Arcane Drops: Randomness and...Patterns?!

[aidan890]

Tenno,

We took a look at this thread: https://forums.warframe.com/topic/941619-eidolon-arcane-rewards-strange-drops/ and can confirm there is indeed a pattern that defies the expected Arcane Drop RNG. Our drop rates table produces % that we'd expect the results to reflect, but this is not the case. 

Some visualization of the unforeseen Arcane Consequence: 

 

There really shouldn't be any white spaces here! This means that, for example, Arcane Energize and Arcane Strike have never dropped together. This was never intended and we have already fixed the issue! 

 

This is the result we expected to see when we added Arcanes to the Eidolons.

 What Happened?

Some diligent Tenno collecting independent research have unearthed a very unfortunate and very unintentional circumstance for Arcane RNG distribution. Essentially, the Arcanes have been dropping in consistent sets of 3 rather than each having independent RNG as we had intended when Arcanes moved from Trials to Eidolons. This means that certain combinations Arcanes are actually much less likely to be rewarded, sometimes even incapable of being rewarded, than our previously released drop rates would suggest. This is what causes the several white spaces in the graph above. 

The issue was caused by the method we used to ensure that all players in a single instance would receive the same Arcane rewards, regardless of when they joined the hunt; There was some initial desync occurring in testing that we quickly fixed. The unintended consequence of this fix was that only certain progressions through the drop tables existed which means that all of the Arcanes were effectively sorted into sets that we had never intended for. 

Technically, the system we had devised was always “working” but with these unforeseen side effects of the initial RNG roll occuring at the onset of the mission. Essentially, if you were to roll a rare Arcane on the Gantulyst, you could predict that you could not get a rare from the Hydrolyst. This was never what we intended.

Is it fixed? Yes!
We have diligently applied some Arcane Healing! A fix for this issue has already been completed server-side that will make each Arcane a truly random drop, and will also keep the rewards consistent across all members of a squad. Previously, a single roll determined all three Arcanes at once, but obviously this is what caused the above issue. 

The issue appears to have also affected console drop rates as well, so the fix will be applied server-side as well. This post will be edited with more specific console information when that info is available! 

EDIT: The fix has also been applied to XB1 and PS4!

[(PSN)s3d_sas]

After reading it I see that you made the RNG of the arcanes much better but what about console players?

edit: nvm I’ve seen that you mentioned us ty <3

[[DE]Grzegorz]

Wew lads.

[(PSN)s3d_sas]

Sorry for the double posts

i thanks @(PS4)OmegaSlayer For his dedication (even that he kept spamming the topic.....)

[Perisurdas]

Oh, looks like my google docs spreadsheet is kind of pointless now... Still, nice to have it fixed I guess, hated getting Victory + Aegis and then running out of time on 4th/5th run.

[Kurambik]

Thank you :)

[Fruchtpudding]

Is this a leftover from the old (2013 or so) seed based system where you could calulate what drops you would get on further rotations from the first few? I remember being able to use a calculator for determining if the squad would be able to get a Banshee part on Xini after 10 waves.

 

Edit: Found it https://www.reddit.com/r/Warframe/comments/1g3bar/tier_3_defense_mission_reward_calculator/

 

[Tojad]

Here's hoping it won' be Maroo style serverside fix

[Wyrmius_Prime]

See guys, it's not always ''Just RNG''. There is no true random, if s system is something developed by a human, it will never be truly random.

[God_is_a_Cat_Girl]

No wonder that back when I was doing Eidolon hunts my part managed to get the same set of arcanes back to back.

 

Would be nice if other stuff like Bounties could be at least looked at to make sure they aren't suffering of similar issues, since I've personally got quite a few infuriating repeated sets of rewards to the point that it didn't feel worth playing.

[Thundervision]

So I wasn't insane after all, hooray!

[-dicht.Goko-]

This is why data mining was a useful endeavour...

[(PSN)s3d_sas]

5 minutes ago, Wyrmius_Prime said:

See guys, it's not always ''Just RNG''. There is no true random, if s system is something developed by a human, it will never be truly random.

Hmm

Any suggestion about a different specie that could do it

[Voltage]

Just now, -dicht.Goko- said:

This is why data mining was a useful endeavour...

But of course we must crush the people who find the flaws in the files. Remember when Snipetron rivens were removed from rotation? Datamining allowed players to see this directly. Sad that we can't have nice things.

[GrayArchon]

2 minutes ago, -dicht.Goko- said:

This is why data mining was a useful endeavour...

I mean, players were able to identify this problem and bring it to DE's attention (where it was promptly dealt with) without using data mining.

[Wyrmius_Prime]

5 minutes ago, (PS4)s3d_sas said:

Hmm

Any suggestion about a different specie that could do it

Haha so funny!

1 minute ago, GrayArchon said:

I mean, players were able to identify this problem and bring it to DE's attention (where it was promptly dealt with) without using data mining.

It wouldn't have taken so long to fix it though. Without datamining it required pretty big sample sizes to gather enough proof that there is something to prove that there was something wrong with the drop pattern.

[AuroraSonicBoom]

You might want to check other reward algorithms you're using as well. I've seen similar drop patterns occur during bounties, especially during one day where I got 11 Eidolon lens blueprints in a row.

 

Just now, Wyrmius_Prime said:

Without datamining it required pretty big sample sizes to gather enough proof that there is something to prove that there was something wrong with the drop pattern.

Completely worth the added security and discouragement of spreading spoilers. Your argument is weak.

[Wyrmius_Prime]

7 minutes ago, AuroraSonicBoom said:

Completely worth the added security and discouragement of spreading spoilers. Your argument is weak.

Spoilers are easy to avoid as long as you don't enter Reddit. On the official forums one would get banned in an instant. What makes the spoilers through datamining even less severe, is the fact that somehow DE manages to spoil everything inside the game anyway. There's no surprises to begin with.

[aidan890]

@AuroraSonicBoom @Wyrmius_Prime Just stepping in as moderator here, can you kindly take this discussion to PMs or elsewhere, as it is veering off topic. 

Thanks!

[Amazerath]

In light of this wouldn't it be worth it to check other reward systems as well? Just to be sure everything is working as expected?

[aidan890]

1 hour ago, [DE]Aidan said:

EDIT: The fix has also been applied to XB1 and PS4!

 

[(PSN)OmegaSlayer]

2 hours ago, --Q--Voltage said:

But of course we must crush the people who find the flaws in the files. Remember when Snipetron rivens were removed from rotation? Datamining allowed players to see this directly. Sad that we can't have nice things.

You should, in all honestly, just stay silent or offer me pubblic apologies after having called me a tinfoil hatted spammer with no evidence.

 

But most importantly thanks DE for checking and taking action and thanks to all those who helped collecting data

[Chipputer]

4 hours ago, (PS4)OmegaSlayer said:

You should, in all honestly, just stay silent or offer me pubblic apologies after having called me a tinfoil hatted spammer with no evidence.

 

But most importantly thanks DE for checking and taking action and thanks to all those who helped collecting data

This is the norm for Voltage. Anything that will get him reputation.

[Toran]

vor 13 Stunden schrieb [DE]Aidan:

 

...

Technically, the system we had devised was always “working” but with these unforeseen side effects of the initial RNG roll occuring at the onset of the mission. Essentially, if you were to roll a rare Arcane on the Gantulyst, you could predict that you could not get a rare from the Hydrolyst. This was never what we intended.

Is it fixed? Yes!
We have diligently applied some Arcane Healing! A fix for this issue has already been completed server-side that will make each Arcane a truly random drop, and will also keep the rewards consistent across all members of a squad. Previously, a single roll determined all three Arcanes at once, but obviously this is what caused the above issue. 

The issue appears to have also affected console drop rates as well, so the fix will be applied server-side as well. This post will be edited with more specific console information when that info is available! 

EDIT: The fix has also been applied to XB1 and PS4!

So technically, before the fix, a drop was a dependent event and the drop chances resulted out of the (unknown to us) probabilities of the 20+ drop groups/combinations.

Have the arcanes' drop chances been changed with the fix?

Bad news for us btw.: we will see abysmal combinations too now  

[Thanathros]

 

More relevant than ever.