Dungeon Drop Mechanics

This is an archive of the mabination.com forums which were active from 2010 to 2018. You can not register, post or otherwise interact with the site other than browsing the content for historical purposes. The content is provided as-is, from the moment of the last backup taken of the database in 2019. Image and video embeds are disabled on purpose and represented textually since most of those links are dead.

To view other archive projects go to https://archives.mabination.com

Rydian wrote on 2015-05-20 05:35
I'm curious on how certain categories of items are picked for dungeons, so I looked through the leaked server files (from G13, mind you, so some drop rates and stuff may be different) and I checked out the Alby Advanced Hardmode boss script specifically. Note that this dungeon acts slightly different from other dungeons in that it generates 8 chests every time and you can craft Yarn Keys to open the extras if possible.

When the boss is generated (door opened), the treasure is selected too. Here's a rough order of how things happen (assuming that OnPuzzleActivate() is calling GetItemRewardDesc() + GetMoneyBagDesc() and they're not called separately).
1. The boss is spawned alongside the spinning wheel and minions.
2. The server counts the number of players in the party. If there's more than 8 (?!) it only counts as 8.
3. The server generates 8 chests and puts gold in them. It picks a min of 13000 and a max of 26000, and then... multiplies those by 0.16. Dungeons are set up to have different gold multipliers (decimators really) according to party count, but for Alby Advanced Hardmode it's unchanged (all the multipliers are 0.16 here). Strangely, the non-HM version is set up for higher gold rewards.
4. The server runs a loop to generate "junk" items for the chests. Magic powder, NPC equipment with slightly-different stats, low-level potions, etc. Puts one result in each chest if I'm reading this correctly. Too lazy to list the items and probabilities 'cause it's all NPC or low-level crap.
5. The server next checks for some features (if special upgrades and extra enchants are enabled) and then sets up stuff to generate the robe and upgrade stone. The following assumes that these features are enabled.
6. The server runs the following loop once per player count in the party.
  6a - Generate a random number to 1000.
  6b - If this number is less than or equal to 40, put a Hebona Robe in chest x (x being the loop count+1).
  6c - Generate a random number to 1000.
  6d - If this number is less than or equal to 30, put a Blue Upgrade Stone in chest x (x being the loop count+1).
7. Generate a random number up to 100, once.
  If it's less than 30, give the Warrior enchant scroll in chest 1.
  If it's 30-54, give a Lute with the Lucky enchant in chest 1.
  If it's 55-84, give the Lethal enchant scroll in chest 1.
  Otherwise, give the Precise enchant scroll in chest 1.
8. Set up the stuff to be run when the boss is gone, such as giving the keys, checking and assigning journal progress.

So there's a few things of interest here.

A - Enchants only show up in chest 1, so you can only get one of them. This is likely what players refer to as a "primary" dungeon reward. The other dungeon scripts would have to be examined for else statements outside of loops to see what else (maybe skill books?) might be a "primary" reward where you can only get one item from that category no matter the chests opened.

B - Chest numbers higher than the party count can't contain a Hebona Robe. So if the chests in the room are always in a specific layout (1 is somewhere, 2 is somewhere else), that means you can figure out which chests cannot have a robe in it because they were not reached in the loop. This could help people save on yarn keys and get effectively higher chances per total chests opened if they can disregard chests that cannot have a robe.

Unfortunately I have not yet figured out if there's a pattern to which chest is which because getting people to cooperate in Mabi and not open chests until I say it's okay is proving quite difficult.

C - The server accounts for more than 8 players in a party, and this was well before Raid parties. It also accounts for less than 1 player, so either this is a failsafe (there's a failsafe gold multiplier too), or it accounts for GM trickery.

Oberynn wrote on 2015-05-20 06:29

This is some sick insight, makes me wish the leaked files were more recent.

Qnjo wrote on 2015-05-20 10:56

Quote from Rydian;1275710:

6a - Generate a random number to 1000.
6b - If this number is less than or equal to 40, put a Hebona Robe in chest x (x being the loop count+1).

So, in other words, drop rate of the robe is 4%?

Calistin wrote on 2015-05-20 14:03

Would this information also definitively prove/disprove the looting pre-boss room chests = lower value rewards?

Citrusjuice wrote on 2015-05-20 14:38

Id be happy to [s]leech[/s] help you out with not opening chests o3o

isn't there a pattern to which chest is what # depending on what the last floor map layout is like?

Kenero wrote on 2015-05-20 14:52

Quote from Calistin;1275732:
Would this information also definitively prove/disprove the looting pre-boss room chests = lower value rewards?

It disproves it. There is nothing in the data that says looting the chests that spawn in hallways causes the dice ratio to affect rare item drop rate to make it lower or higher.

So loot all the end of hallway chests as much as you want.

Fracture wrote on 2015-05-20 14:58

Quote from Rydian;1275710:

Unfortunately I have not yet figured out if there's a pattern to which chest is which because getting people to cooperate in Mabi and not open chests until I say it's okay is proving quite difficult.

Unfortunately I'm going to guess there isn't. In prior experience with usually running with 2-3 people, the chests show up at random points in their circle, and those chests would theoretically be Chest 1 through Chest 3/4. That said, 8-chest dungeons may be different.

Rydian wrote on 2015-05-20 18:18

Quote from Qnjo;1275723:
So, in other words, drop rate of the robe is 4%?

Was trying to avoid saying stuff like that since things could have changed since then.

I mean this is Mabi so it seems doubtful at first that drop rates for older dungeons would change, but there are separate logic blocks for two different enchant sets depending on if an extra enchant feature is enabled and that would change relative probabilities for the existing enchants when more are added, so they actually do have drop pools set up to change when new stuff is added to them.

Quote from Calistin;1275732:
Would this information also definitively prove/disprove the looting pre-boss room chests = lower value rewards?

If you mean the rooms with 9 chests, I saw nothing that referenced their existence in the boss+treasure script. I didn't even think of looking for that, and all of the numbers were defined as-is in the script itself, not inherited from elsewhere.

Quote from Citrusjuice;1275736:
Id be happy to [s]leech[/s] help you out with not opening chests o3o

Well I'm mainly on Ruairi, I totally have to be carried with the Mari alt.

Quote from Citrusjuice;1275736:
isn't there a pattern to which chest is what # depending on what the last floor map layout is like?

Idunno' if there's a pattern for the 8-chest generation yet as I've only gotten to examine one so far.

Anyways, thanks to Xcelled and a mule we managed to get one test done. As theorized, the chests are generated with sequential IDs, so noting their positions should let you know which chest is which. In the test run we did, I wrote this down.
[code]45360614198280199 45360614198280196 45360614198280195

45360614198280197 45360614198280200

45360614198280198 45360614198280201 45360614198280194[/code]

And we end up with this resulting chest layout.
6 3 2
4 - 7
5 8 1

And chest 1 in the bottom-right corner contained the Lethal enchant.

But more testing is needed to see if there's some sort of human-usable pattern. If not, it will at least be possible for people with logging software to determine the order easily.

Citrusjuice wrote on 2015-05-20 22:50

Quote from Rydian;1275760:

Well I'm mainly on Ruairi, I totally have to be carried with the Mari alt.

Idunno' if there's a pattern for the 8-chest generation yet as I've only gotten to examine one so far.

I have accounts on both Mari and Ruairi.

if you can figure out how many different types of layouts there are it's possible to begin testing since chest 1 always drops enchants(?)

X wrote on 2015-05-21 00:17

From a few tests I did, the chests are placed randomly. So, tough luck for those of you hoping to find a magic chart, tough luck.

That said... there is hope.

[Image: http://puu.sh/hUMzj/516c83d167.png]

Rydian wrote on 2015-05-21 11:39

"Normal" dungeon rewards are set up using kind of a weird system, took me a bit to figure out what the hell the code is trying to do... it's using a switch statement inside a while(true) loop and has the comparative number threshold cumulative to the previous one in the switch statement. Seriously like. I don't even.

Par normal reward? You want that Snow Crystal right?

1 - Set a counter to 0.
2 - Roll a random number to 100. Let's say it's 50.
3 - Start a loop that cycles through rewards until breaking past the random number.

3a - Select Magic Powder, add 18 to the counter.
Random number (50) is not less than the counter (18), so go to the next check.
3b - Select Mana Herb, add 18 to the counter again.
Random number (50) is not less than the counter (36), so go to the next check.
3c - Select HP 50 Potion, add 18 to the counter again.
Random number (50) is less than the counter now (54), so return.

4 - The last set item (HP 50 Potion) ends up being the reward.

I'm not sure why they don't just use a damned "if the number is <x, choose x, else if, blah blah" thing like they do elsewhere. It would be a hell of a lot easier to read and edit later on down the line for changing dungeon rewards around. Maybe this is why they just keep adding additional drop tables, maybe nobody left at Devcat can read the damned legacy code to edit the originals!

Anyways.

Random number to 100, it needs to be 90 or higher to get the Snow Crystal.

Snowie Stormflower wrote on 2015-05-21 15:02

Do rewards for Shadow Missions / Theatre Missions use the same mechanic or are those rolls done when the chests spawn rather than when you enter the mission?

Kenero wrote on 2015-05-21 16:38

The dungeon data reveals a few stuff.

First, if you want to get a Hebona Robe, you need more people in party.

If there is only 1 person in party, then only one chest out of the 8 will roll for the Hebona. Making hunting a Hebona inefficent.

By having all 8, even as a filler, will ensure that all 8 chests will roll for the Hebona.

X wrote on 2015-05-21 17:41

Gotta love devCat's troll: "Oh, you're hunting for a Hebona?? Here, have 7 extra keys THAT CAN'T GIVE YOU WHAT YOU WANT."

Rydian wrote on 2015-05-22 03:52

Quote from Snowie Stormflower;1275886:
Do rewards for Shadow Missions / Theatre Missions use the same mechanic or are those rolls done when the chests spawn rather than when you enter the mission?

Guardian of Avon...
Determines the rewards when it generates the chests, upon completion.

Random number to 100, if it's less than 30 generate another pass from the chest.

Random to 100 to determine whether to drop a combo card or not (always drop one if the server is run in test or dev mode). Basic/Int cannot drop. Advanced needs it less than 6, Hard and Elite (placeholder?) need it less than 12.

Generates another number up to 100 if it's going to make a card, there's 5 choices (the type of card). The 6th choice, with a 1% chance, is to make a gold combo card.

Defeat Fomor Commander I...
Generates the chests and chooses the items upon boss death as well. Harder to determine exactly how Shadow Mission drops work because instead of each mission script having a drop table like Dungeons and Theater Missions seem to, the scripts reference a global "shadow mission drop table" file and I CBA to cross-reference the IDs.

But the answer is that it seems all three types of dungeons/missions choose the reward at the end, at least as far as I can see.

1 2 3 Next › Last »