Assumptions when calculating luck for clog items

[eternal totem]

Hey, for context, I'm trying to calculate how lucky you've been for every item in the collection log (where possible). Some items have complicated drop mechanics, and I have to make some assumptions, otherwise those items can't be calculated.

Could people please give input on the following assumptions and let me know if they're valid? In general, I assume the player is fairly competent and I plan to add a tab to the settings panel where most of these assumptions can be modified.

1+2) Nightmare AND Nex: Assume players are in teams of 5, and they MVP 20% of the time. Both options configurable.

3. Assume the players have bought 0 abyssal lanterns from the Guardians of the Rift shop. # purchased would be configurable.

4. Assume players always make the required completion time for CoX challenge mode. This affects uniques + dust etc. Players could configure the % of the time they make time.

5. How many points does an average player get for CoX and CoX CM? I have no clue. This would default to the value y'all choose and be configurable.

6. For ToB and ToB HM, the player always reaches max pet rate. Seems like it's not too difficult as you can die 2+ times and still get max pet rate.

7. For ToB and ToB HM, the player completes raids in teams of 4, and the player receives 1/4th of the total MVP points. Both options configurable.

8. The player keeps 1 infernal cape and 2 fire capes (keep 1, sacrifice 1 to enter inferno) and gambles the rest. I could either make the "default" gamble-all-except-1 and have # not-gambled be configurable, or I could make the "default" gamble-none, and the player can configure how many they've gambled. Thoughts on all these options?

9. The player receives 2 rolls per Wintertodt crate. # rolls per crate is configurable.

10. ToA are done at 50, 150, or 300 invo. Configurable. Not sure team size matters.

11. I have no clue about team size / mvp rate for zalcano. 4 man and 1/4th MVP rate default? Or maybe 3-man crystal pickaxes?

12. Assume players only kill all 6 Barrows brothers. The problem with assuming this is your KC goes up if you just kill a few brothers then open the chest to reset when failing combat achievements repeatedly (I'm guilty here). Maybe # of wasted chests could be configurable... lol...

13. This probably goes without saying, but assume the player has mimics enabled, always does their mimic immediately, and never dies. I think an option for enabling mimics or not is reasonable, but I don't plan to make it further configurable.

[lunar ingot]

8. pet drops for inferno and fight caves varies a lot on task idk if you want to try to take that into consideration
9. 2.5 is probably a better default
10. should have additional 415 invo scale i know thats very popular
    depending on how precise you want to be a pre update and post update elite number might want to be added.

[eternal totem]

That reminds me, actually:
13) Skotizo kills are all completed after the jar drop rate update. Configurable: number of kills pre-update.

Updates are a problem in general because there is no way to know how many KC the player had before each update unless they input that info.

For elite number do you mean before masters were added or before mimics were added, etc?

[lunar ingot]

before the elite mega drop rate was increased

[eternal totem]

I agree that on-task should be taken into account too. Maybe default should be 1st cape not on task, rest of capes only on task

[lunar ingot]

probably a fair assumption, in general it should probably default to what is optimal, same for expected cox points

[eternal totem]

Agreed

First fire cape CAN be on task actually, right? Just not inferno

[lunar ingot]

uhh i think thats correct but most people dont since it requires a slayer unlock that most people probably dont have at the time?

[eternal totem]

Good point

[hard eagle]

1. assume 5 man always for reg nm imo (will fuck over people that soloed tho)
2. id use the shards as a metric to calc nex uniques and luck
3. yup
4. yup, making time in a CM is giga easy unless all learners
5. solo cox is 30-33k points
   Team cm varies depending on team scale and max eff strats.
   60-63k im solo
   35-45k in trio
   30-40k in 5 mans
6. always assume max pet rate
7. for hmt you can assume 5man since they are most efficient, but 4 man is more consistent and easier. Up to you.
8. yup

9. ill let camping answer that kekw

10. 8 man 400/410 is very efficient too

[eternal totem]

2. is pretty big brain. Assuming shards scale linearly with unique chance

[hard eagle]

It does

[lunar ingot]

that is a smart shout thats how most people check to see if theyre dry or lucky

[hard eagle]

Only issue is if people farm a lot of room kc it gets inflated a bit

Iirc its 350 shards per unique

But i never calced it

[eternal totem]

Number of players for ToA doesn't affect personal unique chance per KC, right? It just affects time for the KC? Assuming average contribution. Unless there is an MVP mechanic

[hard eagle]

Assuming avg contribution, correct

But if you mvp a lot, you get a ton of extra pts

8mans just tend to be faster

[eternal totem]

Yeah. I can make MVP% or points configurable

[desert cloud]

11. All 3 Zalcano uniques have different drop requirements. Zalcano shard scales with contribution but not team size, Smolcano scales with neither, and Crystal tool seed scales with both. 3-man or 4-man is a lot better for Crystal tool seed, masses are a little better for Smolcano, it's not clear which is best for Zalcano shard.

[eternal totem]

I started adding these assumptions to the config file, but now I ran into two different major issues:

A) Any KC from before the collection log was added (or before a specific item was added) will have an inaccurate luck calculation. For example, my Zulrah KC all happened pre-collection log, so the calculations are busted. All available options suck.

I could have a section for the user to configure the # of untracked KC for every boss, but that would be really messy, the user might not even know, and it might not even work if any item was added to a boss's log in a later update, OR items that had a drop rate change, etc. So that won't work...

The only possibility I can think of is somehow using the loot tracker (but I think this is nearly impossible to do?) to make sure your KC and tracked items are from the same time period, but if you ever reset it'll break, etc....

So far the best option I can think of is to "do nothing" and just allow the calculation to be possibly inaccurate. And maybe throw a warning somewhere so users know why their calcs are messed up.

B) ALL of the config options we discussed would be client-side only. This means that if you type "!log luck asdf" into chat, nobody else will know what your typical Nightmare team size is, which you configured in your plugin config, for example. The only option I can think of is to completely disable luck calculation for other players' logs for any item where any configurable assumption needs to be made.

This means that you can't see anyone else's:
All 3 raids
Nightmare
Nex
Wintertodt
Zalcano
Barrows(?)
Jad/Inferno pet
maybe more

Actually, another option is to just print the luck as normal but then attach a message that says something like "this is only accurate if <insert assumptions made>". It's a bit messy and could be a long message. Maybe people could even spam with it.

[desert cloud]

I think it's fine for your tool to not handle the case of pre-log KC. You could catch cases where you're over 99.999999% dry and print a warning with it, like (Probably inaccurate)

[eternal totem]

Yeah I think in both cases above the best solution might just be to print a warning if it detects it's probably inaccurate

Can't think of anything better

[hard eagle]

You could compare other items on the same log & see if its logical.
Example: if the person is 4x dry for serp, onyx, fang, bp. Then something sus might be going on

[eternal totem]

True, maybe one individual item could be uber dry but not all of them. There are already plans to have a per-boss luck so I could just reuse that here

[eternal totem]

14. For Venenatis/Callisto/Vet'ion, do you get a KC in the log if you're not in the top 10 players which receive loot? Can I assume that if you have X KC, you had at least X rolls on the table (although the roll may be split between up to 10 players, where # of players should be configurable)?

wondering if this statement is accurate since the wiki is not clear: https://www.reddit.com/r/ironscape/comments/15nrr1f/if_i_kill_callisto_with_another_ironman_who_is/

"If a unique is to drop at a wildy boss all participating players will receive 1 roll at the unique; however, the MVP of the kill will receive 2 rolls. (Only one unique can ever possibly drop per kill)

You'll know you got MVP if you're able to see the bone drop from the boss."

[hard eagle]

That’s correct

[hard eagle]

I highly doubt this. Afaik mvp does nothing besides confirm you did the most damage

Theres also a min roll of 15% of loot potential even if you do 0.1% dmg (1 damage).
Thus far, via callisto mass discord, it feels like this rate also applies to pet, but can’t confirm that as it could be due to spoonage

[eternal totem]

I agree, it wouldn't make sense if you have 2 players and 1 gets 2 rolls and 1 gets 1.

Was worried there's still a 5% chance or something like nex

[hard eagle]

Nah dont think you gotta worry about that imo

[eternal totem]

Alright, I'll have a setting for "average callisto loot potential" or whatever. Put it at say 0.2 if you average 5 mans (or just generally get that percent of the damage)

[eternal totem]

I'm thinking of not using shards anymore to calculate Nex uniques because you can also get shards from other sources (blood reaver, spiritual mages, etc.) and it's also a bit of a rarer drop than I thought (1/16 ish). It's very possible to get a Nex unique before ever getting shards from Nex, which would make calculating spoonage bug out for that case

Maybe it works for people with thousands of KC and it could be used as an alternate method to cross check

[tame magnet]

how else are you gonna estimate nex? its still a much much better estimator to use the shards than kc

[eternal totem]

I think it's only better for people with very very high KC. Otherwise it can be treated like Nightmare (input average team size + mvp rate)

[tame magnet]

oh ud have to input your avg team size ok that works

[eternal totem]

Shards are like a 1/16 if I recall, so anyone with less than say 100 KC might not even have any shards lol

I'll look into maybe combining both approaches depending on your kc

[tame magnet]

i mean if someone has 100 kc but u dont know what team sizes, then the shards still the best estimator

u dont need to use the shard count if u know avg team size tho

[eternal totem]

Yeah, team size is an estimate by the user though, hopefully they're actually correct when they input it. Otherwise I'm trying to figure out the math for like a 95% confidence interval given your number of shards

Like if you've received X shards, how much total contribution (effective solo KC) have you had, as a range

[tame magnet]

hf with that

[eternal totem]

lol thanks, will try

This is definitely taking longer than I thought

[tame magnet]

for raids, are you also having user input for the avg scale/raid lvl?

cause there u can use the total amount of purps as estimator and the math for that would be the same as using shards for nex

well almost same

[eternal totem]

yep all the assumptions are in the first 2 ish posts

It's a shame there isn't a way to track actual raid points etc.