The Tavern

That is not true. You are always killed on a given turn. Games with no particular turn on which you died are either wins or quits. What VAF says upthread is true: If you're taking "average" damage each turn for a succession of turns and you die as a result, you're doing something wrong. Assuming reasonable play, you're most likely to be killed by high damage, low probability events. Harm makes those events much worse.

As for Majang's argument, he's making a facile comparison between two very different trade-offs, one that massively increases max damage and one that has little or no effect on max damage. This is unconnected with questions asked elsewhere in the thread about hypothetical alternative harm amulets. My take on Siegurt's question, beyond what I've already said, is that it's moderately complicated and if you can't get people to recognize the importance of max damage in crawl combat, getting into finer analysis of tails of damage distributions and their impact on crawl tactics is a nonstarter.

I think I understand you now, and I retract my comparison with two-handed fighting. I think the cleaving analogy may indeed get closer to the idea of what harm does; I mean, putting myself into a situation where cleaving actually takes place, not just having a cleaving weapon.
Two-handed fighting seems still a good comparison for characters that cannot easily extricate themselves from fights that are going bad, such as Nagas, Barachim, Formicids, and Chei worshippers. Once they are committed to a fight, they may have to finish it, and two-handed fighting is likely to affect their odds in the same way as harm does.

watertreatmentRL wrote:Average damage tells you something about long term dynamics of combat, but it tells you nothing about whether you can die this turn. This is what maximum damage tells you. Whether or not you can die on the current turn is probably the most important question in crawl combat.

However your *current hp* is also a factor in "can i die this turn" and in turn, how much damage you have taken so far (which is in turn effected both by the average damage of the critters who have attacked you, and how many times they could do so)

If that wasn't true, ac and ev would be completely unimportant, as neither one effects the maximum amount of damage you can take (well ac can in some cases, where gmdr applies)

So *if* you could say "the extra damage done by harm increases the amount of hp i have available on an average combat turn by more than the amount it increases the maximum damage done by dangerous creatures" it would probably be worth it. I don't think that can be the case for even the most contrived situations presently, but i think there are set of numbers for which that could be true for a subset of characters.

My opinion is that the game is somewhat balanced around the current max damage of monsters and that adding a +20% incoming damage modifier makes things a lot more dangerous, to the point that +30% more damage dealt does not make up for it. Personally I think that the difference between bad/balance/overpowered on harm numbers changes so quickly that it isn't reasonable to try to balance it. I also think that currently the only good use of harm is as a swap. As the devs are against amulet swapping, the design philosophy mentions that putting stuff in the game that isn't supposed to be used is bad game design, and because the amulet is a noob trap, it should just be removed.

Also, the existence of harm makes using amulets in the early game suboptimal. Something that makes wearing an entire class of items suboptimal and instead going with a blank slot is flawed.

By "not able to be killed on any given turn" i meant "not at risk of dying on every turn"

@siegurt: I don't see how this is responsive to what I said, but probably the simplest way to think about the effect of harm is that it is as if you have frail 2. This is not quite right, it's off by a couple percent and your heal potions and hp regeneration are a bit weaker than if you had real frail, but it's pretty similar. The reason I bring this up is that no matter how amazing your damage output is, you're still starting the fight with about 20% less hp than you otherwise would.

There's another similar tradeoff in the game: Hepliaklqana. You get frail 1 in exchange for immediate access to a powerful ally that is consistently powerful throughout a 3 rune game. It often makes sense to take Hep, it's a strong god. +30% damage is not as good as the benefits of the hep religion, in my opinion.

watertreatmentRL wrote:The reason I bring this up is that no matter how amazing your damage output is, you're still starting the fight with about 20% less hp than you otherwise would.

Right, and I agree that 30% more damage isn't worth 20% less hit points, but I posit that there *is* an amount less hit points that's worth 30% more damage (or alternately that there's an amount of damage that's worth 20% less hit points) if the amount of bonus damage was "enough to always kill anything in the game at max LOS with a thrown stone and 0 skill" obviously that'd be worth 20% less max hp, and similarly I think that if the amount of hp reduction was "exactly 1 max HP less" it'd be worth it for 30% more damage. I am not trying to argue that the amulet is worth using given it's current numbers.

Yes, there's a continuity argument to be made there. Presumably there are values of the parameters that would make the amulet acceptable from the perspective of best practice crawl and maintain some semblance of balance. The next layer is whether there are values of the parameters that make it an item that is interesting and valuable as an element of the game. This is where the ship of reform runs aground, in my opinion.

edit: On reflection, I'm actually skeptical about the continuity idea here too.

Siegurt wrote:If that wasn't true, ac and ev would be completely unimportant, as neither one effects the maximum amount of damage you can take (well ac can in some cases, where gmdr applies)

This is wrong, AC and EV cannot be completely unimportant because they change probability of getting max damage. But arguably you are right, I ignore AC/EV when I can be killed next turn and I read scroll of blinking, use sanctuary, berserk etc. Yet AC/EV really help to avoid those "you have so low health that you can be killed by max damage next turn" because your average damage taken is much lower during previous turns.

Somebody mentioned that game is somewhat balanced around average damage, I am not sure it is true because otherwise we wouldn't have so high max damage and would have damage formula like {average_damage +- some_fraction_of_average_damage} instead. It looks like devs like to create exciting moments "oh, crap, I lost 45% HP in a single turn, now I need to do something" except it does not work because I've spectated top players behaving like "Well, this is very unlikely to happen again, let's continue to do what we were doing, there is no need to spend scroll of blinking now I think".

VeryAngryFelid wrote:
Siegurt wrote:If that wasn't true, ac and ev would be completely unimportant, as neither one effects the maximum amount of damage you can take (well ac can in some cases, where gmdr applies)

This is wrong, AC and EV cannot be completely unimportant because they change probability of getting max damage. But arguably you are right, I ignore AC/EV when I can be killed next turn and I read scroll of blinking, use sanctuary, berserk etc. Yet AC/EV really help to avoid those "you have so low health that you can be killed by max damage next turn" because your average damage taken is much lower during previous turns.

Yes, if you read what I wrote again, you'll see that's exactly what I was saying, I was saying "If you only consider max damage for a creature and nothing else, then ac and ev aren't important" the obvious implication was "that is not a valid point of view to evaluate the game from"

Perhaps people can move away from defining "optimal play" to be "play that maximizes winrate". It is a mostly arbitrary choice, few people actually play the game that way, and this measure renders 90% of the game meaningless (the game is won by Lair, isn't it?)

How else would you define "optimal play?" FWIW I think there is something to be said for gauging the merits of strategic decisions based on "best practice considering the player is a human and liable to make mistakes in the future." For example: walking around with an elemental vulnerability is not particularly bad if your tactical play is consistently above a certain threshold of good, but if you know you like to tab & autoexplore & not always read messages, perhaps your personal winrate would be higher if you didn't walk around with elemental vulnerabilities except in exceptional circumstances.

In a similar vein, I might win a greater % of games in which I autoexplore (in moderation, as opposed to exploring everything by hand) because I can be impatient and manually exploring an entire 3-rune game would potentially frustrate me and cause my attention to lapse elsewhere. But obviously using autoexplore is "suboptimal."

There are many alternative ways to define "optimal play". One simple way is play which maximizes expected score, instead of expected winrate.

Ideally, you want to distinguish people based on skill. Score doesn't do that perfectly (though it's decent), and winrate is worse. Furthermore, since "optimal" winrate (if people bother to play to maximize winrate) is so close to 100%, there is a lot less granularity to distinguish people. In this situation, the RNG becomes much more important (if you get shafted into a dangerous situation, or if you get unlucky and get paralyzed or banished). This is both boring and silly.

You can only compare two people who hold the same definition of 'winning' or 'better' in mind, though. "Competing with myself/other players for highest score or highest score given a starting combo" is a metagame, tangentially related to the isolate experience of "playing a game of crawl," which has a clearly defined set of rules and in which a player can lose or win.

You can play the game however you like, that is not the point.

"Optimal play" however, is only arguable once one has some standard or measure in mind. Indeed all advice here is based on some implicit standard in mind. If someone asks "is the amulet of harm worth it", they have some kind of standard or measure in mind. If someone says: "no, it sucks", they also have some standard or measure in mind.

To take a concrete example: Worshipping Chei (depending on how you play) can lower your winrate, but can increase your score. Therefore, worshipping Chei could be good or bad based on what standard you have in mind.

It's good for casters if you can one hit. I did that once with a FE with the harm amulet and enhancers. Fireball was either one shot killing or nearly killing everything on one cast in the depths. Clearly, no defense needed. Just mind the noise.

It's also good for throwers or ranged units that can stay out of melee range. It helps a lot for types that excel in melee and using a 1-handed weapon. I'd imagine it's also good for high health types, but those types tend to have a problem with defenses so maybe not the case.

4Hooves2Appendages wrote:Threads where people disagree are usually the most interesting and most useful!

This is absolutely true, but I responded to a report in this thread, and as a reminder: it's fine to disagree with people's ideas, but don't personalize these disputes. Arguing about video game is fun. Stop getting angry about video games. Thanks.

The Tavern

Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?

Re: Are amulets of harm worth it?