Relic drop rates bugged

[greymistbc1]

Reposted from >PC Bugs because I was told to do so..

 

I have been tracking a number of relic drops that just don't happen... certainly not with the listed probability.

My info.. almost 900 hours played, mostly solo, MR 9.

As of yet, I have failed to get the parts for a Prime (anything). I have a bunch of items where I have two or more of every part except one, that absolutely will not drop.

Easiest example, Lex prime. I was farming the Meso T3 Relic for weeks (before it vaulted), found and cracked the relic a lot, (since the barrel is 'common, I did not not refined). I have 5 Paris lower limbs, a bunch of forma, 3 Chroma Prime Chassis, but I can't get the damned barrel to drop. I cracked at least 13 of them... 13 rolls at 25%, looking at right at 97.6% that the Lex barrel should have dropped.)

I have had similar problems with the Fang Prime, Orthos Prime, and a bunch of others.

I each case, the one remaining part has failed to drop, and the odds for each having failed is above 95% each.

Add to this, the fact that the devs keep 'vaulting' relics, and replacing them with higher relics... basically, they are making it so you cannot get any of the prime items without paying plat.

Pay-to-win. 

Ugly.

I spoke to support..

 

          [DE]Kenneth (WARFRAME Support)

..

          As stated before drops and spawns are set to be random.

..

..I don't think they understand what 'random' means.. at this point, my odd of having the parts for at least one of these Prime items exceeds 99.99%... how many "9"'s does it take to admit there is a problem?

[Jiminez_Burial]

16 minutes ago, greymistbc1 said:

Reposted from >PC Bugs because I was told to do so..

 

I have been tracking a number of relic drops that just don't happen... certainly not with the listed probability.

My info.. almost 900 hours played, mostly solo, MR 9.

As of yet, I have failed to get the parts for a Prime (anything). I have a bunch of items where I have two or more of every part except one, that absolutely will not drop.

Easiest example, Lex prime. I was farming the Meso T3 Relic for weeks (before it vaulted), found and cracked the relic a lot, (since the barrel is 'common, I did not not refined). I have 5 Paris lower limbs, a bunch of forma, 3 Chroma Prime Chassis, but I can't get the damned barrel to drop. I cracked at least 13 of them... 13 rolls at 25%, looking at right at 97.6% that the Lex barrel should have dropped.)

I have had similar problems with the Fang Prime, Orthos Prime, and a bunch of others.

I each case, the one remaining part has failed to drop, and the odds for each having failed is above 95% each.

Add to this, the fact that the devs keep 'vaulting' relics, and replacing them with higher relics... basically, they are making it so you cannot get any of the prime items without paying plat.

Pay-to-win. 

Ugly.

I spoke to support..

 

          [DE]Kenneth (WARFRAME Support)

..

          As stated before drops and spawns are set to be random.

..

..I don't think they understand what 'random' means.. at this point, my odd of having the parts for at least one of these Prime items exceeds 99.99%... how many "9"'s does it take to admit there is a problem?

Come back with a sample size of a few thousand and then we can talk about there being a problem.  Also please don't copy paste text in the forums, it results in incredibly ugly white/grey bars for anyone not using the same theme as you.

[greymistbc1]

2 hours ago, Jiminez_Burial said:

Come back with a sample size of a few thousand and then we can talk about there being a problem.  Also please don't copy paste text in the forums, it results in incredibly ugly white/grey bars for anyone not using the same theme as you.

A few thousand?  Since people usually say "a couple" when they mean two,  let's be generous and assume you meant "three" (rather than 4 or more).

So.. 3000 relics earned, and three thousand cracked.

Assume it takes 10 minutes in a mission to earn one, and 10 minutes to crack one, 3000x10x10/60 = 5000 hours of play, or 624 eight hour days.

THAT is the bar you are setting.

So.. another troll, I guess.

I am interested in actual feedback, if someone has something useful (?) that would be awesome.

 

[Jiminez_Burial]

32 minutes ago, greymistbc1 said:

So.. another troll, I guess.

Not a troll, someone who understands statistics.  And I never said you had to be the source of the data points, just that a sample size of a few thousand is the minimum you would need to reach any even slightly viable conclusion in a game with millions of users.

[greymistbc1]

1 hour ago, Jiminez_Burial said:

Not a troll, someone who understands statistics.  And I never said you had to be the source of the data points, just that a sample size of a few thousand is the minimum you would need to reach any even slightly viable conclusion in a game with millions of users.

I've had some basic statistic, certainly not an expert, so help me understand, if you can.

According to the game, the Meso T3 (before it was vaulted) listed the Lex Prime barrel as a 'common' item,  with a 25% chance to drop. Therefore, using the above example, the chance that I would fail to get one, in 13 attempts, is 0.75^13 = 0.0238.  Roughly 2% to fail. 

I have had this happen with _every_ prime item that I am trying to get... always one part that fails to drop.

If there are millions of users,  my results alone may not be significant... so the way to get more data might be... to ask on the forums? To see if anyone else is having the problem? (i.e., my original post).

Alternately, this could be an indicator of a very real, but more localized phenomena, something that doesn't affect millions of users, but rather, say, only PC users who have taken long breaks from the game,  and there is a bug that happens with steam updates for a fraction of that subset.  (idk, just as a semi-plausible example).

Another possibility is that this is intentional and 'working-as-designed' by the devs,  to push people into buying with platinum out of frustration. (If that sounds paranoid, I could tell you stories about Eve Online... the devs are just people.. every bit as likely to do corrupt things as priests and lawyers and everyone else.) 

 

[Jiminez_Burial]

8 minutes ago, greymistbc1 said:

I've had some basic statistic, certainly not an expert, so help me understand, if you can.

According to the game, the Meso T3 (before it was vaulted) listed the Lex Prime barrel as a 'common' item,  with a 25% chance to drop. Therefore, using the above example, the chance that I would fail to get one, in 13 attempts, is 0.75^13 = 0.0238.  Roughly 2% to fail. 

I have had this happen with _every_ prime item that I am trying to get... always one part that fails to drop.

If there are millions of users,  my results alone may not be significant... so the way to get more data might be... to ask on the forums? To see if anyone else is having the problem? (i.e., my original post).

Alternately, this could be an indicator of a very real, but more localized phenomena, something that doesn't affect millions of users, but rather, say, only PC users who have taken long breaks from the game,  and there is a bug that happens with steam updates for a fraction of that subset.  (idk, just as a semi-plausible example).

Another possibility is that this is intentional and 'working-as-designed' by the devs,  to push people into buying with platinum out of frustration. (If that sounds paranoid, I could tell you stories about Eve Online... the devs are just people.. every bit as likely to do corrupt things as priests and lawyers and everyone else.) 

 

Again with the copy pasted text.  If you do that with another response I simply won't bother with a reply.

If you're arguing that this thread was a means to get more data then either you missed the mark completely due to the fact you never requested more data as well as putting this in Feedback rather than General Discussion, or you're being completely disingenuous.  Either way I would suggest that something change.

If you've had some basic education in statistics then you're surely aware of 'outlier' data points, data points that despite not being invalid can't (and shouldn't) be used to make a claim about over-arching patterns or correlations.  You should also be aware of personal bias where people put a larger weighting on experiences they themselves have had rather than looking at the wider picture and valuing each instance equally.  You should also (again) be aware that the human mind is programmed to seek out patterns even when none exist.  What these three things do in conjunction is make you think that your personal experiences that are likely outliers are indicative of the wider situation, and are then justified by the unrelated context that you took a long break from the game.  Without a far larger data to provide statistical context, you shouldn't be making claims of any kind.  And before you say to have never made a claim - "certainly not with the listed probability".

All in all you're falling into all the traps that someone who truly understands statistics wouldn't fall into.  You're also not really seeking any further data because as I said before you're posting in Feedback after you had already contacted Support.  That means rather than seeking wider context you seem to be fixating on the idea that something is wrong.

[greymistbc1]

I was using the 'quote' button within the forums.. not sure why that is causing you formatting problems.

I was posting here, in this thread, because the Support response gave me the link to this thread, and told me to post here.   I would give you the quote, but that would require copy/paste..

I am certainly not an expert in the different sections of the forums, or why "General Discussion" would be more appropriate than "Feedback" as a place to post.

You seem annoyed.. and fixated on hand-waving my concerns away.  My premise is simple... if you fail a 1:4 die roll enough times, it is significant.  "Enough times" is something much less than 3000.

I understand that a 1:4 chance might fail four times in a row... even eight or nine or whatever.  Random events can be 'streaky' or 'clumpy'.  

However, continual failure can indicate that probability is not involved at all, but rather some sort of code error.

[Jiminez_Burial]

47 minutes ago, greymistbc1 said:

I was using the 'quote' button within the forums.. not sure why that is causing you formatting problems.

If the initial comment was copy pasted (which it looks to be) then each re-quote of that is going to appear the same way.  If you then continue on from that re-quote then the same formatting info is applied to everything afterwards until a format break is applied.

47 minutes ago, greymistbc1 said:

I was posting here, in this thread, because the Support response gave me the link to this thread, and told me to post here.   I would give you the quote, but that would require copy/paste..

I am certainly not an expert in the different sections of the forums, or why "General Discussion" would be more appropriate than "Feedback" as a place to post.

I mentioned General Discussion in response to you saying "So the way to get more data might be... to ask on the forums?  To see if anyone else is having the problem? (i.e., my original post)".  If you intended the thread to be a way to gain further data then it would belong in General Discussion (or to a lesser extent Players helping Players).  If you intended it as feedback then the validity of that feedback is entirely dependent on something other than simple statistics taking place.  And regarding the "that would require copy/paste.." comment, it doesn't.  As I have done you can simply type the quote in it's entirety or just the relevant portion assuming it doesn't remove important context.  In cases where the poster is using Light Theme and the reader is using Dark Theme it results in ugly white blocked text.  In cases where the poster is using Dark Theme and the reader is using Light Theme it results in black text on a black blocked background (which as you could imagine makes it very hard to read).  As a result it's worth the extra minute or so (at least in my mind) to simply type things out rather than copy and paste.

47 minutes ago, greymistbc1 said:

You seem annoyed.. and fixated on hand-waving my concerns away.  My premise is simple... if you fail a 1:4 die roll enough times, it is significant.  "Enough times" is something much less than 3000.

I get annoyed when people try to portray their gut feelings as anything other than just that... a gut feeling.  You are making statistical claims that require actual evidence, evidence that would require a far larger sample size than you have provided.  You are making a claim in the affirmative (that the drop rates are not functioning as described by DE) therefor the Burden of Proof lies with you.  In all honesty even 3000 data points would be on the low-side to truly say that anything aside from randomness is at play.  You would also need to have that data supplied by unbiased means so you can't simply post on a forum asking for data since that data would have a bias of affirmation (meaning people who see the thread and have had the same gut feeling as you are more likely to supply data than those who haven't had the same gut feeling).

47 minutes ago, greymistbc1 said:

I understand that a 1:4 chance might fail four times in a row... even eight or nine or whatever.  Random events can be 'streaky' or 'clumpy'.  

However, continual failure can indicate that probability is not involved at all, but rather some sort of code error.

And as stated above you have the Burden of Proof.  You need to supply enough unbiased data to say undoubtedly that something is happening either not as intended, or not as described.  Until you do that (according to proper statistical methodology) then nothing you say holds weight when discussing the global situation.

Edited December 29, 2019 by Jiminez_Burial

[greymistbc1]

Your burden of proof is, afaik, impossible to meet unless one has access to the source data (in other words, I would have to be a dev).

Am I missing something?  Is there some way to get >3000 data points from an unbiased source?  

If not, than requiring proof that is impossible to get is just a higher form of trolling (no insult intended).

[Jiminez_Burial]

42 minutes ago, greymistbc1 said:

Your burden of proof is, afaik, impossible to meet unless one has access to the source data (in other words, I would have to be a dev).

Am I missing something?  Is there some way to get >3000 data points from an unbiased source?  

If not, than requiring proof that is impossible to get is just a higher form of trolling (no insult intended).

Unless you want to break a few privacy laws by installing software that forces a screen capture/video recording when people run a fissure mission, all while the person is unaware that they are part of data collection for this experiment then there isn't an entirely unbiased means of gathering meaningful data.  But that does not mean that your data therefor becomes meaningful, it remains just as meaningless.  If you can't provide the evidence to meet your Burden of Proof, then don't make a claim in the affirmative in the first place.  That also doesn't make what I am doing "a higher form of trolling".  What I am doing is informing you of why your data (and your thread/claim as a whole) is not valid in a statistical sense, which is the only sense that should matter when you're talking about raw percentage numbers.  To troll you would be for me to say your data doesn't matter, but do so in a way that you could reasonably think you can do something to change that.  I am making it very clear - you are not in the position to make a statistical claim with any validity.  Neither am I but I'm not the one making a claim in the affirmative (I'm refuting a claim that someone else made in the affirmative which is not the same thing).

In your initial reply you stated:

5 hours ago, greymistbc1 said:

I am interested in actual feedback, if someone has something useful (?) that would be awesome.

which is interesting considering it's the same thing DE want.  The difference is that you do not have sufficient statistical proof in order to provide "something useful".  The thing of use that I am providing you is a more than reasonable explanation as to why there is no point in making a definitive claim when you can't back that claim up.  Again, that's just me trying to save you from wasting time down the line when you have another gut feeling about mechanics dictated by RNG.  A troll would simply be having a 'Yah-uh, Nah-uh' back and force with you without explanating the situation.

Edited December 29, 2019 by Jiminez_Burial