful UNTIL they hear this now, whereupon they now check features carefully for something that might seem off. A more powerful mimic announced by a warning line might have only it's description be a hint, like an Oka altar that isn't sharp, a door with no visible means of opening it, a wall that snickers with laughter as you get closer to it.

Xom could make this even more fun by sometimes turning it into a game of hot and cold. As you get closer he's like "Warmer. Waaarmer. Getting hot, mortal!"

[Kate]

Like if you get a Heal Wounds potion mimic, make it a different color than what Heal Wounds is in your current game. That way you're crawling through the dungeon and you see a heal wounds potion that shouldn't be oran- heeeey. Or maybe an up staircase with just two of the steps melded into a ramp shape instead, subtle changes like that would encourage the idea that the mimics aren't perfect masters of disguise but will still catch inattentive players, while not harming good ones (because they'll notice things like an elephant statue without a trunk.)

That's pure interface screw, and exactly the kind of thing that Crawl explicitly tries to avoid. (Also wouldn't work in console, not that any other reason is needed).

[Arkhan]

Like if you get a Heal Wounds potion mimic, make it a different color than what Heal Wounds is in your current game. That way you're crawling through the dungeon and you see a heal wounds potion that shouldn't be oran- heeeey. Or maybe an up staircase with just two of the steps melded into a ramp shape instead, subtle changes like that would encourage the idea that the mimics aren't perfect masters of disguise but will still catch inattentive players, while not harming good ones (because they'll notice things like an elephant statue without a trunk.)

That's pure interface screw, and exactly the kind of thing that Crawl explicitly tries to avoid. (Also wouldn't work in console, not that any other reason is needed).

Ah, you got a point there. Visual differences don't really work in text. But still, I mean if they're kept at the .10 power level it's not really that much of a screw because they pose no actual threat but I guess if you're going to make them relevant...

It's odd because one of the design goals is clarity, and this is a monster that's inherently about deception. Too little and they don't do justice to the idea. Too much and it's unfair and breaks the goal. In a game whose goal is about clarity of information if you're going to trick the player then they need to know it's happening (Like the Misled status is there to confer that you are being deceived about what you're seeing, so the player knows they're being tricked but are still receiving clear information.) If you did something similar with a Mimic (a distance warning or maybe a Dread status when one is nearby) would that work? That way they 100% know that something that they're looking at is attempting to deceive them- so it's not the fault of the interface anymore, and now it's up to the player to either throw rocks at everything or deduce that a scroll named IAMMIMIC is in fact a mimic.

[danharaj]

Hi, just had another completely asinine experience with feature mimics. It was soooo fun fighting an enemy that is resistant to everything and will ruin your day even if you're not making any technical errors before it interrupts your auto-explore and forces you to use consumables or die. I absolutely love enemies that leave no counterplay and punish you for doing absolutely nothing wrong. Look. There are plenty of interesting and challenging enemies in Crawl at all depths. None of them have to rely on absurdly powerful constriction, free attacks, omni-resistance, and hideously skewed depth scaling all at once. When you have to pile all that on a monster and also make it completely invisible to the player (monsters even treat door mimics like they're actual doors! Unbelievable!), there is a fundamental flaw that can't be made 'interesting'. There is nothing fun about mimics.

I am boycotting the August tournament if feature mimics are left the way they are on 0.11 release.

[evilmike]

I am boycotting the August tournament if feature mimics are left the way they are on 0.11 release.

I don't like the current state of mimics and I'd be quite upset if they made it into 0.11 as they currently are. I think I agree with your opinions on them (minus the hyperbole). I'm also looking forward to the upcoming changes that galehar has talked about.

That said, "fix this or I will boycott the next version" threats are ineffectual. This sub-forum is for game design discussion. Focus on talking about the flaws of mimics and how to address the problems. Feel free to express discontent or frustration. But if you want to be taken seriously, do not make spineless threats about refusing to play a free video game.

[KoboldLord]

Eh, my current character got hit by a volcano portal mimic, right after dumping my scrolls in the tunnel leading up to it. The character in question was a steam draconian venom mage with no relevant wands and no useful spellbook drops so far, so I could do literally nothing against it. I ran away until I found a rat and got it to follow me too, went around the corner so the rat got in the mimic's way, and then went back up the stairwell. No problem. I came back down another stairwell to continue exploring the rest of the level and retrieve my scrolls. I didn't run into the mimic again at this point, but my plan was to briskly walk to the stairs again. On the way back from Lair 8 I had something that could kill it, so I did.

At the time, I was mostly glad it wasn't something actually dangerous, like a spiny frog. Killing those things with nothing but steam breath kind of sucks.

[BlackSheep]

I actually like that mimics don't immediately disappear if I use a ranged attack on them now.

[galehar]

I admit that I overdid it when I introduced depth scaling. The scaling will be changed by normal tiers, and they'll probably be nerfed a bit along the way. inept and normal mimics will have a hold attack (damage-less constriction), ravenous ones a weak constriction (about half damage). Inept don't have poison attack neither any resist.

Those changes are now in trunk with also a bunch of adjustments to constriction:
* Regarding to_hit, it's a normal attack (affected by +Acc, shroud of golubria, can be blocked with a shield,...)
* implementation is saner
* constriction damage can be adjusted per monster instead of depending on HD
* blink doesn't work anymore, but it counts for 2 escape attempts (teleport still works just fine)
* damage scaling over time is changed: dam += dam * stepdown(duration, 5) / 5. Means damage doubles after 5 turns and then again after 10 more.
* Tentacled monstrosity and greater naga's constriction is reduced a bit, otherwise the new formula makes it crazy. For them, it's a bit slower to build up. For weaker constrictors, it no longer increase to high damage after a long time.

Feedback is very much welcome, especially if it's backed up by testing.

[KoboldLord]

That sounds great. I'll fire up a naga, and hopefully I'll be able to play it without the same sense of shame I felt on my last one.

[dd]

Wait, you mean I can't blink anymore to escape constriction?

I don't know... it's going to be tough when you get that one snake level that's an open area where all the nagas gang up on you the moment you walk up stairs... And tentacle monsters were tought to kill even with blinking, now they're going to be brutal.

I'm going to have to test this to be sure but I kind of fear this will be problematic for some characters, especially small races...

[galehar]

Wait, you mean I can't blink anymore to escape constriction?

Sure you can. It counts as 2 escape attempts, so you're likely to escape with 2 blinks, 3 at worst. For 2 MP, it's still efficient. It's just not guaranteed anymore.

And tentacle monsters were tought to kill even with blinking, now they're going to be brutal.

You don't want to melee those guys. Ever.

I'm going to have to test this to be sure but I kind of fear this will be problematic for some characters, especially small races...

On the other hand, small races will have a much easier time dodging those constriction attacks. Once grabbed though...

[danharaj]

I just played a whole game with the mimic changes except the blink tweak. All in all, mimics are now neutral on the fun-scale unless you bump into them while in danger. That's probably a decent enough spot for them to be for 0.11. I can see positive experiences with them occurring, and most of the negative experiences seem to be toned down to something more benign. I'm sure vault creators will be come up with more vaults that use them effectively. One suggestion I'd like to make is to add a way for vault creators to mark features as unmimicable (or perhaps make that the default). Feature mimics might have undesired interactions with some vaults that designers should be able to account for now.

In general I am not a fan of constriction's increasing damage. It makes meleeing creatures with constriction much less reliable, with damage received wildly swinging. So much, in fact, that if you get a particularly bad streak of misses on one, you can easily end up taking twice or even more times as many damage as usual. I once died to a naga warrior after killing over a dozen without any problems because of a bad set of rolls. Monsters being consistent threats, IMO is a big deal. I feel like constriction would be better if it didn't deal any damage but also kept you in place longer, with high STR making it easier to break out. I feel like constriction really doesn't matter for high damage, high ac, high hp characters, and strongly punishes low hp, high ev characters.

I think constriction deserves its own thread because it affects multiple areas of the game.

[evilmike]

One suggestion I'd like to make is to add a way for vault creators to mark features as unmimicable

Already exists. It isn't used much (because it isn't needed much), but it's there.

As a side note, if you have trouble with constriction, I suggest trying to break free by moving away. I think players underuse this right now, because the original implementation of constriction made it almost impossible to succeed doing this. Moving away is actually a reasonable tactic now, and should succeed after a few attempts for normal characters. For best results, try to do this before damage ramps up to insane levels.

[dpeg]

There is actually a lot of variance in damage, I'd say (consider spells). What's wrong with having variance also for constriction damage? It is certainly much nicer to have a bad series of constriction rolls than to have a single bad iron shot roll -- you can react to the former. I would expect players to develop a feeling for what's a "good" and a "bad" constriction. Since we don't want players to fill tables with numbers, the game could say that: "The naga warrior constricts you some more.", "The naga warrior constricts you some more!!", "The naga warrior severly constricty you. OUCH!"

About who's punished: wasn't there concensus that mages always get the better deal than meleeists?

[danharaj]

As a side note, if you have trouble with constriction, I suggest trying to break free by moving away. I think players underuse this right now, because the original implementation of constriction made it almost impossible to succeed doing this. Moving away is actually a reasonable tactic now, and should succeed after a few attempts for normal characters. For best results, try to do this before damage ramps up to insane levels.

As I play more with constriction my intuition for how dangerous it is gets sharper. It's just at first it was hard to intuit when constriction was getting really dangerous. When to escape is a very delicate decision to make. Sometimes you might be better off trying to kill the thing because escape would take longer than killing it, and those last few extra turns can easily kill you.

I will need to play with this new iteration of constriction to see how it feels to give useful feedback. One idea I think could also work is making it so constriction damage doesn't increase if you try to escape. Making the player do nothing for a turn is quite strong already! Effectively paralyzing the player in melee is a very good effect. I feel like that is the aspect of constriction that is most valuable.

There is actually a lot of variance in damage, I'd say (consider spells). What's wrong with having variance also for constriction damage?

About who's punished: wasn't there concensus that mages always get the better deal than meleeists?

There are more ways to mitigate dangerous spellcasters than there are ways for melee characters to deal with constrictors. Some melee types are fragile but avoid getting hit. A recent few spriggans, halflings, and sludge elves of mine come to mind. Constriction is rather stressful for these characters but on my heavy hitters like a recent MiBe, I hardly even noticed or cared. Even against tentacled monstrosities.

[crate]

I had never in the past noticed constriction damage ramping up ... if you are taking a lot of damage you needed to move away and start over (note that tentacled monstrosities have horrendous ac and ev and don't use shields so it was rarely a problem against them unless you have a character who shouldn't be meleeing them in the first place), which you can do easily against nagas. Naga warriors just have large variance in how long they take to kill and how much damage they deal, because they use equipment (including shields) and have reasonably decent AC/EV.

[dpeg]

minmay: This could be different with mimics?

[galehar]

That sounds great. I'll fire up a naga, and hopefully I'll be able to play it without the same sense of shame I felt on my last one.

I have to admit that I have focused more on monster constriction than player's. What you reported last time (low str naga killing everyone with constriction) should be fixed, but if you're playing a stronger one, I'm not sure where the balance is now...
Player's base constriction damage is 2d(str/5). I had to reduce tentacled monstrosity's base damage to 9 because of the new formula. This should work. Although we could also just use str/3. I'm not sure why there is variance for player and not for monsters.

I feel like constriction would be better if it didn't deal any damage but also kept you in place longer, with high STR making it easier to break out.

Inept and normal mimics have damageless constriction. Str already has a huge impact on escape attempts. Formula is (4+n)d(8+str/4) >= 5d(8+HD/4) (n is the number of escape attempts).

I feel like constriction really doesn't matter for high damage, high ac, high hp characters, and strongly punishes low hp, high ev characters.

Now that it's treated as a normal attack, dodgers will have a much easier time dodging those grabs. But if they get too confident and let them be be grabbed, they'll still be in trouble.

What's wrong with having variance also for constriction damage?

I don't think it's called variance, because constriction damage from monsters isn't randomized. It just increases over time. The only randomisation comes from AC. Not sure why, you should ask the guy who came up with those formulae
It used to be a simple + duration, so a ball python would deal 1 damage the first turn, and 16 the 15th turn. A tentacled monstrosity would deal 12 the first turn and 27 the 15th. Now, it's a stepdown(duration,5), so a ball python deals 1 damage on turn 1, 2 on turn 5 and 3 on turn 15. A tentacled monstrosity deals 9 on first turn, 18 on turn 5 and still 27 on turn 15 (there might be little mistakes here, I haven't charts available, but you get the idea).

One idea I think could also work is making it so constriction damage doesn't increase if you try to escape. Making the player do nothing for a turn is quite strong already!

I don't think it's a good idea. It's not very intuitive and doesn't even make much sense. But failed escape attempts are not wasted time, far from it. It makes further attempts much more likely to succeed (as indicated by the message).

[dpeg]

Okay. But once the constricting monsters have become staples and the dust has settled, one can start thinking about further applications: a monster that blinks and constricts, stuff like that.