Fix Drop Rates

[DarkKnight271]

It shouldn't take 15x4 relics to get an uncommon item to drop. And it shouldn't take that many for a common item to drop. You all need to fix the drop rates of relics.

[trst]

Doesn't sound like anything is broken.

Actually it sounds like they're working as intended and you got unlucky.

[DarkKnight271]

2 minutes ago, trst said:

Doesn't sound like anything is broken.

Actually it sounds like they're working as intended and you got unlucky.

You do understand the difference between common, uncommon and rare in a normal game where drop rates work properly right? I shouldn't have a greater chance of getting a rare drop than an uncommon or common drop. And it shouldn't take as many or more relics to get an uncommon drop than a rare drop. That's broken. If you don't agree, that's fine, that's your opinion, but you are wrong.

[LSG501]

8 minutes ago, DarkKnight271 said:

You do understand the difference between common, uncommon and rare in a normal game where drop rates work properly right? I shouldn't have a greater chance of getting a rare drop than an uncommon or common drop. And it shouldn't take as many or more relics to get an uncommon drop than a rare drop. That's broken. If you don't agree, that's fine, that's your opinion, but you are wrong. 

Well to be fair, your first post wasn't exactly the clearest post in the world.... 

I worked out it was something about drop rates and relics but other than that it could have been complaining about anything....

[trst]

4 minutes ago, DarkKnight271 said:

You do understand the difference between common, uncommon and rare in a normal game where drop rates work properly right? I shouldn't have a greater chance of getting a rare drop than an uncommon or common drop. And it shouldn't take as many or more relics to get an uncommon drop than a rare drop. That's broken. If you don't agree, that's fine, that's your opinion, but you are wrong.

Except that's not how they work? Drop rates do correspond to their displayed rarity and the rates fluctuate based on how refined a relic is.

Experiencing situations where you fall above or below an expected number of attempts to get a desired outcome is part of the nature of rng driven systems. You can go a hundred runs and not see a single 20% drop or you can get a 1% drop after a single run, that's just how it is.

If you think this is broken then you're not asking to "fix" this system; you're asking for a different system entirely.

[Jiminez_Burial]

7 hours ago, DarkKnight271 said:

It shouldn't take 15x4 relics to get an uncommon item to drop. And it shouldn't take that many for a common item to drop. You all need to fix the drop rates of relics.

I don't think you understand how statistics work.  There's something called a statistical outlier.

Edited January 1, 2019 by Starfreak911

[seprent]

1 hour ago, DarkKnight271 said:

It shouldn't take 15x4 relics to get an uncommon item to drop. And it shouldn't take that many for a common item to drop. You all need to fix the drop rates of relics.

just because you did X thing Y times where the odds say i should give you the object doesnt mean it will give you the object so what you gettting commons? rares? other uncommons you didnt want? Sounds so far the only thing broken is your luck DE cant fix that it sucks we all been there saying they are broken when your just in reality unlucky doesn't magically make them broken now if they were not dropping anything at all when you got 10 reactant or dropping stuff from a different relic entirely from the one you chose THEN it be broken 

also you wanna talk drop rates we can talk drop rates arbitration where hold on lemme check *one drop table check later* where 92% of the drop table is endo in one shape or form and the other 8% is the other 4 (2%x4) things you actually want and thats just rotation A for B its 90% endo to 10% (2.5%x4) mods endo C is 86% endo to 14% (3.5%x4) now are they broken? no are they annoying Yes if your unlucky 

Source: http://n8k6e2y6.ssl.hwcdn.net/repos/hnfvc0o3jnfvc873njb03enrf56.html

This is the drop table for everything in warframe

(for those who are concerned with clicking the link the wiki links to it too https://warframe.fandom.com/wiki/Drop_Tables )

Edited January 1, 2019 by seprent
Forgot to quote

[(XBOX)Tanta Cinta]

Warframe math is weird. Just keep running until it drops.

[deothor]

1 hour ago, trst said:

Actually it sounds like they're working as intended and you got unlucky.

I know how RNG works, but man.. Sometimes, or most the time i run Sortie, RNG proves it has issues following the % statistics.

For the last 4 months, I received about 10~ rivens* tops (not including lotus gifts or those I bought with plat). That's 120 sorties. 10~ of rivens means +/- 9-10% Sortie rewards I got was a riven.

Where's the problem? They have, supposedly 27% chance as a reward. Oh and Anasa, which also has about ~28% chance, drops, not even joking, 5-6 times per week.

I know that common "good for everything" RNG argument "hurr durr your example doesn't have strong enough base. You need bigger test sample, like 1000000000000000000 or 100000000000000000000000000000000000000000000000000000000000"... Well... Yea..... I don't. It's supposed to a game. I'm not supposed to look at it as some kind of experiment. The RNG should be leashed to act a little bit more with % chances in a shorter run. There's that type of RNG in game, where the longer you don't get something, the bigger your chances are to get it. This should be implemented in here.

Or that other, second most common RNG argument "you were unlucky". Yeah, great, nothing feels better for having bad "luck" with RNG that is almost obviously screwed up or rigged in the first place.

 

* At first I was starting to be genuinely pissed off at how rare rivens had became, so i started counting them... But because they were so rare and time between getting them was kinda long, I ended up forgetting what was the last counted Riven.... Meh. Either way definitely it wasn't above 15.

Edited January 1, 2019 by deothor

[Fire2box]

when mesa prime released me, a streamer and 2 other community members were farming for akjgara. We got the a majority of the pieces we needed even a gold or two. But we did like 14 stagggered runs for the silver/uncommon part and only got one. 

 

Meanwhile when Tiberon prime dropped with Zephyr I had traces and got tiberon rare part relics and droped them back to back with the same streamer when only I had the relics for them in the group. We got tiberon prime within the first hour. 

 

There's unlucky AF and then there's lucky AF. Can't have the good without the bad. 

Edited January 1, 2019 by Fire2box

[trst]

21 minutes ago, deothor said:

I know how RNG works, but man.. Sometimes, or most the time i run Sortie, RNG proves it has issues following the % statistics.

For the last 4 months, I received about 10~ rivens* tops (not including lotus gifts or those I bought with plat). That's 120 sorties. 10~ of rivens means +/- 9-10% Sortie rewards I got was a riven.

Where's the problem? They have, supposedly 27% chance as a reward. Oh and Anasa, which also has about ~28% chance, drops, not even joking, 5-6 times per week.

I know that common "good for everything" RNG argument "hurr durr your example doesn't have strong enough base. You need bigger test sample, like 1000000000000000000 or 100000000000000000000000000000000000000000000000000000000000"... Well... Yea..... I don't. It's supposed to a game. I'm not supposed to look at it as some kind of experiment. The RNG should be leashed to act a little bit more with % chances in a shorter run. There's that type of RNG in game, where the longer you don't get something, the bigger your chances are to get it. This should be implemented in here.

Or that other, second most common RNG argument "you were unlucky". Yeah, great, nothing feels better for having bad "luck" with RNG that is almost obviously screwed up or rigged in the first place.

 

* At first I was starting to be genuinely pissed off at how rare rivens had became, so i started counting them... But because they were so rare and time between getting them was kinda long, I ended up forgetting what was the last counted Riven.... Meh. Either way definitely it wasn't above 15.

But that is the problem: your sample size is not large enough to determine if the rates are correct or not. But all of this is simply how these systems work; the system that Warframe uses.

People can argue for a different system like a pseudorandom one (like you mentioned) or even a token system like people have asked for in the past (how the Arbitrations shop works). But this thread is about the system being "broken" which it isn't.

Edited January 1, 2019 by trst

[(PSN)OccultSif]

It's RNG, it's a percent to get certain drops with everything in this game. Do I get mad when I run radiants and only get the common drop? Yeah, do I get mad when I get 3 or 4 rare and uncommon drops from intacts? Yeah, but it's just RNG nothing I can complain about.

[deothor]

Just now, trst said:

But that is the problem: your sample size is not large enough to determine if the rates are correct or not. But all of this is simply how these systems work; the system that Warframe uses.

This argument would apply even if I either had a sample of 1000, 10000, or 100000000000. But noone will ever have it. I really doubt there's a guy with excel recording everything he got, be it relics, sortie or eidolons for thousand days. We gotta work with what we have. Sure I can forget about legendary core, or other low-chance rewards. But dude, we are talking about 27% here. It can't be as scewered as this. Not even half of what should have been.

But again, this isn't math (statistics) class, but a game.

[HellVOps]

So we went from relics rng to sorite rewards rng?

Oh and your 27% riven drop chance is false, 27% (your number)is the total sum of all the catgories with each having its own seprate chance to drop e.g. pistol riven 2%, rifle riven 6.79%, these what you should be looking for.

http://n8k6e2y6.ssl.hwcdn.net/repos/hnfvc0o3jnfvc873njb03enrf56.html#sortieRewards

And everything is math, your mods,abilities to name a few.

[deothor]

7 hours ago, HellVOps said:

So we went from relics rng to sorite rewards rng?

isn't RNG a RNG?

7 hours ago, HellVOps said:

Oh and your 27% riven drop chance is false, 27% (your number)is the total sum of all the catgories with each having its own seprate chance to drop e.g. pistol riven 2%, rifle riven 6.79%, these what you should be looking for.

That would be correct if I wanted a specific riven. But in case of me wanting just any riven, then a sum of all riven chances, 27% is correct number.

7 hours ago, HellVOps said:

And everything is math

Everything is a math if numbers you operate on are correct. Or when you know about all variables. If for example you'd have lower chance of getting a riven depending on how many rivens you already own, then no math from your side will ever cover it.

 

I know 120~days isn't alot-alot for test sample, but rivens and anasa don't even bother to act as if they work on the % numbers (27 & 28) we know they should.

IMO DE has just ninja-changed the values and didn't bother to tell us that.

 

 

[(XBOX)zThulsaDoomz]

35 minutes ago, deothor said:

IMO DE has just ninja-changed the values and didn't bother to tell us that.

 

 

Um.....no?

[DarkKnight271]

14 hours ago, trst said:

But that is the problem: your sample size is not large enough to determine if the rates are correct or not. But all of this is simply how these systems work; the system that Warframe uses.

People can argue for a different system like a pseudorandom one (like you mentioned) or even a token system like people have asked for in the past (how the Arbitrations shop works). But this thread is about the system being "broken" which it isn't.

It most definitely is broken. For starters, let's dispell the notion that Random Number Generators in a computer system are actually random. They're not. Until we have Quantum Computing I doubt they ever actually will be and even then it's not guaranteed.

That being said, does luck play a part in this, sure, but if you run a full group, all with the same radiant relic 15 times in a row any you only see three of the six possible choices come up and the "rare" drop showed up more than one of the uncommon drops that never showed up at all it tells me that there maybe something funny going on behind the scenes. Like the drop rate for that specific item may be off.

It this was truly random and adhered to statistics then sample size shouldn't matter, a sample size of 10 should have the same statistics as a sample size of 10,000. So if that uncommon drop was a 20% chance it should have dropped twice in the lower sample size and 2,000 times for the larger. If it drops 0 times in the smaller but still 2,000 times in the larger there's something wrong.

[Talinthis]

In that case you should be happy. Sell the gold rares and buy like 10 uncommons if you want them.

rng doesnt follow strict statistic percentages for drop sample sizes. Chances are it wont drop 2 out of 10 times every time like your example.

Back when i played ragnarok online I got 4 cards that had less than a 1% drop chance in less than an hour. Just because I got them doesnt mean the math is broke, and that is exactly what you are saying here. Also in ff XIV I failed a 20% materia infusion 35 times

[rune_me]

1 hour ago, DarkKnight271 said:

It this was truly random and adhered to statistics then sample size shouldn't matter, a sample size of 10 should have the same statistics as a sample size of 10,000. So if that uncommon drop was a 20% chance it should have dropped twice in the lower sample size and 2,000 times for the larger. If it drops 0 times in the smaller but still 2,000 times in the larger there's something wrong.

Not even a little bit. 

You can flip a coin 10 times and get tail all times. That happens and is cetainly statistically likely. You can not flip a coin 10.000 times and get 10.000 tails. That is so improbable that you won't ever see it happening. Which is why sample size is important, and why a sample size of 10 certainly don't yield the same results as a sample size of 10.000.

The larger the sample size, or the more you iterate the experiment, the more the result will approach the statistically most likely outcome.

[DarkKnight271]

8 hours ago, rune_me said:

Not even a little bit. 

You can flip a coin 10 times and get tail all times. That happens and is cetainly statistically likely. You can not flip a coin 10.000 times and get 10.000 tails. That is so improbable that you won't ever see it happening. Which is why sample size is important, and why a sample size of 10 certainly don't yield the same results as a sample size of 10.000.

The larger the sample size, or the more you iterate the experiment, the more the result will approach the statistically most likely outcome.

Except, in your example, you have to take into account hundreds of thousands of variables which is why you need a larger sample size. In a video game there are no random variables. Statistics are pretty much only used because of you need to find an average without being able to account for all of the variables of life. Unlike in real life, when you're in a computer environment, all variables are programmed in and specific. Like I said above there's no such thing as a truly random number generator inside of a computer system. It's a 100% impossibility at this time. So if an item has a drop rate of 20% then it should be 20% across the board. So if the item only drops at a 20% chance when you get to 10,000 or 100,000 chances but not when you're under that, the drop rate is broken.

[DarkKnight271]

9 hours ago, Talinthis said:

In that case you should be happy. Sell the gold rares and buy like 10 uncommons if you want them.

rng doesnt follow strict statistic percentages for drop sample sizes. Chances are it wont drop 2 out of 10 times every time like your example.

Back when i played ragnarok online I got 4 cards that had less than a 1% drop chance in less than an hour. Just because I got them doesnt mean the math is broke, and that is exactly what you are saying here. Also in ff XIV I failed a 20% materia infusion 35 times

I don't care about other games as they don't matter to this conversation.

As to selling the rares to buy the uncommons... yeah, that's not possible. Those "rare" drop so regularly it would take something like 50 to be able to buy even one of these "uncommon" drops.

[rand0mname]

3 часа назад, DarkKnight271 сказал:

I don't care about other games as they don't matter to this conversation.

As to selling the rares to buy the uncommons... yeah, that's not possible. Those "rare" drop so regularly it would take something like 50 to be able to buy even one of these "uncommon" drops.

What uncommon costs more then 50 rares? Even if you sell them as prime junk, it is a lot of plat.

Edited January 2, 2019 by rand0mname

[(PSN)Mono-Pop]

On 2019-01-01 at 9:34 AM, DarkKnight271 said:

You do understand the difference between common, uncommon and rare in a normal game where drop rates work properly right? I shouldn't have a greater chance of getting a rare drop than an uncommon or common drop. And it shouldn't take as many or more relics to get an uncommon drop than a rare drop. That's broken. If you don't agree, that's fine, that's your opinion, but you are wrong.

You know how the law of averages work in general, right?

If something has a 10% chance to drop, that doesn’t mean you’ll get it in 10 attempts.

It means you’re likely.

I don’t believe for a second you got 3 people to open 15 of the same relic with you.

[rune_me]

9 hours ago, DarkKnight271 said:

Except, in your example, you have to take into account hundreds of thousands of variables which is why you need a larger sample size. In a video game there are no random variables. Statistics are pretty much only used because of you need to find an average without being able to account for all of the variables of life. Unlike in real life, when you're in a computer environment, all variables are programmed in and specific. Like I said above there's no such thing as a truly random number generator inside of a computer system. It's a 100% impossibility at this time. So if an item has a drop rate of 20% then it should be 20% across the board. So if the item only drops at a 20% chance when you get to 10,000 or 100,000 chances but not when you're under that, the drop rate is broken.

No my example has nothing at all to do with taking thousands of variables into account. A coin flip has only two outcomes and there's a 50% for either. It doesn't have to be a real coin, it can be purely a thought experiment. Or if you want to use software, you can code it to be more true to a game. 

The rule still holds true. The more times you iterate the experiment, the more the result will approach the statistically most likely outcome. If you only do it 10 times, you will not get a useful result, because you could end up with a result that varies widely from the most likely. If you do it 10 million times, though, you will get a result very close to 50-50, simply because it would be so statistically unlikely to give anything else that it will never happen.

If you get rares all the time while opening relics with your friends, it doesn't mean the odds are broken. It just mean you got 10 tails in a row. Unlikely, sure, but not at all impossible. 

[(PSN)drpunk-yo]

I'm all for fixing the drop rates.

Like, I joined a rad share for a Mesa part. The host only had a Flawless equipped. Loading into the mission I called them out on it. They said others had Oked it. Whatever.

It was his relic that gave us the rare part!!!!!!!!!!!!! How's that fair?!?!?!?!?!?!?!?!?!! ☹️

P.S. While this did actually happen, my salt is sarcasm.