ACR Loot system

[AcadiusLost]

OK, took a bit of time and did some testing with debugging code to the DM channel from this script, and I've gleaned some observations.

Killed: 68 ABR orcs, from CR0.5 to CR3.5

20/68 made the 25% loot chance (29.4%, reasonable)

of those 20, only 4 actually dropped loot.

Of those 4, only 1 dropped some of it's loot as gold.

Total gold dropped: 2gp on one of those orcs, a CR2.5 one, even.
(other loot included a rusty bastard sword, 2 heavy crossbows, and some near-valueless junk)

What happened with the other 16? The loot system is totalling carried wealth on the dying creature, and testing it against the random normalized loot value roll. If the creature wealth is higher, no loot is dropped.

CR0.5 Orc: calcwealth=16, lootroll=14, no drop
CR1.5 Orc: calcwealth=60, lootroll=47, no drop

This is meant to keep carried gear and weapons which drop at death from being added on top of ACR loot. However, none of these items factoring into the wealth calculations are set to drop. So, the "phantom dissapearing loot" is either nullifying, or in the best case, subtracting from the generated loot.


Another unanticipated quirk of the 25% flat percentage comes from mixed groups of enemies- if you encounter 3x CR0.5 kobold minions, with one CR3 kobold barbarian, you will most likely get a loot drop from at least one of them. However, 75% of the time, that drop will be from a minion, and the higher-CR loot reward becomes rare indeed. You've still faced a higher-CR challenge, but the chances of getting rewarded for it decrease.

The main intentions of the 25%/4x drop value calculation were to avoid fields of low-to-worthless valued items cluttering up areas during any large, low-CR combat (a frequent occurance in PWs). Also, we wanted to push up the ceiling on individual awards, so more useful items like potions could drop occationally from low-CR targets. We could shift things to a 33%/3x value, but this is just another arbitrary stab at a good balance.

Alternatively, we could look at skewing the %/x relationship by CR- as CR increases, have the multiplier decrease until it reaches a 100%/1x level at an arbitrarily significant CR (5? 9?) by which time no sane DM or builder will be spawning floods of such beasties.

Last point- currently humanoids have a 25% chance of dropping between 1 and 25% of their loot as gold. In practice, this means gold drop becomes a rare event (6.25% of the time) and is just as likely to be in the 1-3 GP range as more significant levels, due to the random 1-25% range.
I'd recommend putting the GP chance for humanoids up to 33%, and run the range from 5% to 35%, so gold, when found, is less often like 1-3 gp.

I'm interested in feedback on this from Standards- none of my suggestions change the total value of loot drops proposed and approved earlier, though changes in the % dropped as gold do have some affect on PC wealth, as they don't have to go through a 50% or less buyback mechanic.

In the meantime, I'll try to figure out if we can tell whether an item in a creature inventory is going to drop or not via scripts, so that subtraction and comparison function can be amended.

[AcadiusLost]

Minor update, looks like the GetDroppable() function is working for our purposes, so I'll commit and switch TSM over to a fixed version for the major bug fix (subtracting nodrop inventory items as if they were going to drop).

We can hold on the rebalancing ideas (%/multiplier changes, gp% changes) until some discussion has transpired, in the meantime though, I think the priority is getting the loot system working again.

[indio]

Agreed. Nice one.

[AcadiusLost]

Droppable fix committed, and applied to TSM. Will have the 1.48 ACR hak on the worldgate in the next few days.

[darrenhfx]

Thanks AL, this should help fatten up teh lewts.

[ç i p h é r]

Good fix, but I'm curious why there are so many "non droppable" items on creatures?

[AcadiusLost]

The way I understand it is: if all low-CR mobs dropped their weapons/shields, the value of them would exceed what is allowable by our current loot/wealth standards. (It would also clutter up areas with loot bags from any large combat). I know it was the original intention to force full drops in the interest of fairness, but the practical side (at least with our current value formula/standards) didn't work out.

Also, having everything drop the gear it had available for use is incompatible with the 75% swipe, 4x value transformation that is currently being used to push spawn drops into ranges where consequential items can drop (healing potions, etc).

We can still amend this as we go along; return items to droppable states, drop them even if they're above loot value (but flag it in the logs, for example), and only roll for random loot with any amount above and beyond the blueprint drops is left over. This would also mean going to a 100%/1x model I'd think, and losing the dynamic range allowed by the 4x multiplier. I'd think this would make more sense if we decrease the availability (and predictability) of spawns for farming, but as it stands now on TSM, you don't have to look far to find packs of monsters.

As an aside, is there a canon basis for nullifying loot rolls on all undead? Seems to me it'd only really make sense to block for mindless low-CR undead (skeletons/zombies)- maybe I'm wrong on this one. We do seem to have an awful lot of undead on TSM.

[ç i p h é r]

So in other words, TSM builders are marking gear non droppable to encourage more random loot? Is this philosophy being applied universally? It just seems to me to circumvent part of the design of the loot system - and whatever arguments lead up to it - which is to allow gear drops up to loot value. The two will almost never be exactly equal, so some random loot is at least possible even with gear drops.

About Undead, yes there is canon basis for restricting drops (see Monster Manual), but I don't think there's a universal rule. Certain types of Undead might be exceptions - like Vampires for instance - but for the most part, they are mindless creatures and don't covet material things. As a consequence, they don't drop any treasure, just the gear they happened to have on them when they died. Marking their items as "non droppable" effectively makes Undead no-lewt creatures.

p.s. Purely random loot drops are going to produce really odd results if most creatures don't have any gear to drop. Creature specific drops may be needed sooner rather than later.

[AcadiusLost]

I realize it's been a while since this was addressed. As a reminder, it was made quite clear that builders were not to allow full gear drops of any sort, since they'd throw the system off balance.

We're also using a transformation currently that forces 75% of all spawns to drop no loot whatsoever (random chance) in order to boost the drop value by a factor of 4. (Alara's idea, which was supported by Rusty and others).

If we have all spawns dropping their weapons (which was my recommendation), we can't really model that anymore by the same formula.

The system we have is the system Standards wanted. Does it make the best sense? Hard to say, but it's certainly been through the ringer for compliance all along. Can't really blame the builders for doing things in exactly the way they were instructed to. It would be a possibility to force a switch of the droppable flag on weapons (or just equipped weapons), but what happens when that item value is greater than the normalized random generated loot value for the spawn? You can't drop half a crossbow, and if you drop the whole, you're throwing off the curve. If we drop the 4x anchor that goes with the 75% swipe, it'll be even less likely that any given spawn is going to be able to "afford" it's own weapon.

In the meantime, I've found a mistaken use of RACIAL_SUBTYPE_BEAST (instead of RACIAL_TYPE_BEAST) tested against GetRacialType() that's giving all the animals onserver ABR loot- so I'll get that fixed for now, would be nice to know ASAP if there are other things that need to be fixed while I'm in there.

[indio]

I haven't been part of the standards team's development of loot drops, but the easiest solution to this is to simply drop only gold. Forget mundane items and mob equipment (realism be damned when it lags the server) and just drop gold relevant to the strength of the mob.

[JaydeMoon]

I tried suggesting that. BB had reasons why we can't do this. I was content to let it go at that, however read something in the Wealth per DM Run Session standards that, along with what Indio asks above, prompts me to bring it up again (italics are mine):

Where...
Average XP Awarded: This refers to all sources of XP for a session, including combat XP and special rewards. It is the average across all party members.

So standards states that the Wealth tables in these standards applies to combat XP.

However, I don't know if the DMs have the ability to remember exactly how many spawns a player may have fought whether they got a lot of XP for it or a little, or different amounts depending on who was what level, or sliding scales, what have you.

Best if the monster just figured all that out for you and dropped loot on the spot.

The trick is making it so that they don't drop a static sum (like 5gp for EVERY kobold). Otherwise you get lvl 10 PCs hunting kobolds for the gold.

The trick is having a script that somehow determines exactly how much XP was awarded by the death of the kobold and giving out a sum of gold based on the levels of the players.

Again, it may not be technically feasible.

An example is:

1 kobold vs 1 lvl 1 PC. XP is 4. Gold should be 5 (4x1.25).
1 kobold vs two lvl 1 PCs. Each PC gets 3 xp. Gold should be 7 or 8 (6x1.25 rounded up or down).
1 Kobold vs lvl 8 PC. PC gets 0 XP. Gold should be 0 (0x1.25).

Vampire vs 2 lvl 9 PCs. Each PC gets 15 xp. Gold should be 75. (30x2.5)

Vampire vs a lvl 7, a lvl 9, and a lvl 11 PC. Lvl 7 gets 5 XP, Lvl 9 gets 10 XP, lvl 11 gets 5 XP. Gold should be 50. (total is 20 XP x 2.5 (multiplier for median lvl of 9).

Etc.

Maybe that would cause a lot of lag or something, I don't know.

[AcadiusLost]

We had a system that worked entirely in gold briefly, that was my placeholder until the proper loot system could be put in. It's nonideal for a number of balance reasons:

1. Though the net worth of 30 gp vs a 30-gp-longsword is identical, the latter will sell for (at best) 55% of that value with a maximum merchant and maximum appraise roll. So, switching to a straight-gold system actually doubles effective reward- and focuses the reward 100% into effective, useable wealth towards whatever topshelf item PC X is saving up for. I know we're not in the business of thwarting "success" here, but however you slice it, it's a big change to reward levels. [edit: this also means the random loot may be more effective given to others, or used, than just sold- I'd call that another plus for random loot.]

2. Gold is weightless, rusty scythes aren't. If a hypothetical farming PC is running a loop of spawn points to build up loot, there is never a need to haul back to town to sell, or choose to pick up Q instead of R because Q is lighter. A PC with 87 rusty weapons in inventory has clearly been killing and looting a lot of loot-dropping mobs. A PC with lots of gold could have gotten it any number of ways, can't really tell at a glance what's up with them. [edit: also means store gold limits, buy limits, and potential investment in Appraise/diplomacy skills all fail to affect the ability of a PC to accumulate wealth]

3. IC immersion. Hit the monster, coins drop out. This just doesn't work for me. While the oddities a random loot system bring are sometimes also disruptive, at least they're interesting and can lead to some RP, or even just conversational topics, conjecturing about why these kobolds might have been carrying a lock and a shovel, etc.


On loot by XP earned:

With regard to balancing loot value to XP earned instead of CR, this is potentially do-able, but problematic IC- because XP is calculated based on challenge level, and subdivided proportionally by character level, that means if we go on an individual level (award based on how much XP the PC who landed the killing blow got for the combat), it would heavily reward soloing (the XP isn't divided among the party, ergo higher XP, higher loot). If we base it on the total challenge vs party, that means whenever Joe the Higher Level Character is around, all loot drops dry up. The engine spots mr. 5th level and decides the kobolds weren't enough challenge collectively to warrant any loot. I'm not sure this is at all desirable, and since it's a change in an in-game, in-character factor (how much sell-able gear was on those orcs), it's something that is more likely to affect how people RP.

("I'm going out hunting, I may be some time. Don't follow me, last three times I took you with me, I found nothing of value. The Gods must be displeased that I am relying on the aid of others." )

[Brokenbone]

Just a point to JM re: "Wealth per DM Run Session", you do take combat XP into account.

But also key is that it's a DM Run Session.

I.e., if I send PCs off into the woods to solve a mystery, and part of that adventure is "killing a lootless wolf pack which results in some instant combat XP", which is aside from the DM award at the end, I'm entitled to take into account ALL that xp when making an award from say, a grateful employer, or a future bank. I.e., I plan to award 200 for the couple hour session, the wolves generated a bunch of onesie / twosie amounts for say, 50xp between the whole pack, that'd mean I can base an award on 250xp (though again, me as the DM waved a wand for 200, and other scripted systems did the rest).

...

Anyhow, I wasn't involved in the invention of factors, percentages, or the rest for the ACR Loot system in NWN2. It sounds though like some very alert and detailed troubleshooting is going on (i.e., noticing that undropped gear was counting against a "greater/lesser" gold figure, PLUS fixing it), where I can pretty much only say "great thinking and well done."

...

Mild specific points: animals, mindless undead, and many other critters have fairly clear Monster Manual entries that describe their non-acquisitive natures. Cipher seems to have already ID'd that. Of course, this is something meant to apply to spawns... a DM could easily put together a little session where some animals happen to be enjoying the remnants of a dead (but geared up) adventurer, or where a bunch of mindless skeletons all happen to be sealed into a tomb, along with fabulous treasure and plunder (ie, it's not "theirs", they being quite beyond caring about material things, but you're going to have to get past them to get it).

[JaydeMoon]

1. Then make the drops = 50% of that amount. (XPxFactor)/2.

2. We added weight to gold in NWN1, is this not possible in NWN2?

3. IC Immersion: That's just as broken if you hit the monster and the items I see it carrying do not drop out. If we want to go for IC Immersion, then what it is visibly carrying, it should drop. Got a knife? Drop a rusty knife. EVERY time.


On the last, the engine already spots that for XP and thus we already have that problem. We don't know that it will be further exacerbated by tying loot to it.

The engine spots mr. 5th level and decides the kobolds weren't enough challenge collectively to warrant any XP. I would think that the people that would be stopped from partying up because loot will be minimal are the same people who are already soloing for XP gains. Maybe not *shrugs*

The long and short of it is that there is no completely IC way to do this that accounts for all of the different factors here and there. Either your immersion is broken because they drop only gold. My immersion is broken because only one in 4 drops a rusty dagger when ALL of them were weilding rusty longswords. The best compromise is to work towards ensuring that the balance is maintained.

[AcadiusLost]

I've already said, I favored having the weapons drop all the time. It was my recommendation, as that's the part that's seen (and felt) IC. What I'm presenting is the rationale and balance concerns that went into the Standard decision (as best I understand them). Maybe it's a point on which we could say "If a weapon is over the wealth roll for a creature, it's OK, let it drop anyway". What is needed, though, is a consensus in Standards over how it should be done, because I don't want to be changing the code (and hence the loot drops ingame) back and forth every time someone's got a "better idea".

If we make weapons drop all the time, we should also drop the 75% swipe 4x multiplier- means most everything kill will drop something, but the average value of the something is universally going to be lower. And it means carrying a lot more somethings around for the same amount of combat. I was going to say this would mean we'd stop seeing potions drop on all but the highest-CR spawns, but then I followed a hunch and checked- sure enough, there are no potions on our ACR loot table anyway.

Gold weight: Seemed like this system was pretty universally disliked in NWN1-ALFA, this is the first I've heard of an interest in going back to something like that (499 coins= weightless, 501 = 5.5 lbs instantly added to your PC). It also gets us into tradebar debacles and the need for unexploitable gem merchants, appraisal scripts... the list goes on.

In terms of XP altetations when working with others: this is the way I see it. Gold, gear, gems- these are all in-character qualities. Your PC sees these things, has a legitimate IC interest in them. They can and do get divied up in-character. Your PC is aware of them. XP, on the other hand, is entirely metaknowledge. One can try to fake some kind of bizarre pseudo-IC justification ("Obi Wan is holding you back! You should be farming padawans without him if you are to progress, or you shall never escape his shadow") but the actual math and number crunching on XP shouldn't be anything any PC should be doing in-character. Night and day difference, for me at least. I'd hope it is so for everyone who plays here.

I'm not defending the loot system we have now as ideal, far from it. I simply don't have the time and interest to endorse and implement anything complex on a whim. We've got a huge list of things that need to be fixed or developed that seem a lot higher-priority than starting from scratch on one of the few systems we've got (more or less) working.

I gave some suggestions of changes that could be made simply to the system (as in ~10 minutes' work) that I thought might improve things without changing the long-discussed much-argued, then finally "resolved", total drop values. These were ways to step "sideways" that might make the system more satisfactory to the average player.

Those suggestions got no feedback- instead we've got one person who seems to be bristling at the builders doing what they were told (setting nodrop weapons), suggestions to bin the whole system and go back to straight GP drops, and discussion of changing the fundamental base metric for the value generation to take into account challenge level. This sort of argument is why I never wanted to be a part of Standards. The only meaningful input I ever have to contribute is to give reality checks when suggestions are made that simply aren't practical or possible with our technical resources.

At this point, I'd say Standards can start their own thread to go back and forth about how much when and why for rewards. I'd lock this thread in the meantime, but it seems the ACR staff head doesn't have moderation rights on the ACR forum.