Level feeling with 3.3.1?

[bio_hazard]

I'm starting to come around to prices not changing depending on the current condition of the player.

Prices should certainly certainly take into account the future value to a player. It wouldn't make sense for better equipment to cost less because the player can't activate/can't wear without massive penalty/can't swing more than once. Should big armor cost less for mages? Should a high ego weapon that they rarely use in melee cost less for a mage? This doesn't sound right to me imho.

I think the "sell back" problem could be solved by either further lowering what shopkeepers will pay, tagging bought items to be more deeply discounted, or tagging BM items so they can't be sold in another store.

Honestly, it's an exploit that hardly needs fixing- if someone wants to hoard lose memories enough to make a little money, they are probably only going to be able to do this a couple times in a full game, and they are giving up an inventory slot and a home slot to do so (not counting restore potions which are otherwise useful).

[Magnate]

JFTR, I don't support a move to current-utility pricing either. I just wanted to show that as the author of the current oft-derided pricing regime, I'm totally open to discussing and potentially implementing alternatives.

At the moment we have a pricing system that reflects a pretend market economy, where shopkeepers do not change their prices according to who you are - they are constant for all adventurers in the pretend world. Which means it has to use absolute values for things whose utility varies greatly according to what class and level you are, what other gear you've found, and your personal preferences.

Just in case anyone wasn't aware, the pricing model was designed around what one might call "ultimate utility" (instead of current utility) - i.e. it rates things according to how useful they are for an endgame character with endgame kit. This is why, for example, regen is rated the same as +1 WIS, and why base4 resists are much cheaper than high resists despite being much more important. It might seem obvious that that's wrong for much of the game, but it is at least consistent with the rest of the modelling.

The assumption of 5 blows is another example of problems caused by the endgame assumption. I'm hoping that the new v4 combat system will allow me to design a way around it, but it might not.

So since we are again discussing pricing (and this is the first such detailed thread for several months), I'm very interested in whether people think either:

- we should choose some other benchmark than "endgame kit" for setting the values of different attributes, or

- if there is a way of calculating the value of more stuff empirically. Just as we currently calculate the value of damage-altering attributes by the amount of additional damage they offer against the range of monsters in monster.txt (leaving aside the fundamental issues with blows, the values of slays relative to each other are empirical), I have long aspired to calculate the values of resists in the same way (using the power of monsters whose attack it resists). What about regen or telepathy or stats or anything else? Can anyone suggest ways to calculate their value without arbitrary constants?

[Estie]

The pricing for endgame kit is pretty accurate, but as Derakon has stated, once you are at that point prices dont matter anymore. While I think variable prices that change with character properties are generally a bad idea (think nethack: Adinohopos outrageous prices for food when you are hungry), the only idea I have would be to have 2 kinds of shopkeepers: those who evaluate according to early game needs, and those who do according to late game (as its now). When the game starts, its all early game type and they gradually change towards endgame type as the game progresses.

Weird prices however arent a big concern, its when pricing is used for other evaluations like level feelings that it can become one. I like the fact that evaluation of properties changes as the game progresses; I am rather afraid that the new combat system might streamline those bumps away. Maybe thats just me though.

Evaluation of weapons:

Currently, early game everyone wants str and dagger, lategame all stats are the same and everyone wants MoD.

This should become: early game hobbits want dex and dagger, trolls want str and lance; as for lategame, there was dispute. Derakons combat revamp goes in this direction and we will see how it turns out.

Evaluation of armour:

AC from armour comes with 2 penalties: weight which results in negative speed and mana reduction (ignoring the minor to hit malus).

The problem is that both penalties are very severe early game and a trifle lategame. Speed from 0 -> -2 is bad, speed from 35 -> 33 not so much. Mana from 30 -> 15 is bad, mana from 250 -> 235 not so much.

This means that early game, the robe is the best armour type, while lategame its the one with highest AC.

Is this situation good or bad ? Some consequences are:

- lategame, everyone wants the same stuff; Isildur drops, it gets equipped by mage or warrior.
- early game, noone wants the rare out of depth heavy armour junk (unless to sell it....)
- not even warriors who suffer from only 1 of the 2 penalties; I find that the extra weight allowance from high str is much better used by carrying more staves of id and potions of ccw than by getting a bit of AC.
- lategame, you cannot give the melee guy more protection than the ranged guy. Melee is intrinsically more hurting than ranged, but the only way to give some advantage to balance melee is to increase the damage compared to ranged.

[Magnate]

Originally Posted by Estie

The pricing for endgame kit is pretty accurate, but as Derakon has stated, once you are at that point prices dont matter anymore. While I think variable prices that change with character properties are generally a bad idea (think nethack: Adinohopos outrageous prices for food when you are hungry), the only idea I have would be to have 2 kinds of shopkeepers: those who evaluate according to early game needs, and those who do according to late game (as its now). When the game starts, its all early game type and they gradually change towards endgame type as the game progresses.

Weird prices however arent a big concern, its when pricing is used for other evaluations like level feelings that it can become one. I like the fact that evaluation of properties changes as the game progresses; I am rather afraid that the new combat system might streamline those bumps away. Maybe thats just me though.

Well, the adaptation of the power rating system to the new combat system is quite specific to v4 and unlikely to affect V for quite a while, so let's put that to one side (since my main objective in this discussion is to see if there are immediate improvements to ratings in V). I quite like your idea of two different kinds of shopkeepers, but as you say, this isn't really about prices, it's about the algorithm that lies underneath them, which is used for level feelings and randarts as well. For randarts it's appropriate that endgame ratings are used (that's why the algorithm is biased that way in the first place). For level feelings, how about simply using a depth slider? So for example, instead of assuming 5 blows when rating a weapon, why don't we assume 1 + depth / 20 blows, capped at 5? Then we can actually do a useful adjustment based on weight (something like assume +0.5 blows at each of 9lb, 6lb and 3lb, capped at 5). Not using actual stats, but an approximation so that light weapons are rated more realistically compared with heavy weapons during the early portion of the game.

This would mean that weapons got more expensive as you got deeper, but that doesn't seem like a terrible price to pay to fix the basic problem.

- lategame, everyone wants the same stuff; Isildur drops, it gets equipped by mage or warrior.

This is a fundamental issue with V combat and won't be changed for quite a while. Once we implement player evasion/absorption in v4 this might be different.

- early game, noone wants the rare out of depth heavy armour junk (unless to sell it....)
- not even warriors who suffer from only 1 of the 2 penalties; I find that the extra weight allowance from high str is much better used by carrying more staves of id and potions of ccw than by getting a bit of AC.

Is this still true, even though heavy armours are now twice as good as they were, relative to light armour? This was changed in 3.1.2 and there were quite a few positive comments (as well as the inevitable detractors). I remember seeing dumps with people wearing heavy plate for the AC, even if it cost them some mana etc.

- lategame, you cannot give the melee guy more protection than the ranged guy. Melee is intrinsically more hurting than ranged, but the only way to give some advantage to balance melee is to increase the damage compared to ranged.

Sorry, I don't understand this one. Is it related to armour weight, or something else?

[Estie]

For an evaluation of early game weapons, the big spikes are the very light weapons: dagger, MG, whip, rapier, perhaps trident. Cutlass is already too heavy for many blows. Simply adding a constant amount for those would make a good approximation. So have the base dagger 1d4 cost whatever it costs now +1k; then modify for enchantments as usual.

Of course it'd look a bit silly, longsword (+0,+0): 100$, dagger (+0,+0): 1000$, but heh were pricing for usefulness. More elaborate systems that take weight into consideration and smoothen the gap between light and non-light can be constructed, but I dont think its necessary. Starting with cutlass, the decrease of blows gets compensated for by increase in dice, more or less.

Improving AC on heavy armours has not done much in making them more valuable. The thing is that the threat to life and wellbeing comes mostly from elemental attacks, not physical.
Ive said this before, imo the game lacks archers. Im not sure how hard/easy this would be to program, but I imagine giving some monsters like, say, the uruk a limited ranged attack (maybe 10 shots, after that he becomes melee orc) that requires to hit and does high damage; not 1-hit high, but 2-3 hit. The ai uses the ranged attack for as long as its available.
This would btw go a long way in making the game harder :P

Current physical ranged attacks are a) weak and b) alternated by phases where the monster tries to get CLOSER into melee range; if anything, it should try to keep its distance, but I am not suggesting to go that far.



The other thing is that speed penalty in the early game is really harsh. If the choice is to take on a speed 0 melee orc unique with 30 AC and speed 0 or 80 AC and speed -2, I am not sure what I would pick, maybe the 80 ac is a slight advantage maybe not, but in all non-melee situations, the higher speed wins. So id definitely not lug the mail and rather phase a bit more for the orc boss.

Here is an idea: you could give characters an armour allowance for weight.
So if they pick up a 40lb mail, their weight and speed takes that into account as before. If they, however, equip said mail, the speed calculation substracts a certain amount from that weight. This can be easily justified as it is easier to carry a platemail worn than it is in a backpack.

The exact amount can depend on class, str, con, whatever. It should probably show up on the character screen somewhere, alas.
Note that the "weight" entry doesnt change; only the speed entry might.


Protecting the melee guy:

What I mean is that if both warrior and mage use the same platemail, they are defensively both on the same level. So the advantageous ranged attack of the mage must be balanced by doing less damage than the warriors melee attacks.

In classic D&D, mages are not allowed to use plate and thus have low ac. They have to avoid melee or perish.

Physical vs elemental damage aside, imagine this:

Suppose that we'd give all metal armours immunity where light armours have resists. So a leather of res fire would correspond to a plate of imm fire; furthermore, this would be the only way to get immunities.

Now add a harsh mana penalty for metal armours, so mages with immunities cant cast spells. In this situation, you have a vulnerable mage who has to do alot of avoidance work and a sturdy warrior who can shrug off damage to get toe to toe. The warrior doesnt have to have the best damage; but you lose universal usefulness of items: if Isildur drops, the mage is disappointed.

Doing exactly this would not be a good idea; I just wanted to show one way of making variable defenses. Currently, in lategame only offense varies.

[buzzkill]

Originally Posted by Estie

Suppose that we'd give all metal armours immunity where light armours have resists. So a leather of res fire would correspond to a plate of imm fire; furthermore, this would be the only way to get immunities.

There's something here. While probably not for V, what if we moved more toward shades of grey, so that heavier armours generally give better levels of resistance. Thus, a robe of rFire might typically grant 25% resistance, leather 40%, chain 66%, and plate 85%. Throw some variability and/or magical pluses (of *resistance), and maybe an immunity could occur occasionally. This might work quite well of FA-type cumulative resists, and maybe even merit a run in v4 with single equipment resists.

[LostTemplar]

Or simply do a huge buff to monsters melee so it can kill.
Adding archers will help also.

[Lord Tom]

This is a great discussion about the power algorithm and pricing weapons according to it; very interesting. Acknowledging that no system will be judged perfect by all...

I'd agree that pricing for a given weapon should not change with the character's state -- there's the abuse issue and I think it opens up even more variables to optimize in terms of pricing accuracy.

I think rather than pricing gear for a certain game scenario (eg "the endgame"), it'd be better to think of a rough average across a successful game's completion.

So light weapons are markedly more valuable early, with heavier weapons preferred later due to their higher dice. Since the heavier weapons' higher dice is already factored into their cost, just make a small upward adjustment for light weapons.

So call the average blows for a heavy weapon (say) 3, for a really light weapon (say) 4 -- players will get 5 blows with both in the endgame, but will reach 5 much earlier with the light weapon so the average blows count is higher.

I'm guessing that blows are included in weapon power calcs to calibrate them in comparison to damage-doing-non-weapons like wands, etc?

As a final thought, to improve maintainability perhaps monster experience could be calculated by a similar power equation taking into account hp, capabilities, etc?

[Magnate]

Originally Posted by buzzkill

There's something here. While probably not for V, what if we moved more toward shades of grey, so that heavier armours generally give better levels of resistance. Thus, a robe of rFire might typically grant 25% resistance, leather 40%, chain 66%, and plate 85%. Throw some variability and/or magical pluses (of *resistance), and maybe an immunity could occur occasionally. This might work quite well of FA-type cumulative resists, and maybe even merit a run in v4 with single equipment resists.

Ah, good - I'm glad you like that idea. Percentage resists are definitely on my list for v4, but probably not until we implement different damage types properly (see #1580). I'd support the proposal that the higher resist percentages are restricted to heavy armours.

EDIT: Oooh, this was my first-Intel-CPU'th post ;-)

[Estie]

Grats 4k Magnate

When I said doing exactly that wouldnt be a good idea, that was partly because a more staggered approach like % resists would be better. But the main issue is still that its a bit artificial - heavy armor should protect from blows, not from fire, right ? - however, it would be a way to go if the threat potential from enemy attacks stays so heavily focused on elements.

Thats the bigger issue that hefty archers would change. Unless that happens, I also dont see a split of evasion and absorbtion having much impact. Reducing incoming physical damage is low on the current priority list, so introducing alot of variety there seems a bit superfluous.