v4 - a few characters

[CliffStamp]

Yes, but for me they are the same, the play is the character, and it just feels like wizard mode when you get that much information.

I would prefer that it actually was fuzzy and became more clear in use, ideally it would get more specific with use and eventually end up at 57% or whatever but it would have to be after a lot of blows landed.

One of the things I liked a lot was watching weapons almost develop new abilities in use as more and more things came clear. I actually stopped using ID scrolls on weapons and it reminded me of TOME where you had sentient weapons which gained abilities through EXP.

I realize the abilities were always there sure, but still, it was just another aspect of development.

It is similar to as you explore the dungeon things are revealed. If you pay with an enlightenment class you lose a lot once you get that as all you start doing is full detect / scan (teleport as needed), stairs, down and repeat.

But I do realize there are people who want all the mechanics, and love to read intricate spoiler files and people who love to write them, i.e., Jarulf and Diablo.

[Magnate]

I have to say I quite like Cliff's idea of the precise numbers becoming available only eventually, with the old system of adjectives in use beforehand. But this is an area where we'll never please everybody, so we shouldn't try. I have to say I dislike the enforced vagueness of Crawl and prefer the transparency of modern-day Angband if it's a straight choice.

On a side note, you can do some serious damage by hitting a shield hard enough, so I've tended to think of shields as adding to abs rather than Ev. But yes, we've always intended to extend the ev/abs system to @ eventually, but it's much more complicated because of all the equipment changes - it adds a whole extra stat for both player and gear.

[Derakon]

So I guess the question then is, is there a way we can reasonably support both kinds of players in the same game? Intuitively, it seems like the drastically different levels of knowledge the players would have would make balancing the game for both players an utter nightmare, even though we would of course accept that the "opaque" players (to whom attributes, numbers, etc. would be less available) would have a harder time than the "transparent" players (playing the game basically as it is now).

Some parts of the game are trying to reclaim mystery without sacrificing rule transparency -- that is, while you know exactly how things work, you don't know exactly what your situation is. Fuzzy detection is a big step in that direction, and there have been some suggestions for making monster detection and telepathy fuzzier too. There's obvious gradations there -- level 1 is "there is a monster here", level 2 is "there is a 'D' here", and level 3 is "Ancalagon is here". You could even assign them to pvals on items, letting you gradually get better and better telepathy / detection as you upgrade your gear / cast stronger spells.

Doing away with any ID that was not ID by use would also be a big step towards making the game more mysterious, though assuming we keep the rune-based approach you would still end up knowing basically everything about every item once you got into the late game.

The real issue you seem to be worried about, though, is the exposure of numbers. What are the big offenders? Off the top of my head:

1) Damage dealt by your attacks (and to a lesser extent, blows per round)
2) Damage dealt by enemy attacks (melee and spells)
3) Chance to hit / be hit by enemies
(?) 4) Probability of using a magical device / casting a spell successfully

Intuitively it seems like we ought to be able to track how well the character knows each of these attributes and pin it down to three broad categories: the character knows nothing, they know roughly how big the number is, and they know the number precisely. Presumably you'd hit the second category as soon as you used the item / got hit by the monster / etc., and the third category after many repetitions.

It gets more difficult when many of these are on sliding scales, though. For example, we want unenchanted Chain Mail armor to be roughly as good against hill orcs as Adamantine Plate Armor is against ancient dragons -- but the two items are numerically completely different. If we use an absolute scale for the broad categorizations, then we end up telling the player that their starting armor is universally crap even though it makes a material difference at this stage. Similarly, 15 damage/round is fantastic at level 1 and very rapidly gets outclassed. My best idea for dealing with this so far is to have the game track statistics on what the player has been fighting lately and use those stats to form the descriptions -- but that's a pretty heavyweight way to go about doing things, and would probably still result in oddities.

[ekolis]

It gets more difficult when many of these are on sliding scales, though. For example, we want unenchanted Chain Mail armor to be roughly as good against hill orcs as Adamantine Plate Armor is against ancient dragons -- but the two items are numerically completely different. If we use an absolute scale for the broad categorizations, then we end up telling the player that their starting armor is universally crap even though it makes a material difference at this stage. Similarly, 15 damage/round is fantastic at level 1 and very rapidly gets outclassed. My best idea for dealing with this so far is to have the game track statistics on what the player has been fighting lately and use those stats to form the descriptions -- but that's a pretty heavyweight way to go about doing things, and would probably still result in oddities.

Wasn't there one variant that actually adjusted the skills display (fighting, digging, perception, etc.) based on some dynamic factor? So a starting character might still have "excellent" in most of the skills, but only if the skills improve will they remain "excellent" later on... I think it may have been Sangband, and the dynamic factor may have been the max depth. This actually kind of makes sense, as if you dive too deep, the monsters will get stronger and overwhelm you, but if you stay shallow, you will overwhelm them! (Of course, it does get a bit silly if you dive initially but then stay shallow for a while afterwards...)

[Magnate]

It gets more difficult when many of these are on sliding scales, though. For example, we want unenchanted Chain Mail armor to be roughly as good against hill orcs as Adamantine Plate Armor is against ancient dragons -- but the two items are numerically completely different. If we use an absolute scale for the broad categorizations, then we end up telling the player that their starting armor is universally crap even though it makes a material difference at this stage. Similarly, 15 damage/round is fantastic at level 1 and very rapidly gets outclassed. My best idea for dealing with this so far is to have the game track statistics on what the player has been fighting lately and use those stats to form the descriptions -- but that's a pretty heavyweight way to go about doing things, and would probably still result in oddities.

TL;DR - trying to describe @'s combat stats without reference to a specific monster is a lost cause.

I think the right approach is the way we currently deal with to_hit - we don't try to show an absolute rating, we merely show it in the monster recall, as it applies to that specific monster. Exactly the same approach would work for both damage and armour (both ev and abs, eventually):

You have a high chance of hitting this monster with your current weapon.
Your blows will seriously hurt this monster (or "will kill it in a few rounds", or whatever).
You have a high chance of dodging this monster's attacks.
Your armour will not protect you from much of its damage.

The last pair could, if we wanted, vary by the monster's attack, so you could have them up to four times, saying how much of the physical and elemental/other types of damage you would absorb/resist. We'd need a syntax for that, lest it get very wordy.

[Derakon]

The last pair could, if we wanted, vary by the monster's attack, so you could have them up to four times, saying how much of the physical and elemental/other types of damage you would absorb/resist. We'd need a syntax for that, lest it get very wordy.

This kind of thing should, I think, be reserved for a future in which the monster memory is shown in a grid system (like the character sheet) instead of the current paragraph system.

...which, come to think of it, is even easier if you support mouseover tooltips. "What's that 38% mean next to the "Hit: 14d8"? Oh, it's my damage reduction. Cool." And if you play without a mouse, you either check the docs or just deduce it from experimentation.

[CliffStamp]

There are always going to be extreme points at the ends of any idea and there is little point in trying to accommodate them both unless you want a list of options. The best you can do is try to make some kind of consistent behaviour which is carried throughout all of the game.

The fuzzy detection to me is a great thing as it encourages exploration and radically reduces the tenancy to item scum/farm. Using a fuzzy display of armor/weapons would encourage actual use of weapons/armor and not simply ID/ditch.

How about something simple such as :

-when you are attacked or attack you learn what happens in that attack and that is reflected in the description

-as you attack/are attacked more the average starts to become more clear

-a full ID would show this immediately (the expected value)

That way people who wanted to explore could have it happen naturally and people who wanted just to ID could bulk up on scrolls/staves.

This would however not be trivial to code as you would need another tag in the weapons/armor for the displayed values.

This is the way I play it now, I don't really calculate all the slays, etc. . I just try out the weapons and see how long it takes them on average to kill.

[Magnate]

How about something simple such as :

-when you are attacked or attack you learn what happens in that attack and that is reflected in the description

-as you attack/are attacked more the average starts to become more clear

-a full ID would show this immediately (the expected value)

That way people who wanted to explore could have it happen naturally and people who wanted just to ID could bulk up on scrolls/staves.

I could be wrong, but I think there's a post from Eddie circa 2009 that suggests almost precisely this. And I suspect you're right that it wasn't implemented because of the difficulty of displaying different values to the 'real' values. Nice idea though - no reason we couldn't do it one day.

[LostTemplar]

Maybe add an option, that removes all combat info from weapon description, also hides all numerical character stats (replace HP/mana with percent value, or just some description e.g. "badly hurt", etc.), but give information about damage on hits, similar to wizard mode.

[d_m]

EDIT: tl;dr, I'm mostly agreeing with Cliff and Eddie circa 2009

I'm someone who'd like an option to make the numbers more transparent. But I think the point about fuzzy detection is well-taken.

Basically, I feel like information that the character probably would have should be given to the player. It's pretty easy to tell when someone is faster or slower than you, so I think that (at least) should be obvious for any seen monster (and really, I think exact speed should be shown). I think a bit more specificity about how hard the player hits (i.e. how much damage was done) would be good... right now those numbers are pretty disassociated from combat.

In my perfect world, the game engine (rules) would be assumed to be open, and so whatever statistical modeling/estimate the player might do could be done automatically. So, after hitting a monster a bit you might guess that it has between 120-150 hp (based on damage plus fuzzy health bar), and eventually narrow in on its real total. This is a generalization of how the monster recall currently works.

There are some tricky questions, in terms of what knowledge the player starts with versus what is learned. You could even support a "no numbers" style where the game engine is assumed to be known, but all values are hidden, and so the game figures out various percentages/symbolic amounts for you by calibrating against your actual play.

But in any case, I think when we provide numbers (e.g. weapon to-hit/damage bonuses) we should provide/assume knowledge of the game engine and try to provide some other useful context (i.e. how the weapon is actually performing).

[CliffStamp]

But in any case, I think when we provide numbers (e.g. weapon to-hit/damage bonuses) we should provide/assume knowledge of the game engine and try to provide some other useful context (i.e. how the weapon is actually performing).

I can understand the logic there, but it doesn't fit how I see the character interacting with the world. Imagine you are a novice warrior, you read a scroll of ID on a weapon, it gives you a list of numbers - would you then be able to tell from that description of the enchantment how much damage it would do? Or forget about Angband for a second, if I was to give you statistics on a bullet (fps, impact energy, etc.) could you figure out the damage it would do to a target?

I would argue that you should be able to know the weapons main dice but the to-hit, and to-dam (or whatever you want to call it) are things that would only come with use. The class/race modifiers would then influence how rapidly this took to become less fuzzy. I high level / high skill warrior could likely figure out the exact damage ranges with just a little use, but a novice mage certainly would not.

This to me would move the game more away from ID crunching, would again differentiate characters and add to immersion in game play. But I don't see any way to make the coding trivial because you would need to store the "known" values for weapons/armor for each item used until they were destroyed (left on a level and abandoned). Quite frankly I would like to see the same for rods, etc. as well but again that is a lot of coding.

It is also an argument that making this change really could be sufficient grounds to claim variant as this is a pretty big chance in game play vs the Angband which came from UMoria.

[Magnate]

I can understand the logic there, but it doesn't fit how I see the character interacting with the world. Imagine you are a novice warrior, you read a scroll of ID on a weapon, it gives you a list of numbers - would you then be able to tell from that description of the enchantment how much damage it would do? Or forget about Angband for a second, if I was to give you statistics on a bullet (fps, impact energy, etc.) could you figure out the damage it would do to a target?

I would argue that you should be able to know the weapons main dice but the to-hit, and to-dam (or whatever you want to call it) are things that would only come with use. The class/race modifiers would then influence how rapidly this took to become less fuzzy. I high level / high skill warrior could likely figure out the exact damage ranges with just a little use, but a novice mage certainly would not.

This to me would move the game more away from ID crunching, would again differentiate characters and add to immersion in game play. But I don't see any way to make the coding trivial because you would need to store the "known" values for weapons/armor for each item used until they were destroyed (left on a level and abandoned). Quite frankly I would like to see the same for rods, etc. as well but again that is a lot of coding.

It is also an argument that making this change really could be sufficient grounds to claim variant as this is a pretty big chance in game play vs the Angband which came from UMoria.

I think you and d_m are in more agreement than it might appear! It's not too hard to add disp_to_d and disp_to_d to the o_ptr structure to store the "currently known" values, and I agree that it would be a lot more immersive. I don't think it would really be that big a change, for those of us who have got used to ID-by-use. It's really just a halfway house between not knowing the plusses until you cast ID, and knowing them on the first hit (which is what we have now).

I like your idea of extending this kind of gradual learning to devices too. That's a little harder to do, because devices have all sorts of variables (damage, range, duration, etc.) which would need to be learned gradually, instead of just plusses.

In fact, that same logic could be extended to pvals - I'm thinking particularly of slays, but it could also be applied to other pvals, like stealth or speed or stats. But that really would be quite a big change - not knowing exactly how fast or strong you are is likely to discomfit a lot of players.

[LostTemplar]

Or forget about Angband for a second, if I was to give you statistics on a bullet (fps, impact energy, etc.) could you figure out the damage it would do to a target?

All this numbers, e. g. damage dice, hitpoints, etc. are only parameters of the model, in real life they does not exist, so the only real number, that characterise a weapon effectiveness is how many hits it takes to kill an average orc, or something like this.

[Derakon]

At this point we start running into the issue of how "real" all the gameplay abstractions are. In other words, we can pretend that we are inexactly representing a realistic world, in which case the adventurer only knows how heavy, sharp, etc. his weapon is and so on. But then said adventurer must start wondering why the orc he's fighting is just as spry at 1HP as it was at 80HP, or how exactly he survived being crushed by falling rock, etc. Or we can pretend that we are exactly representing an abstracted world, in which case the adventurer knows that his world runs on a system of hitpoints and numerical damage and special-cased rules for earthquakes and so on. Neither case is ideal, and I think no matter how you decide to represent information to the player "based on what the character has figured out", you're going to run into holes because of these disconnects.

[LostTemplar]

But then said adventurer must start wondering why the orc he's fighting is just as spry at 1HP as it was at 80HP

Well, you missed my point, IMHO HP is just a model parameter, not something, adventurer can wonder about, adventurer just hit orc 3 times and orc dies, no hp no damage, no dice, etc. all this is just a part of a computer model. In other words I dont care too much about how things are modelled, if an orc dies in 3 hits, anything else does not matter (e.g. dwarf fortress game has a very complicated model), and I dont need any numbers to figure out, that one weapon kill orc in 3 hits, while another takes 5 hits, and so make a decision, that first one is better.