ACR Loot system

[ç i p h é r]

Ok, I'm really confused here. There isn't any need for builders to mark items as "non droppable" for the loot system to work. The system manages gear drops. Since it was designed in part with standards input, I'm really puzzled as to why they would direct builders to mark items as "non droppable". Is this just the result of not fully understanding how the system was going to work or was it an interim measure until a working system was put into place?

Anyway, I guess it's moot. The "how and why" of it doesn't really matter as long as everyone understands that marking items as non droppable isn't necessary (it should be reserved for items that a builder doesn't want PCs to ever possess). The question is, should we go back and make all items droppable or just force them to drop via script? Clearly no-gear-drops was not what anyone intended. And bear in mind, nothing greater than the calculated loot value will ever drop, so the difference between gear value and loot value is the effective value of the random loot generated.

[AcadiusLost]

Clearly no-gear-drops was not what anyone intended.

Actually...

So, hesistantly, I suggest we proceed along the lines of a CR:GP randomised lognormal deviating drop value, excluding all monster gear except ammunition...

We can't include total monster gear drops without wasting hours trying to balance stores and monster equipment or breaking the reward system.

I assumed simply directing builders to have their mobs set to not drop gear was the simplest solution, tho' central enforcement is always a bonus.

and

However, I guess a rough heuristic could be to just boost average value per CR mob, and add, say, a 75% "swipe" condition - i.e. a 25% condition of spawning a drop in the first place. That way you could do something pretty close considering those 10 orcs, due to the virtue of stochastics, maintain the average wealth, just not the same per individual orc.

It's a good thought, I'd also prefer an approach towards that direction, if it doesn't add a notable strain or burden.

These quotes from our discussion here:
http://www.alandfaraway.org/phpbbforum/ ... hp?t=36297

The Standards side of this discussion is here:
http://www.alandfaraway.org/phpbbforum/ ... sc&start=0

Some highlights on these points from the Standards thread:

I'd suggest just ignoring the equipment - imho, a critter does not always have to drop his gear. But giving on the actual loot tables of the critter a high chance to stuff that is his equipment - if the random roll of value supports it. So if you fight kobolds with short swords, not everyone drops a crossbow... but every once in a while one that got a value higher than 20gp drops a shortsword as part of his overall lootiness.

I suppose I might as well agree. Given our XP gain rates, we have to remove equipment drops. The only concern I have is that this may necessitate seperate loot tables being drawn up for every type of mobling, rather than being able to use more general ones.

as I posted there, got the idea that we can apply a heuristic to decrease the amount of clutter: Increase value of drops, decrease likelihood of drop occurring. If you triple n, but add a condition that will see 66% of drops not occurring - i.e. that there's on a chance of 34% that a drop happening - you have the same average, but a better bundling and an easier way to get equipment included in the drop.

The placeholder system was coded in regular communication with Standards, and the decision to block carried items/loot was not a spurious one, nor was it's implementation due to a failure to understand the principle of the system. The final conclusions of Standards were not entirely in line with the original spec, which is I believe, why you are feeling some disconnect here. You were a part of those discussions, though- it's just been the better part of a year since it was argued and resolved.

With regards to putting weapon drops back into the system we have in place now- we could use SetDroppable() on the weapon(s) if !GetIsPlotItem(), and plug them into the calculations as you've coded- but when we get to the test of "V > g" they are going to almost universally fail for most all our spawns, especially if we are dropping "V" 4-fold in order to run the calculations on every creature killed (instead of only 25% of the time). So, what happens, by the system? The weapon doesn't drop (again), and lower-value ABR loot is substituted- just like we have happening now.

Working through an empirical example, with my 68 killed ABR orcs from above. In my case only 20 got loot rolls, and of those 20, only 4 had "V > g", that's with the 4x multiplier in place (to go along with the 25% chance). So, under the fixed system, those 4 would be the only ones who would be able to drop their weapons. If we went to a loot roll for each creature killed (68 rolls) and lost the 4x multiplier, only 2 of the 20 would have cleared the "V > g" hurdle. Expanding that out, we'd be looking at only 6 or 7 of the 68 orcs who could potentially drop their weapons, the rest would have ABR loot of a value less than a standard weapon substituted. Also note that more than half of those orcs were >= CR 1.45, 30% of them were >= CR 2.5. So, we're not talking about a "low-CR mob only" problem here.

I guess I'm not seeing how that would make things better IC- it still doesn't solve the immursion frustration of being slammed with a warhammer that evaporates into thin air a few seconds later.

The main controls on droppability are in the creature blueprints for most of this, where you are given a checkbox for "droppable" for each item in the inventory tab, which applies the property to the instance of the item, rather than to the blueprint. If we do want to change this systemically, best to do it scriptwise, otherwise it's a huge toolset burden.

[ç i p h é r]

Well we certainly do have a serious disconnect. My discussions with Rusty about the loot system just prior to coding it included gear drops.

Here's the very latest spec I worked from:

Code: Select all

G = total gear carried by creature (not armour)
g = individual items of gear carried by creature (not armour)
V = calculated value of loot drop
L = randomly generated loot
D = actual dropped loot

D must never exceed V
Where G = V, L = 0.
Where G is less than V, L = V - G.
Where G is greater than V and g is less than V, L = V - g.
Where G is greater than V and g is greater than V, V = L = D.

As you can see, gear is very much a part of the calculation. I don't know how it evolved to this point, frankly, and I'm not sure it's worth piecing that together. The question is, is anyone not satisfied with this algorithm?

The benefit to having a centralized algorithm is that we can easily adjust it if need be. So, if V is > g less than 5% of the time, we can adjust V to increase the probability of gear dropping. Agreeing on what that number should be though is probably easier said than done, especially since g can vary a great deal from one creature to the next.

I agree with your proposed scripted solution to getting items back in. It's a good tradeoff between expediency and flexibility.

[Rusty]

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

In these circumstances ( G > V ), the drop shd be all randomly generated, no? Is this not working the cause of the problem?

[Mayhem]

It seems crazy to me that Weapons and Armour, 2 things you can actually see on the creature that is attacking you - two things that actively effect how dangerous it is - are being deliberately set as the things you will never find on their body.

[Rusty]

You should be able to find weapons, assuming that V > g(weapon).

[Veilan]

Indeed, I think the quotes AL highlighted - thanks - state pretty well that gear drops should be included, just, to make our mathematical model easier, not on every single mob, and not from the get go as we were still testing and developing the system. Gear drops are relatively big chunks of value in a single piece, usually too large for a single creature consistently, hence a modifier with a percentile chance of the gear dropping to get an approximation.

[Mayhem]

Indeed, I think the quotes AL highlighted - thanks - state pretty well that gear drops should be included, just, to make our mathematical model easier, not on every single mob, and not from the get go as we were still testing and developing the system. Gear drops are relatively big chunks of value in a single piece, usually too large for a single creature consistently, hence a modifier with a percentile chance of the gear dropping to get an approximation.

If a creature is armed with gear above its wealth value to the extent that you can't let the player who kills it take that gear, then it needs to be re-armed with something cheaper.

[JaydeMoon]

+1, in theory.

[AcadiusLost]

It would seem that the desire to make a given creature more likely to drop it's weapon is in conflict with the swipe/multiply transformation of Alara's that we're currently applying. Indeed, in order to keep the math working out, we actually need to insure that nothing drops on a majority of kills, in order to counterbalance the boosted loot roll on the remaining minority.

The simple way of thinking about it would be, if you were going to have to fight your way through a sewer tunnel containing 12 rusty dagger-wielding goblins, would you rather find 12 rusty daggers, or one CLW potion and 3 minor gemstones (assume for our purposes that the value was the same). The tradeoff is consistency versus greater dynamic range. If we don't plan to skip a certain percentage of the kills, the "ceiling" on each individual loot drop gets low indeed.

I'm agnostic on these issues, though I do kind of feel the rarer and bigger the drops, the more temptation there is to keep on trying for those who are hoping to come out ahead (in a "playing the slot machines" sort of fashion). So, I'd be fine with shifting the %/multiplier (as proposed in my initial suggestions) or dropping it entirely to make spawns drop only what they are "worth". Remember as well that a random normalized distribution comes into play, so unlucky low-CR rolls will still have their shortswords swapped for a battered beerstein with some frequency.

[Mayhem]

The simple way of thinking about it would be, if you were going to have to fight your way through a sewer tunnel containing 12 rusty dagger-wielding goblins, would you rather find 12 rusty daggers, or one CLW potion and 3 minor gemstones (assume for our purposes that the value was the same). The tradeoff is consistency versus greater dynamic range. If we don't plan to skip a certain percentage of the kills, the "ceiling" on each individual loot drop gets low indeed.

Conversely, if you barely survive fighting the orc that's battering the heck out of you with a greatsword, you have every right to expect to find that greatsword on its body.

If the weapons are worthless, your point stands, but in many cases they are not. A rusty dagger presumably has a loot value of zero, so would not figure towards the value of any treasure drop.

(zero value caluclated thus: Dagger 2gp, halved for rustyness, -1 gp, buy back 50% = 1/2 gp, rounded down to zero)

[AcadiusLost]

(assume for our purposes that the value was the same)

I didn't take the time to work out equal-value equivalents for the example because I thought the principle of the question was pretty clear. As I seem to have failed in that regard, I'll restate it.

All other things being equal, is it definitely worth getting lots of small somethings (which may or may not be what the mobs were wielding), if it means you'll never (or almost never) get a "high-roll" treasure item?

Making all mobs drop /something/ means the individual drops will be a lot smaller and less varied than we have now. That's the trade-off.

[Brokenbone]

Occasional "big" item has some appeal on the one hand, since it may be LESS likely to be liquidated, which automatically cuts a drop's value in half, or worse.

I.e., find 50gp worth of simple weapons... clubs, daggers, staves, other cheap stuff. Someone else finds a potion of CLW worth 50gp. One guy is likely to sell all his crap at the store (buyback yielding him maybe 25gp?), rather than carry a golf bag of redundant weapons, other will keep that 50gp piece of loot on his belt.

On the other hand, the "slot machine", keep playing til you hit it big factor, is also to be considered. Monsters are reliably worth more XP in NWN2 than in NWN1 (least that's my hypothesis even with broken, overly diminished dimrets, vs. NWN2 implementation), so they're already pretty tantalizing. If you have a relatively good idea that 1 in 10 or 1 in 20 clobberings means a healthy useful payout, behaviors may be influenced.

I think of the infamous and long-gone Blacktooth Orcs of NWN1 Shadowdale in support of that theory. Now that I understand the guts of that module, they were tied to a pretty diverse loot merchant, where luck of the draw usually would mean like, a fish, a rag, or other junk items, but sometimes would mean a ring of protection, leathers +1, full plate, many different potions or other expensive items, etc. Basically if you killed a field of 20 of them, at least one had the odds of being a dead Daddy Warbucks. Since the mod was unfixable under the then-existing management, certain terrain became "no visits absent a DM", on pain of ostracizing.

Anyhow, yeah, there's definitely a tension between "lots of reasonable little stuff" vs. "nothing sometimes, but big payouts to make up for it now and again." If one goes for a dry spell too long, I don't know if it'd just be human nature to expect "sooner or later I'm going to roll high and get compensated." Some kind of gambler's fallacy in there someplace (re: law of independent trials? Dunno).

[ç i p h é r]

Well, we can tweak the algorithm to periodically forgo all gear and generate a "big loot" item. That'll help keep things from becoming too predictable or uninteresting.

But going back to what Mayhem said, if the loot value (V) of a creature is less or significantly less than the value of the gear it's carrying (G), the loot drop formula seems fundamentally unfair. PCs in this scenario are obviously fighting a creature that can carry gear of greater value than players could ever recover, meaning that the risk/reward ratio is too skewed.

If this is indeed happening, I think there's a good case to adjust V upward (increase the multiplier, maybe even based on % deviation from typical gear value). Conversely, builders could equip creatures appropriately up front based on the relationship between creature CR and gear value. For simplicity, I think it's generally around 20 * CR, currently. This would be consistent with builders setting the CR properly of creatures they equip, but it's also least likely to get done or done right [every time].

[Veilan]

I'm not particularly wed to one system either, I did however believe the "swipe" system to allow for mobs to from time to time drop what they carry.

The assumption that you should earn everything the creature you fought upon defeating it is valid in CvC, but has no other place in a persistant world, where DMs do not have control over the number and kind of encounters PCs face. It's cool if you get gear a mob carries over time, and again, I think that's facilitated with a percentile multiplier system - however, the notion that equipment value of mob = loot value of mob has scant merit on its own.