ToaLewa
BdW-Veteran
Treasure System Details
A detailed explanation on how the Treasure System functions and a clarification on "different treasure".
Summary
Over the past few weeks, we have been reorganizing our treasure profiles for a variety of different creatures in order to help streamline the system and make it easier for us to add additional items when and if necessary without having a drastic impact on the items already in these treasure groups. This has been reflected in our Release Notes as "Different Treasure." Some players have started to perceive this as a "nerf", when in reality, the prime goal has been reorganization. [more]
Due to the way our treasure system is implemented, everything is broken into groups, which have a percentage change of triggering, and everything in that group then has a percentage chance to drop a set number of times. When you have multiple treasure groups and multiple sub groups, these numbers can get quite overwhelming and even the smallest change can have a large impact on all other items.
As anyone who has taken a math class can tell you, adding percentages of percentages against other percentages of percentages to gain a new percentage can get quite confusing--much like this sentence! Our new system merges many of these subgroups and thus makes it much easier for us to manage and modify accordingly as we move forward and add additional items into the treasure system when needed.
This system can be somewhat complicated, so if you're feeling up to it, please continue for a larger, more detailed breakdown of how the treasure system functions.
Details
Treasure is organized into 'treasure groups.' Treasure groups are used for anything that generates treasure, primarily monsters and quests. A treasure group is a listing of all the items that that group can drop; each one with a weight assigned to them that controls the likelihood of that particular item being selected. If I have a treasure group with just two items in it, and give item A a weight of 100, and item B a weight of 50, that means there's a 66% chance item A will be selected (twice the 33% item B would have). Simple, right? It gets more complicated when you assign the group to an actual monster. Each monster can have as many treasure groups assigned to it as you want, and each of those groups gets a weight of its own. In addition, there's a control for the maximum number of times a given group can be used, and a setting that controls the total number of items that will drop. So, in a simple case, let's say we have a SimpleMonster, who is set up with two treasure groups configured like we described above. Group A has a weight of 100, group B has a weight of 50. What are the odds of getting a given item?
Group A, Item A: 44.435%
Group A, Item B: 22.217%
Group B, Item A: 22.217%
Group B, Item B: 11.108%
Already I'm breaking out a calculator, and this is a very simple situation. Now, I mentioned the goal here has been primarily reorganization, and it is. As items were being introduced throughout beta and things were being added/removed very heavily, people kept adding new treasure groups and spreading them out through the various monsters. The groups used were organized around what that group dropped, so you have a very big list of groups, even with the fairly sparse drops we currently have: "Crystals 20-30", "Techniques 10-15", "Raw Resources 10-20", "Wolf Fangs", etc. There's probably 200-300 of these in total, each of which contains that list of items with weights, ranging from a single item ("Wolf Fang") to dozens ("Crystals 20-30").
When you look at a typical monster with all this assigned to it, things get pretty hairy. Here's an example of a typical monster (a level 30 Saris Ghost) using these groups:
WEIGHT NAME
25 "Lvl 40-44 Dropped Techniques"
20 "Crystals Lvl 20-39"
1 "Crystals Lvl 40-59"
5 "Imperial Bounty Marker"
15 "Imperial Lore Token"
60 "Lvl 40-44 Blighted Processed Resources"
100 "Lvl 40-44 Blighted Raw Resources"
60 "Lvl 40-44 Processed Resources"
100 "Lvl 40-44 Raw Resources"
200 "Ghost Vapor Saris"
5 "Soul Fragment"
--------------------------------------------------------------------------------
591 XXXX
Now, remember each of these groups has between 1 and 50 items inside of it, each of which has its own weight to look at. So, the odds of a Soul Fragment dropping are 0.84%. The odds of a Lore Token dropping are 2.538%. The odds of a Dark Garnet Weapon Crystal dropping are 0.06%. The odds of Saris Ghost Vapor dropping are 33.8%.
Now, this isn't that big of a deal when you're analyzing a single monster. But there's no guarantee these same groups or settings are there for the next monster. Maybe the level 21 Saris Ghost has different settings than the level 20 Saris Ghost -- after all, this is a pretty big list to set, and all these were being added and changed constantly during beta, by several different people. The real problem is when you want to make a change -- say I want to add Imbued Armor Splinters to Saris Ghosts. There's two ways to do this. First is to add Imbued Armor Splinters to one of the existing groups, then I won't have to touch any of the monsters (and risk typoing/breaking something). The problem is there's no real system to what groups are on what monsters, and in fact no guarantee that the group I edit isn't being used for a quest or other treasure-generating system as well. That's pretty unsafe given my goal -- I stand a decent chance of messing something else up when I add this.
So, I'm probably going to make a new treasure group called "Imbued Armor Splinters" and add it to the list of all the saris ghosts. Say I give it a weight of 50.
Now the total of all the weights here is 641. All those percentages I just quoted got smaller by 9% or so to make room for the armor splinters. On top of that, I should go through any ghosts I add splinters to one by one, because maybe one of them has got a much smaller total weight than any others, it's possible a weight of 50 on that monster means they'll drop 75% of the time.
For example, here's a Saris Skeleton, level 10:
WEIGHT NAME
25 "Lvl 10-14 Dropped Formulae"
5 "Imperial Bounty Marker"
15 "Imperial Lore Token"
60 "Lvl 10-14 Blighted Processed Resources"
100 "Lvl 10-14 Blighted Raw Resources"
100 "Lvl 10-14 Raw Resources"
5 "Soul Fragment"
25 "Undead Fragment"
--------------------------------------------------------------------------------
335 XXXX
Woah! So here we've got something a third the level in game, but the percentages are radically higher for the rare stuff. Those numbers for the Soul Fragment, Bounty Markers, and Lore Tokens sure look familiar, don't they? Odds are they were globally assigned and got caught just the way I described. The odds of getting lore tokens from this level 10 skeleton are nearly twice what they are from the level 30 Ghost, at 4.47%. This one doesn't drop crystals at all, which is probably fine at level 10, but it also never drops processed resources for some reason (probably just a mistake, most things do).
Basically what I'm saying is, this is very difficult to maintain, and considering that most monster resources end up being assigned to all monsters of a particular race, and most monsters drop level-appropriate resources, and crystals, and so on, there's a ton of duplicate information. Another monster of the same level as our saris ghost probably has most of the same groups, just with "Werewolf Bone" or something instead of "Ghost Vapor Saris". Considering that there's 3300 monster templates to maintain, this much duplicate information causes a lot of confusion and trouble, particularly since changing any of it could modify quest rewards or other things somehow.
So, I've been reorganizing these. Now any given monster has, generally, two treasure groups: one common group (containing all the misc stuff that any given monster may drop -- crystals, resources, etc) and one monster-specific group (containing all the things that particular type of monster drops). This reduces our Saris Ghost to this:
WEIGHT NAME
100 "Monster_General_030-039"
50 "Monster_Ghost_Saris"
--------------------------------------------------------------------------------
150 XXXX
Here, the odds of a Saris Ghost Vapor dropping are 24.23% (Because there's several items in Monster_Ghost_Saris). The odds of a Dark Garnet Weapon Crystal dropping are 0.66%. These numbers shifted because the total number of items in the various groups changed, and because the amount of overlap between the old groups and the new groups varies (In this case crystals, in the general group, became more common -- they were extraordinarily rare in the previous group list). Other things, like dropped quests, got added to the Saris Ghost list in the last few weeks.
Even regardless of odds, it's must easier to keep consistent than the old groups were. More importantly, it means we can start introducing dropped items with some degree of confidence. If I want to add Imbued Armor Splinters to Saris Ghosts, I can do that without modifying 120 monster templates -- I edit a single treasure group instead, and can easily see the odds I'm changing as I do it (since they're largely in one place). If I want to tweak drop odds, I can run through and increase the number of saris-ghost specific drops without doing a ton of addition to see what my result is. In addition, I know that these groups are only being used for monsters, so I don't risk breaking quests/tasks if I make changes.
At this point I've reassigned the treasure groups on the majority of the game's monsters, though there's still a few remaining using the old disorganized groups (Zombies, in particular). As I've gone through, I've tried to keep monster-specific drops to a relatively similar chance, with some stumbling here and there (Golem fragments, mostly, which brought to light the bizarre number of techniques using them, including armor). The payoff of all this will be dramatically less breakage-prone and easier ways to add interesting loot to our monsters when opportunities to do so arise.
If you made it this far through my essay, congratulations. I hope I went at least a little ways towards explaining what "different" actually means, and why it's an appropriate word for the job. This reorganization has already opened the door for easier bug fixing and treasure additions, and hopefully will enable us to make exciting additions in the future.[/more]
A detailed explanation on how the Treasure System functions and a clarification on "different treasure".
Summary
Over the past few weeks, we have been reorganizing our treasure profiles for a variety of different creatures in order to help streamline the system and make it easier for us to add additional items when and if necessary without having a drastic impact on the items already in these treasure groups. This has been reflected in our Release Notes as "Different Treasure." Some players have started to perceive this as a "nerf", when in reality, the prime goal has been reorganization. [more]
Due to the way our treasure system is implemented, everything is broken into groups, which have a percentage change of triggering, and everything in that group then has a percentage chance to drop a set number of times. When you have multiple treasure groups and multiple sub groups, these numbers can get quite overwhelming and even the smallest change can have a large impact on all other items.
As anyone who has taken a math class can tell you, adding percentages of percentages against other percentages of percentages to gain a new percentage can get quite confusing--much like this sentence! Our new system merges many of these subgroups and thus makes it much easier for us to manage and modify accordingly as we move forward and add additional items into the treasure system when needed.
This system can be somewhat complicated, so if you're feeling up to it, please continue for a larger, more detailed breakdown of how the treasure system functions.
Details
Treasure is organized into 'treasure groups.' Treasure groups are used for anything that generates treasure, primarily monsters and quests. A treasure group is a listing of all the items that that group can drop; each one with a weight assigned to them that controls the likelihood of that particular item being selected. If I have a treasure group with just two items in it, and give item A a weight of 100, and item B a weight of 50, that means there's a 66% chance item A will be selected (twice the 33% item B would have). Simple, right? It gets more complicated when you assign the group to an actual monster. Each monster can have as many treasure groups assigned to it as you want, and each of those groups gets a weight of its own. In addition, there's a control for the maximum number of times a given group can be used, and a setting that controls the total number of items that will drop. So, in a simple case, let's say we have a SimpleMonster, who is set up with two treasure groups configured like we described above. Group A has a weight of 100, group B has a weight of 50. What are the odds of getting a given item?
Group A, Item A: 44.435%
Group A, Item B: 22.217%
Group B, Item A: 22.217%
Group B, Item B: 11.108%
Already I'm breaking out a calculator, and this is a very simple situation. Now, I mentioned the goal here has been primarily reorganization, and it is. As items were being introduced throughout beta and things were being added/removed very heavily, people kept adding new treasure groups and spreading them out through the various monsters. The groups used were organized around what that group dropped, so you have a very big list of groups, even with the fairly sparse drops we currently have: "Crystals 20-30", "Techniques 10-15", "Raw Resources 10-20", "Wolf Fangs", etc. There's probably 200-300 of these in total, each of which contains that list of items with weights, ranging from a single item ("Wolf Fang") to dozens ("Crystals 20-30").
When you look at a typical monster with all this assigned to it, things get pretty hairy. Here's an example of a typical monster (a level 30 Saris Ghost) using these groups:
WEIGHT NAME
25 "Lvl 40-44 Dropped Techniques"
20 "Crystals Lvl 20-39"
1 "Crystals Lvl 40-59"
5 "Imperial Bounty Marker"
15 "Imperial Lore Token"
60 "Lvl 40-44 Blighted Processed Resources"
100 "Lvl 40-44 Blighted Raw Resources"
60 "Lvl 40-44 Processed Resources"
100 "Lvl 40-44 Raw Resources"
200 "Ghost Vapor Saris"
5 "Soul Fragment"
--------------------------------------------------------------------------------
591 XXXX
Now, remember each of these groups has between 1 and 50 items inside of it, each of which has its own weight to look at. So, the odds of a Soul Fragment dropping are 0.84%. The odds of a Lore Token dropping are 2.538%. The odds of a Dark Garnet Weapon Crystal dropping are 0.06%. The odds of Saris Ghost Vapor dropping are 33.8%.
Now, this isn't that big of a deal when you're analyzing a single monster. But there's no guarantee these same groups or settings are there for the next monster. Maybe the level 21 Saris Ghost has different settings than the level 20 Saris Ghost -- after all, this is a pretty big list to set, and all these were being added and changed constantly during beta, by several different people. The real problem is when you want to make a change -- say I want to add Imbued Armor Splinters to Saris Ghosts. There's two ways to do this. First is to add Imbued Armor Splinters to one of the existing groups, then I won't have to touch any of the monsters (and risk typoing/breaking something). The problem is there's no real system to what groups are on what monsters, and in fact no guarantee that the group I edit isn't being used for a quest or other treasure-generating system as well. That's pretty unsafe given my goal -- I stand a decent chance of messing something else up when I add this.
So, I'm probably going to make a new treasure group called "Imbued Armor Splinters" and add it to the list of all the saris ghosts. Say I give it a weight of 50.
Now the total of all the weights here is 641. All those percentages I just quoted got smaller by 9% or so to make room for the armor splinters. On top of that, I should go through any ghosts I add splinters to one by one, because maybe one of them has got a much smaller total weight than any others, it's possible a weight of 50 on that monster means they'll drop 75% of the time.
For example, here's a Saris Skeleton, level 10:
WEIGHT NAME
25 "Lvl 10-14 Dropped Formulae"
5 "Imperial Bounty Marker"
15 "Imperial Lore Token"
60 "Lvl 10-14 Blighted Processed Resources"
100 "Lvl 10-14 Blighted Raw Resources"
100 "Lvl 10-14 Raw Resources"
5 "Soul Fragment"
25 "Undead Fragment"
--------------------------------------------------------------------------------
335 XXXX
Woah! So here we've got something a third the level in game, but the percentages are radically higher for the rare stuff. Those numbers for the Soul Fragment, Bounty Markers, and Lore Tokens sure look familiar, don't they? Odds are they were globally assigned and got caught just the way I described. The odds of getting lore tokens from this level 10 skeleton are nearly twice what they are from the level 30 Ghost, at 4.47%. This one doesn't drop crystals at all, which is probably fine at level 10, but it also never drops processed resources for some reason (probably just a mistake, most things do).
Basically what I'm saying is, this is very difficult to maintain, and considering that most monster resources end up being assigned to all monsters of a particular race, and most monsters drop level-appropriate resources, and crystals, and so on, there's a ton of duplicate information. Another monster of the same level as our saris ghost probably has most of the same groups, just with "Werewolf Bone" or something instead of "Ghost Vapor Saris". Considering that there's 3300 monster templates to maintain, this much duplicate information causes a lot of confusion and trouble, particularly since changing any of it could modify quest rewards or other things somehow.
So, I've been reorganizing these. Now any given monster has, generally, two treasure groups: one common group (containing all the misc stuff that any given monster may drop -- crystals, resources, etc) and one monster-specific group (containing all the things that particular type of monster drops). This reduces our Saris Ghost to this:
WEIGHT NAME
100 "Monster_General_030-039"
50 "Monster_Ghost_Saris"
--------------------------------------------------------------------------------
150 XXXX
Here, the odds of a Saris Ghost Vapor dropping are 24.23% (Because there's several items in Monster_Ghost_Saris). The odds of a Dark Garnet Weapon Crystal dropping are 0.66%. These numbers shifted because the total number of items in the various groups changed, and because the amount of overlap between the old groups and the new groups varies (In this case crystals, in the general group, became more common -- they were extraordinarily rare in the previous group list). Other things, like dropped quests, got added to the Saris Ghost list in the last few weeks.
Even regardless of odds, it's must easier to keep consistent than the old groups were. More importantly, it means we can start introducing dropped items with some degree of confidence. If I want to add Imbued Armor Splinters to Saris Ghosts, I can do that without modifying 120 monster templates -- I edit a single treasure group instead, and can easily see the odds I'm changing as I do it (since they're largely in one place). If I want to tweak drop odds, I can run through and increase the number of saris-ghost specific drops without doing a ton of addition to see what my result is. In addition, I know that these groups are only being used for monsters, so I don't risk breaking quests/tasks if I make changes.
At this point I've reassigned the treasure groups on the majority of the game's monsters, though there's still a few remaining using the old disorganized groups (Zombies, in particular). As I've gone through, I've tried to keep monster-specific drops to a relatively similar chance, with some stumbling here and there (Golem fragments, mostly, which brought to light the bizarre number of techniques using them, including armor). The payoff of all this will be dramatically less breakage-prone and easier ways to add interesting loot to our monsters when opportunities to do so arise.
If you made it this far through my essay, congratulations. I hope I went at least a little ways towards explaining what "different" actually means, and why it's an appropriate word for the job. This reorganization has already opened the door for easier bug fixing and treasure additions, and hopefully will enable us to make exciting additions in the future.[/more]
