My monster danger ranking

[phloomp]

Then I have a question. If Lair entrance might be guarded by Caustik Shrike, would you move it up in your list?

Not OP but answer is definitely "yes"

[duvessa]

I created the list by inserting one monster at a time.

Demons: summons and early banishment (the swamp end is irrelevant)
Hell hound: early banishment and the one lair entrance
Death knight: terrible hangedman vaults, could easily be a lot lower

Keep in mind that all the monsters you listed are low enough that they are pretty much harmless.

If Lair entrance might be guarded by Caustik Shrike, would you move it up in your list?

Yes, obviously. I'm aware this question is supposed to make me look stupid or fraudulent or whatever.

[VeryAngryFelid]

Not OP but answer is definitely "yes"

I see. I don't have polymorph, shafts, rare vaults, banishment for Abyssal monsters in mind when I estimate monster danger.

[VeryAngryFelid]

I'm aware this question is supposed to make me look stupid or fraudulent or whatever.

Obviously, you are wrong. I was trying to make sure I understand your factors. Actually you wrote in the OP that you took vaults in account but I wasn't sure how much.

Edit. After your calling Caustic Shrikes harmless I can do nothing but return to crawl to git gut, I died to them a couple of times and don't exclude dying to them again. Thank you for discussion.

[BabyRage]

I guess it would be nice, if you explain how you evaluate monster danger before the list, to avoid confusion for some people.

[and into]

I guess it would be nice, if you explain how you evaluate monster danger before the list, to avoid confusion for some people.

That's already in the OP, though?

I mean duvessa didn't provide a comprehensive and rigorously outlined set of criteria but, you know, this is the joke subforum of a computer game website. What do you people want?

[BabyRage]

What do you people want?

A comprehensive and rigorously outlined set of criteria would be nice. This thread doesn't seem like a joke to me.

[and into]

The heart wants what it wants

[BabyRage]

Crawl is serious business.

[lethediver]

Maybe someone should make a list of just how hard a monster is to fight if you insist on fighting it and not resetting the fight or running/avoiding. They could call it... HD... And even add it to the wiki for all to see...

[Shard1697]

I don't think HD is that accurate for that either.

[duvessa]

It'd look very similar to the original list.

[Sar]

Maybe if the list looked like "maximum damage a monster can do in a turn, also every faster than player monster gets bumped 20 positions up and every monster with a elec attack gets bumped up 10 positions" would be close to what people want to see?

[Shard1697]

Elec can be wrapped into maximum damage a monster can do per turn, distortion is the bigger deal imo. Also ranged damage should be considered, as should summons. (in which way...?)

[Arrhythmia]

OP i disagree with you on several points but i am very impressed by your dedication to this.

[Sar]

Elec can be wrapped into maximum damage a monster can do per turn

I singled it out because it's a "rare" resistance.

[duvessa]

I thought you singled it out because people are irrationally afraid of spark wasps

[Sar]

Shock serpents, also, and sixfurries (though you singled these out as well). Oh, and eels. We're trying to make a list that most people will agree with, remember?

[VeryAngryFelid]

Spark wasps are not fun, I had one 5 or 6 tiles away while adjacent to stairs so I pressed an arrow and then < without a pause. I died from about half HP because the spark blinked and was already adjacent to me when I pressed <
No wonder I have irrational fear of them after that

[Brannock]

The frustration factor of losing a character that you've invested hours in is exponentially higher than the frustration factor of losing an early character.

I believe it should not be considered valid argument. Do you want to see "Let me win, I have been playing this character for 3 hours already" button? I hope not.

No, of course not, but making a postgame monster that has as high a kill rate as a D:1 leopard gecko is something that's probably not going to happen as long as it takes several hours for the average player to make it to postgame. On top of that, thanks to how power scaling and resource options work in Crawl, players are going to have better and more consistent ways of dealing with monsters that otherwise would be as dangerous as an earlygame monster.

Sigmund has a 23.75% killratio, and Lom Lobon has a 1.17% killratio. Do you think it'd be good for Crawl if Lom Lobon kills a quarter of all players it encounters? Lom Lobon is already intimidating and already forces many characters to carefully approach the fight and use their resources to their greatest extent, or else try to evade the fight through a different approach to it (which still carries its own risks).

[Arrhythmia]

OP i understand why you don't want to rank all the permutations of DS/Dr, but could you please rank their jobs in their own heirarchy

[Sar]

well at least now we know where do your allegiances lie!

[lethediver]

So like... Shouldnt we maybe combine some of the less dangerous monsters on the list, or cut them from the game entirely?

And/or remove 2-3 floors from each branch?

And/or reduce # of consumables?

And/or remove food?

[claws]

If no_tele_into prevents shafting as it should (which hasn't been reported as an issue in quite the while), it is indeed impossible to land via-shaft into a diamond obelisk's range outside of Zot. Apis haven't been partially-natural-holiness for two months now.

I'm not sure why felids are listed so high of all of the player-species dummies- they've got no_poly_to like every not-supposed-to-be-a-real-monster player species monster (outside of historical artefact status for octopodes), derived undead were excluded from the list, and that class of monster is solely placed at this point as decoration behind stone glass in entry vaults / one Pan vault (Xom can't swap through glass, though I suppose there are miniscule chances of trap generation). Counting that class in general when the list says "there are too many monsters in crawl" feels slightly unfair, but whatever, I'm pushing quick fixes just for that octopode poly / no_trap_gen bits right now.

Outside of technical matters the only bit I will highlight remark on is blizzard demons making it to the second tier by late Lair branch ice cave and early banishment abyss spawns alone (and the latter can't be that influential for it if they're decently above hellions).

[BabyRage]

"Some people want to fuck humans who are their blood relatives and that's just fine"

Or some elves amirite?

[advil]

Discussions of difficulty/skill/danger/etc in this forum are often very strange. Here's a rant about this that I've been waiting to post somewhere, I'm not sure the bottom of a duvessa cyc thread is exactly the right place but here goes.

I think a lot of strong players who want to talk about difficulty/etc (with minmay/duvessa as probably the main exemplar, but not the only one) ignore, fail, or refuse to consider two factors:

Point 1. There's a tremendous amount of implicit knowledge, unconscious mental habits, etc that go into developing skill at a game like crawl over time. Part of what makes a player good or bad is this unconscious body of information/habits and how they apply it in the game. This is true of things like chess or go, things like playing the piano, and I don't see why it shouldn't be true of crawl. But this means that "difficult" interpreted as "difficult for me" is in many ways not useful, coming from a strong player who doesn't seem aware of the sheer amount of implicit knowledge that they have about how to play the game. Moreover, without explicit acknowledgement of it, most people don't take "difficult" to mean something different than "difficult for the person talking". If you think about what these concepts are used for in a game like chess or go where there's an in principle but unknown analytic solution to the game, they are never used like this. Rather, there are notions of "difficulty" like, for example, the decision space is very large and the causal consequences of early decisions are very far away from those decisions. Example from go: Carpenter's square. This notion of difficulty is not egocentric because, though strong players could mitigate the difficulty, it acknowledges that it is something that everyone is going to have to put in a lot of time and effort on to gain skill, and the acquisition of that skill doesn't change anything about the difficulty. It's also not tied to the analytic solution (which I suspect is what this list is supposed to be about) -- because if you had the analytic solution to go, the whole game would be easy. E.g. in the sense that duvessa uses it, the carpenter's square would be "easy" to some players. But to talk about the carpenter's square this way is, at best, not useful, and at worst, actually quite rude (if you are one of the people who knows the carpenter's square; I'm not).

Most of the time when duvessa (not just him) says something like "X" isn't dangerous, he seems to mean, "X isn't dangerous to me", which I would contend is of fairly limited use, and not what most other people mean when they use words like this. I wish the latter would at least get acknowledged.

Point 2 Crawl has a lot of moving parts, and situations in crawl can be very complicated. They don't have to be -- you can have a stripped down simple situation where it's just you, 1v1 with the adder in a hallway, no jackals or geckos around. This seems to be what a lot of people, duvessa included, would prefer these discussions to be about. But it's very unclear to me what a "normal" situation is like in crawl, even for a very strong player -- I suspect that for non-popcorn fights the average, if there can be such a thing, is not this ideal, because stuff happens. Probably a notion of difficulty like the one I suggested above for go/chess would be useful in crawl as well, and then if you're trying to evaluate monsters, a question is what scenarios can they generate (so howler monkeys or whatever might be quite dangerous by this metric, at least in theory).

One thing that is clear to me from observing my own play over time is that much of the unconscious knowledge I outlined in point 1 is aimed at making situations be simple, so the average changes with skill. A "normal" situation for me now is very, very different from a normal situation when I had ~2 wins, and the change between the two is gradual and hard to notice in its small increments, except when I think about this in the big picture. And I know from occasionally spectating that "normal" is yet again very different for strong players than me. One example: when strong players give advice like "just walk away", there's a lot of skill that goes into the ability to (a) set up encounters via movement and terrain so that that is possible in the first place, and (b) pull off the walking away without getting into a messier situation. This is "easy" by some metric in that strong players can do it very consistently. But again, a lot of this skill is basically unconscious knowledge built up over practice that it's hard to be aware of applying, like the particular details of movements of your hands while you play the piano. So there's a very real sense in which "dangerous in the ideal" is not a very useful notion of danger for most people who are interested in discussing this topic, or at least not the only one. It's also not how most players will approach the notion of danger intuitively in a game like crawl.

I get that this list is probably aiming at some sort of analytic list in order of danger, which is interesting in and of itself; minmay is probably strong enough that this is not too far off for him (though even he actually acknowledges non-perfect play for himself). This post is actually a lot more interesting than most of these discussion imo. But like most of these discussions the notion of "danger" isn't going to be the default intent of the participants, and it would be better to acknowledge this from the get-go and reframe accordingly.

[VeryAngryFelid]

No, of course not, but making a postgame monster that has as high a kill rate as a D:1 leopard gecko is something that's probably not going to happen as long as it takes several hours for the average player to make it to postgame. On top of that, thanks to how power scaling and resource options work in Crawl, players are going to have better and more consistent ways of dealing with monsters that otherwise would be as dangerous as an earlygame monster.

According to http://crawl.akrasiac.org/scoring/killers.html giant gecko killed 1.73% characters, sounds good to me. A postgame monster with similar kill rate would be fine, especially since it is postgame monster.
I remember 2 deaths to giant geckos and 4 deaths to Lom Lobon (twice to Tornado bug) if it matters.

Sigmund has a 23.75% killratio, and Lom Lobon has a 1.17% killratio. Do you think it'd be good for Crawl if Lom Lobon kills a quarter of all players it encounters? Lom Lobon is already intimidating and already forces many characters to carefully approach the fight and use their resources to their greatest extent, or else try to evade the fight through a different approach to it (which still carries its own risks).

This problem is caused by Sigmund, not by Lom Lobon. For some unknown reason (fun from quaffing unknown potions of mutation?) devs decided that having a unique with high number of unavoidable deaths is good for the game. Add "cooldown" to confuse like is done for banishment, chain-confuse is not fun. Unlike those confusing eyes who are in special branch and player can choose MR vs rC+ vs rCorr vs slaying for quick kill, Sigmund does not add any decision-making, you just meet it and die if unlucky.

[dpeg]

advil: Excellent! Extra bonus for persistent mention of the Carpenter Square If you need further go analogies, check the taisha joseki.
What you say explains also why luring is so strong: it's the steady, reliable (but unfortunately also cheap and tedious) method to ensure messy situations don't occur, and things stay ideal -- one on one. I guess that's level 2 after learning to flee.

VeryAngryFelid: Sigmund not leading to decisions sounds a bit far-fetched to me.

[VeryAngryFelid]

VeryAngryFelid: Sigmund not leading to decisions sounds a bit far-fetched to me.

Maybe I am too good player but it is the case for me.

If I am not confused and I have a scroll of tele/blink, I read it.

If I am not confused and I am not berserker, I retreat to the nearest stairs (no matter if upstairs or downstairs) and leave the level. In case of upstairs I reenter the level from different stairs and manually explore the level, diving to the first downstairs I find.

If I am not confused and I am berserker, I activate Trog's Hand and then either retreat or fight depending on having sInv.

If I am confused, I quaff curing or unknown potions. If I am lucky to quaff confusion, then I read tele/blink. If I don't have those, I choose between reading unknown single scrolls (if I have just 1-2 unknown scrolls and many scrolls are known) or retreating to nearest stairs.

Where are the decisions?

[dpeg]

VeryAngryFelid: e.g. it is an actual decision to park Sigmund, for example. In fact, you gave a whole decision tree!

Your playing level is so high that -- for you -- the choice is entirely scripted. Therefore it doesn't feel like an actual thought process. However, as advil says, it took you time/effort to become the expert system you are now. I believe that you actually mean something different: you are asking for *new decisions*, i.e. choices not yet in your decision tree. Of course, a single, early-game monster cannot achieve that (not even Sigmund), and it is likely that Crawl will never produce the depth you (and I) want.

[Malevolent]

VeryAngryFelid: Sigmund not leading to decisions sounds a bit far-fetched to me.

Maybe I am too good player but it is the case for me.

If I am not confused and I have a scroll of tele/blink, I read it.

If I am not confused and I am not berserker, I retreat to the nearest stairs (no matter if upstairs or downstairs) and leave the level. In case of upstairs I reenter the level from different stairs and manually explore the level, diving to the first downstairs I find.

If I am not confused and I am berserker, I activate Trog's Hand and then either retreat or fight depending on having sInv.

If I am confused, I quaff curing or unknown potions. If I am lucky to quaff confusion, then I read tele/blink. If I don't have those, I choose between reading unknown single scrolls (if I have just 1-2 unknown scrolls and many scrolls are known) or retreating to nearest stairs.

Where are the decisions?

If this is your decision-making process when you see Sigmund, then I would go ahead and say your strategy is suboptimal (for one, it is too inflexible).

[VeryAngryFelid]

If this is your decision-making process when you see Sigmund, then I would go ahead and say your strategy is suboptimal (for one, it is too inflexible).

You are welcome to provide your general strategy and tell me where my strategy is suboptimal.
I neglected to mention obvious things like reading scroll of fog, getting summons from scrolls/spells/evokables etc., since those depend on terrain. Of course my strategy will be different when Siegmund lands adjacent to me due to teleportation trap but still it is pretty obvious what to do every time. It is not like I think long and hard what to do. It happens because I have only a few options and one of them looks much better than all others and I don't have time to experiment trying different options because of confusion-chain.

[Malevolent]

If this is your decision-making process when you see Sigmund, then I would go ahead and say your strategy is suboptimal (for one, it is too inflexible).

You are welcome to provide your general strategy and tell me where my strategy is suboptimal.

Well, I wouldn't waste a teleport scroll, and especially not a blinking scroll, if I could win the fight with a high probablility with the use of no consumable or a less important consumable. Like, if I'm a strong melee character, I won't need Trog to just take him on, provided I have some sources of healing (several curing or 1 curing/HW, at least) and something that lets me chop him up quickly (like a potion of might or a particularly strong weapon, like one of venom/electrocution). If I have some wands, I also won't hesitate to sacrifice a few charges for him. (Often 1 successful paralyzation or confusion is enough, for example.) I don't usually think much about it either, but I do take into account most of the options I have - and I'd rather risk sacrificing a few less important consumables and then have to teleport anyway than to miss out on the good chunk of XP Sigmund provides. Also, while chain confusion is bad, Sigmund also tends to do a lot of other stuff like magic darts, invisibility or simply moving - thus with a curing potion or two in reserve, it's not generally an issue. Of course, all this assumes that I'm just playing normally and not trying to streak (in which case avoiding immediate death is a much higher priority - generally I'm quite willing to increase the chance of dying on the first few floors a bit, in exchange for reducing the chance of dying for quite a few floors afterwards).

[tabstorm]

If this is your decision-making process when you see Sigmund, then I would go ahead and say your strategy is suboptimal (for one, it is too inflexible).

You are welcome to provide your general strategy and tell me where my strategy is suboptimal.

Well, I wouldn't waste a teleport scroll, and especially not a blinking scroll, if I could win the fight with a high probablility with the use of no consumable or a less important consumable. Like, if I'm a strong melee character, I won't need Trog to just take him on, provided I have some sources of healing (several curing or 1 curing/HW, at least) and something that lets me chop him up quickly (like a potion of might or a particularly strong weapon, like one of venom/electrocution). If I have some wands, I also won't hesitate to sacrifice a few charges for him. (Often 1 successful paralyzation or confusion is enough, for example.) I don't usually think much about it either, but I do take into account most of the options I have - and I'd rather risk sacrificing a few less important consumables and then have to teleport anyway than to miss out on the good chunk of XP Sigmund provides. Also, while chain confusion is bad, Sigmund also tends to do a lot of other stuff like magic darts, invisibility or simply moving - thus with a curing potion or two in reserve, it's not generally an issue. Of course, all this assumes that I'm just playing normally and not trying to streak (in which case avoiding immediate death is a much higher priority - generally I'm quite willing to increase the chance of dying on the first few floors a bit, in exchange for reducing the chance of dying for quite a few floors afterwards).

"Your strategy is suboptimal and too rigid, and mine is better, because I expend resources for experience instead of using them to avoid dying. But, if you are trying to win, then your strategy is actually better" - you
Come on man.

[VeryAngryFelid]

By the way potion of might is a more precious consumable than scroll of teleportation for me. It is more rare also.

[duvessa]

If no_tele_into prevents shafting as it should (which hasn't been reported as an issue in quite the while), it is indeed impossible to land via-shaft into a diamond obelisk's range outside of Zot. Apis haven't been partially-natural-holiness for two months now.

I'm not sure why felids are listed so high of all of the player-species dummies- they've got no_poly_to like every not-supposed-to-be-a-real-monster player species monster (outside of historical artefact status for octopodes)

I was just out of date on these then.

Ice cave blizzard demons aren't a threat to actually kill someone playing patiently, but they are a threat to take away resources. Similar situation to stuff like black mambas, except obviously the blizzard demon encounter is a lot rarer, which is why they're lower on the list!

[shnurlf]

Hell hound? Weak in Volcano, weak in Lair ending. Or do you mean they can guard entrance to Lair like wiki claims?
Then I have a question. If Lair entrance might be guarded by Caustik Shrike, would you move it up in your list?

To be fair to the both of you, 2 Hell Hounds do sometimes appear as guards to Lair, but they're completely inconsequential. They've never given me difficulties with the characters I encountered them. But then again, I've never run into them as a caster.

[VeryAngryFelid]

Oh, I confused Spark Wasps with Shock Serpents, the latter discharge when you hit them.

[BabyRage]

Gross.

[Lasty]

[list][*]Juggernaut. I have voiced my dislike of this monster's design before, so I will just remind you how easy it is to stop them from following you across stairs, or neuter them with level 1 summons. Restoring their old damage would not change their place on the list.

Yeah, I think these were a failed experiment.

[Malevolent]

[list][*]Juggernaut. I have voiced my dislike of this monster's design before, so I will just remind you how easy it is to stop them from following you across stairs, or neuter them with level 1 summons. Restoring their old damage would not change their place on the list.

Yeah, I think these were a failed experiment.

I'm not so sure. I've played plenty of characters for whom they were a relevant threat (as much as Depths monsters can be). They seem to be designed to counter the most simple and braindead, yet powerful kind of build you can play - pure melee (sadly, they're not actually good against Troglodytes because of the bros - maybe giving them Cleave could upset their weakness against summons?).

[Sar]

I thought they had cleave already?

[PleasingFungus]

I thought they had cleave already?

No, they just punch things.

[VeryAngryFelid]

If juggernauts are stopped by summons, we can fix it:
1) Give juggernaut passive "Charming aura" ability which makes every hostile monster in view friendly except PC. If PC uses rod of shadow or summon butterflies, those summons become friendly to the juggernaut
2) Give juggernaut "Friendly swap" ability which allows juggernaut to swap position with any friendly monster in view (smite-targeted of course). If PC uses rod of shadow or summon butterflies, juggernaut uses the ability to get adjacent to PC.

If juggernauts are stopped by blinking, we can fix it by giving 50% chance to fail when using the spell adjacent to juggernaut.
"Are you sure you want to cast Blink while adjacent to juggernaut?".

[Sar]

I don't like your ideas a lot, Sandman (sorry, they seem very fiddly!), but I kind of like an idea of giving juggs some kind of a magic-breaking ability. Also, if they don't have cleave I think it was at least suggested. Maybe they should?

[VeryAngryFelid]

I don't like your ideas a lot, Sandman (sorry, they seem very fiddly!), but I kind of like an idea of giving juggs some kind of a magic-breaking ability. Also, if they don't have cleave I think it was at least suggested. Maybe they should?

That's ok, I understand the ideas are too revolutionary for crawl. But I like juggernauts and try to trigger discussion about how to save them, they are one of 3 depths monsters I respect with any character. Are they removed in trunk?

Cleaving does not help much with the way juggernaut attack works IMHO. You can cast butterflies 3 times during the time it takes juggernaut to kill them (that's assuming juggernaut does not miss which is not a given provided how high EV butterflies have).

[Sar]

Yeah they are in the game still.

[duvessa]

If juggernauts are stopped by summons, we can fix it:
1) Give juggernaut passive "Charming aura" ability which makes every hostile monster in view friendly except PC. If PC uses rod of shadow or summon butterflies, those summons become friendly to the juggernaut
2) Give juggernaut "Friendly swap" ability which allows juggernaut to swap position with any friendly monster in view (smite-targeted of course). If PC uses rod of shadow or summon butterflies, juggernaut uses the ability to get adjacent to PC.

If juggernauts are stopped by blinking, we can fix it by giving 50% chance to fail when using the spell adjacent to juggernaut.
"Are you sure you want to cast Blink while adjacent to juggernaut?".

A fix that requires breaking other things is not a good fix.

[Arrhythmia]

OP pretty pretty please just rank the Ds and Dr classes among themselves. don't bother with the colours, just the classes.

[Midn8]

On CAO about 5% of deaths are from gnolls. We could just look at that list.