Monster loot.

[LostTemplar]

I am about to design monster inventory for my variant, so I want some other people opinion and advice.

Currently I consider the following possible models for monster loot:

1) Angband-like, where drop is generated when monster dies, based on random choices.

2) Realistic model: monster drops and uses items, it have in inventory. (monster can carry also some useless for it treasure, but any item, it have can be used in combat, e.g. if monster have a wand, it may zap it)

3) Same as 2, but with some random, e.g. monster drop the weapon it uses, but with some chance (consider, that weapon may not fit the character or being broken in combat, etc.)

Since all the functions and data structures, related to items, monsters, and player character have been rewritten anyway (or will be). I dont have to stick with Angband legacy.

So I want some arguments from people here about what loot model you think the best and why.

[Therem Harth]

Options 2 and 3 allow for thievery, which I think is a good idea. Of those I think option 3 is better, if only because it feels more realistic (e.g. stuff a monster is carrying may break during a fight).

Re monsters using items, I really like the idea, with the caveat that you might want to limit item use to intelligent monsters.

[Derakon]

Pyrel plans to have monster inventories have their real items in them (i.e. monster drops are determined at the time monsters are created), and monsters that inherit from the appropriate templates will have equipment slots. Whether or not they're smart enough to equip or use items is a different matter.

"Monsters equip items" is a much harder sell, to me, than "monsters use items". The latter requires them to have access to an extra set of "spells" that derive from their item lists and use those items up as appropriate -- comparatively straightforward. The former requires you to decide how a monster wielding a weapon has their attack modified, what the "natural" AC is for a monster, what it means for a monster to get a STR bonus, etc. Plus it makes monsters that much harder to balance. All that for making the monster somewhat harder to defeat in ways that will be basically mysterious to the player, since the equipment a monster is using is not obvious (presumably unless you (l)ook at the monster).

[fizzix]

I think monster drops should be specified with the following fields: drop_type, probability, number, rating, usable

drop_type can either link to a specific item, or to a grouping of items. For example, you could give Sauron an X% chance of having the one ring. Or you can give a snaga a 30% chance of having a blunt weapon. Or you can give a gelatinous cube any item in the game since it's whatever it happened to roll over on the ground.

probability is the chance of the drop occurring.

number is the amount of the items that it drops. So the gelatinous cube could have 2d3 items or something like that.

rating allows you specific a drop to be good, or great, or even special. You could give some uniques a high chance to drop an artifact. There could be a lot more gradations here, for example, you could split this into two fields, one a broad categorization and the other a power rating.

Last comes usability. I imagine this would be unused for Angband, but I can envision variants making great use of it. For consumables this is straightforward. The monster uses the potion and recovers HP. Or it uses the wand, drains a charge, and causes an effect. In general usable consumables should be restricted. Giving novice mages any wand is probably too much, but giving them a 15% chance of having a weak wand, (select from magic missile, slow monster, stinking cloud, confuse monster) seems like it could be fun.

(how to decide whether the monster uses the charge or not, is a good question.) For armor and weapons, it becomes more difficult, but I think there's some good area to explore here. Imagine a situation where you've configured snagas to have an 60% chance of dropping orcish leather armor, and an 100% chance of dropping an orcish light weapon, both with the usable flag. The only time the player will see a difference is if the armor or weapon is magical. In this case the monster will get a bonus or penalty to their attack.

There's going to be some difficulty to sort out with excellent armor or weapons. what happens when the orc is wearing leather armor of resist fire and you cast a fireball on them. How will your monster memory update? Does that change if it's the first orc that you encounter? I have no idea how to answer that question. I guess you'd have to give the player a note, "the orc's armor glows" so the player knows that resist fire is granted by the equipment.

Of course you have to find someone to update the 700 or so monsters with all this data. Luckily, these are the types of things that I love to do...

[fph]

Nethack has monsters-use-items. It looks like a great idea until you meet a gnome with a wand of Death. (Yes, it does what the name says.)

[LostTemplar]

It looks like a great idea until you meet a gnome with a wand of Death.

I dont think, such a thing will be big problem.

Well, I have two central ideas for my variant, that causes rewriting of many things.

First is full scale monster vs monster fight (e.g. invaders vs local army, dragon vs village, etc.). And somewhat smart monster spawning function.

Second is a "player have no discount" principle, so player character is just one of monsters, allmost nothing special. Currently only player needs food, and only player can learn. (Monsters just spawn with appropriate experience level) anything else is symmetric, and uses exatly the same functions.

Also some lesser goals are:
I will try to make even the first level interesting, no boring start.
No random or many choices before the game even start.
Player's actions will have many long-term but not permanent effects on the game world.

Obviously it will be impossible to just go and slay everything.

My current local goal is to make monster vs monster fights look nicely.

Of course you have to find someone to update the 700 or so monsters with all this data.

Well I rewritten all the monsters anyway (edit file is crurrently ignored, I plan to use it in the future for descriptions only), so I currently have just 5 sample races (each monster have also level, so one race can fill the whole dugeon)

Every monster have a "weapon", however it is not an object now, just some info fields. some weapons, like claws will definitely be not droppable, others can be transformed into objects.

[half]

Imagine a situation where you've configured snagas to have an 60% chance of dropping orcish leather armor, and an 100% chance of dropping an orcish light weapon, both with the usable flag.

This would seem to be a situation in which each snaga drops 1.6 useless items (perhaps in addition to some useful ones). While it adds immersion, I doubt that it is worthwhile from a gameplay perspective. I always imagine that the monsters do drop additional items (rather than simply disappearing on death) but that they are mostly useless given your other items, so not shown to the player.

Sil has a couple of situations in which monsters drop themed items: enemies with artificial light (Easterlings, Giants) are more likely to drop a torch; enemies with bows are more likely to drop arrows. As both of these are consumables, they are less often useless and you can even try to hunt down the appropriate type if you need what it carries.

[LostTemplar]

drops 1.6 useless items

Btw why more useless then random drops ?

If e.g. 100 weapons are generated during the game , 99 of them will be useless, unless combat system is very complicated.

Realistic drops actually strongly reduce number of useless items, e.g if you want a bow, go kill an archer, if you dont want a bow, avoid archers, etc.

Well, there is a separate issue, how to make more then one weapon per game usefull, there are a lot of possible solutions, most obvious is that weapons break a lot and / or obsolete fast. Or add some other then wield uses for weapons (sell, melt, sacrifice, set as a trap, give to ally, etc.)

My original thought about my option #2 was an 'anti farming' behavior, So, just one powerfull, nicely armed monster kill is enough to completely equip character, killing additional monsters barely helps, if you want better equipment, kill stronger monster, etc.

More ore less, so quality of kills matter more then quantity (for equipment).

[Scatha]

Btw why more useless then random drops ?

If e.g. 100 weapons are generated during the game , 99 of them will be useless, unless combat system is very complicated.

Realistic drops actually strongly reduce number of useless items, e.g if you want a bow, go kill an archer, if you dont want a bow, avoid archers, etc.

Well, there is a separate issue, how to make more then one weapon per game usefull, there are a lot of possible solutions, most obvious is that weapons break a lot and / or obsolete fast. Or add some other then wield uses for weapons (sell, melt, sacrifice, set as a trap, etc.)

A lot of roguelikes make more than one weapon per game useful by having some be better, and some be available earlier. If the player is expected to upgrade their weapon around five times over the course of the game, that means six weapons have been useful. Having a combat system which means that the weapons aren't totally ordered can significantly increase this number (note that such a system needn't be very complicated, it just requires you to spend your complexity on a specific goal).

Say you expect 10 weapons to be useful over the course of the game. Then an easy way to increase the proportion of found weapons which are useful is to decrease the number of found weapons. This is likely to decrease the number of useful upgrades, but not by very much. As an extreme example, if one expected to find just 10 weapons in the game, perhaps 8 of those would function as an upgrade at the time found. I don't recommend going that extreme, but generally reducing the number of drops total will increase the average usefulness of drops.

It's easy to combine this with flavoured drops if you want: just make archers more likely to drop bows and arrows than anything else, without increasing the likelihood that they drop anything at all.

Also, to make sure that these flavoured drops aren't more useless than random drops you probably want to avoid putting in a cap on how good the items can get.

[LostTemplar]

Yes, I mean number of weapons, used is of order of one.

note that such a system needn't be very complicated

I thought about weapon swaps here, so in complicated combat system it may be worth to carry e.g. 10 weapons.

Well, I will have upgardes anyway, since all items will have quality (in general percentage of how good it is, compared to the perfect item), also items will have different types.

I can even strictly never drop anything, that is worse then best item, already created (however this will look silly, and eliminates the use of identify), also option "squelch anything worse, then I already have" may be added simply. So too much junk will not be the problem definitely, however I like some junk.

The real problem is about allowing 100% drop of certain items or not.
E.g. it is logical, that one archer kill will give you a bow. However it may be 'too easy' gameplay-wise. (too much bows problem can be easily handled, so not worth considering).

[fizzix]

I've been of the opinion that the solution to useless drops is a good squelch system. If the player can trivially choose to ignore all the normal orcish armor, then those drops won't be annoying. In a game that forgoes squelch, like sil, you definitely need to be cognizant of too much junk making the game annoying.

At the very least, the point of figuring out how to deal with monster items in a new project like pyrel should err on the side of allowing as much latitude as possible to the user (read game designer)