When did you realize the relic system is fraudulent and what do you suggest it should be replaced with ?

[MqToasty]

23 minutes ago, nslay said:

OK, so we agree that the chance of not getting the thing approaches 0 as the number of attempts approaches infinity. 

So it's impossible (probability= 0) to not get it after an infinite quantity of attempts. This can't happen. There you go, you necessarily get the thing after a finite number of attempts... 

Finite does not mean pity timer. It is not a fixed value. It is not a "small" value... finite is any value < infinity. And you know that it's impossible to not get the thing for an infinite quantity of attempts.

Okay, I think we need a different perspective to understand why you are absolutely not guaranteed to get the thing after a finite number of attempts.  Since each roll is independent, the probability of success for each individual roll is exactly 10%.  It will not change regardless of how many successes or failures you've already rolled in the past.  So with this in mind, let's examine several scenarios:

1. Your first roll: 90% chance of failure.  You will most likely not get the drop.
2. Your 10th roll after 9 failures: 90% chance of failure.  You will most likely not get the drop.
3. Your 1000th roll after 999 failures: 90% chance of failure.  You will most likely not get the drop.
4. Your nth roll after (n-1) failures: 90% chance of failure.  You will most likely not get the drop.

So even if you roll an infinite amount of times (meaning you just keep rolling and never stop), each successive roll still has a 90% chance of failure.  You will still most likely not get the drop in that roll.  It is only in retrospect that you can look at a sequence and go "that's a highly unlikely outcome".  Yet "highly unlikely" just means the probability is very, very small.  It will never actually reach 0.  And if you got a bad streak so far?  That does not affect your next attempt, still at 90% chance of failure.

[16Bitman]

34 minutes ago, SheikaVoid said:

Apologies, but, from an engineer's perspective, it really doesn't matter if the outlier in the RNG is 1/infinity, and that the chance is an asymptote approaching zero infinitely.  Mathematically, I'd never express 1/infinity as 0, but I would say, in terms of the function of computer programs and number generators, that a statistical outlier of 1/infinity is functionally nonexistent, and I don't see the need for the distinction when the functional behavior of computer programs such as these almost guarantees that the item will be acquired.

I agreed with your point in an earlier post.

You, as an engineer, are pretty much the personification of the "Practical" side of the argument.

Calculating measurements beyond mm or even μm, is entirely pointless for most practical purposes.

Absolutely nobody but the mathematician, cares for this super low measurements.

 

"I don't see the need for the distinction when the functional behavior of computer programs such as these almost guarantees that the item will be acquired."

There is no need. This is all just to explain math. Which at certain points stops having a practical real world value. There are certain points in equations in which you should just stop and round.

 

37 minutes ago, nslay said:

OK, so we agree that the chance of not getting the thing approaches 0 as the number of attempts approaches infinity. 

So it's impossible (probability= 0) to not get it after an infinite quantity of attempts. This can't happen. There you go, you necessarily get the thing after a finite number of attempts... 

Finite does not mean pity timer. It is not a fixed value. It is not a "small" value... finite is any value < infinity. And you know that it's impossible to not get the thing for an infinite quantity of attempts.

Infnity is not a number, it's just a concept.

"0.9 ^ INFINITY = 0.0000000(TheNumber of Zeroes here are also infinite, but they are still bookended by a value above 0 at the end)"

 

In a practise, you are right.

In practise, people quit after one or two decimals.

Practise is the only thing that matters for our real world.

Trying to see what is beyond the practical possible, is nothing but a thought exercise and a waste of time.

 

But

In theory, which quickly stops being practical, which hits the thought of purely fictional examples as we can no longer find a single person on our planet that fits the example, you never hit the 0 chance to fail.

 

 

Math is fun. Math is fun because there are no opinions. Equations are either right or they aren't math, but just random numbers and symbols.

 

This is the very best way I can explain this and I have nothing to add, I am not a math teacher and my capability to explain ends here.

[Prexades]

Classic gamblers fallacy from the OP.

If you have a 10% chance to get an item from a relic, It will still be 10% on the second attempt, 10% on the third attempt and so on. By increasing the number of relics you increase the probability of one roll being the the one you seek and the probabability will increase the more relics you use, but it will never reach 100%.

 

 

 

[SheikaVoid]

Really, putting theory aside, there's little more to this entire debacle than, as Prexades said, Gambler's Fallacy and misunderstanding statistics.

Over time, even with low droprates, you will almost certainly receive a drop in practice, but, because it's just 10% (or, I think OP's was 8.5%) per roll, your chances are just X% per roll.

There's some argument to be made for weighted RNG, which adjusts depending on number of rolls, but that's another topic for discussion.

As an aside, I think that, in practice, weighted RNG is not necessary in Warframe, but, perhaps, there's an argument to be made for an increase of dx per roll, bringing it in line with what people "expect", though this can only come as a result of a system where you're rolling for a specific item.

[nslay]

3 minutes ago, 16Bitman said:

I agreed with your point in an earlier post.

You, as an engineer, are pretty much the personification of the "Practical" side of the argument.

Calculating measurements beyond mm or even μm, is entirely pointless for most practical purposes.

Absolutely nobody but the mathematician, cares for this super low measurements.

 

"I don't see the need for the distinction when the functional behavior of computer programs such as these almost guarantees that the item will be acquired."

There is no need. This is all just to explain math. Which at certain points stops having a practical real world value. There are certain points in equations in which you should just stop and round.

 

Infnity is not a number, it's just a concept.

"0.9 ^ INFINITY = 0.0000000(TheNumber of Zeroes here are also infinite, but they are still bookended by a value above 0 at the end)"

 

In a practise, you are right.

In practise, people quit after one or two decimals.

Practise is the only thing that matters for our real world.

Trying to see what is beyond the practical possible, is nothing but a thought exercise and a waste of time.

 

But

In theory, which quickly stops being practical, which hits the thought of purely fictional examples as we can no longer find a single person on our planet that fits the example, you never hit the 0 chance to fail.

 

 

Math is fun. Math is fun because there are no opinions. Equations are either right or they aren't math, but just random numbers and symbols.

 

This is the very best way I can explain this and I have nothing to add, I am not a math teacher and my capability to explain ends here.

OK... how about we use the law of large numbers.

Suppose the probability to get thing is 10%. Let X_i reflect the outcome ... 0 is not getting the thing while 1 is getting the thing.

 By the LLN, lim i to inf 1/i Sum X_i = 0.1

Now suppose somebody never ever gets the thing. That means X_i = 0 for i=0, 1, 2, ...

 But this contradicts LLN since then 1/i Sum X_i converges to 0. This is absolutely impossible since it actually converges to 0.1.

There must be some X_i = 1.

[Elenortirie]

If relics are fraudulent, then I am unsure if I want to see OP's opinion on the system they were implemented to replace.

For those who don't remember, drop tables for void keys were extremely diluted and that was when amount of prime gear was about half what it is rn.

[SheikaVoid]

3 minutes ago, nslay said:

OK... how about we use the law of large numbers.

Suppose the probability to get thing is 10%. Let X_i reflect the outcome ... 0 is not getting the thing while 1 is getting the thing.

 By the LLN, lim i to inf 1/i Sum X_i = 0.1

Now suppose somebody never ever gets the thing. That means X_i = 0 for i=0, 1, 2, ...

 But this contradicts LLN since then 1/i Sum X_i converges to 0. This is absolutely impossible since it actually converges to 0.1.

There must be some X_i = 1.

Weak or strong proof?

[nslay]

Just now, SheikaVoid said:

Weak or strong proof?

Bad proof lol.

[SheikaVoid]

Just now, nslay said:

Bad proof lol.

Lol.

[SheikaVoid]

1 minute ago, nslay said:

Bad proof lol.

I have to agree though, because if you take Khinchin's Law, that seems to be the expected, theoretical result at a value for X.

[16Bitman]

6 minutes ago, SheikaVoid said:

Over time, even with low droprates, you will almost certainly receive a drop in practice, but, because it's just 10% (or, I think OP's was 8.5%) per roll, your chances are just X% per roll.

 

On 2021-06-03 at 4:15 PM, White_Matter said:

4 radiant relics mean I should have 34% chance to have at least 1 rare item.

Since the reactions to this statement, is what sparked my posts, I have to point out that OP did not make a mistake and used the correct 10% drop chance for their calculation.

 

They just rounded it down to a number with no decimals.

Their calculation was NOT 8.5% + 8.5% + 8.5% + 8.5% = 34% to drop

It was this:

0.9 ^ 4 = 0.6561

1 - 0.6561 = 0,3439

0,3439 * 100 = 34,39

Which we can call 34,39% chance to get a rare part.

Now OP just rounded all decimals away to provide the 34% chance in the post I quoted.

 

Additive math in chance calculation for an event occuring at least once, is wrong.

[SheikaVoid]

Just now, 16Bitman said:

 

Since the reactions to this statement, is what sparked my posts, I have to point out that OP did not make a mistake and used the correct 10% drop chance for their calculation.

 

They just rounded it down to a number with no decimals.

Their calculation was NOT 8.5% + 8.5% + 8.5% + 8.5% = 34% to drop

It was this:

0.9 ^ 4 = 0.6561

1 - 0.6561 = 0,3439

0,3439 * 100 = 34,39

Which we can call 34,39% chance to get a rare part.

Now OP just rounded all decimals away to provide the 34% chance in the post I quoted.

 

Additive math in chance calculation for an event occuring at least once, is wrong.

Ah, good to know, I had assumed, from other comments, and having not read diligently enough, that they had simply split the difference and added, but that gives me a bit more hope.

[Sevek7]

20 hours ago, Ailia_Grimm said:

I already told them that the chances weren't additive to each other, even with 4 relics, the chances for each one are still individual

Maybe I'm just expressing myself incorrectly ¯\_( :/ )_/¯

I just read through this entire thread. Looks like you're having an argument with an imaginary person. Nobody in the entire thread has suggested that the probabilities are additive, everybody agrees that the individual relics are rolled separately and each maintains the same chance of a rare (10% on radiant).

If we lived in a world where the probabilities were additive, then they would add to 10%+10%+10%+10% = 40%. As you correctly pointed out, we don't live in this world. Therefore we need to use the correct formula (as provided by @16Bitman in the long post that you freely admitted you didn't read). When using this correct formula (chance of success = 1- chance of failure, where chance of failure is 0.9^4 [Notice: not additive!]), you get a ~34% chance of getting at least one rare part in a radshare with 4 players. Notice, again, this is not additive, 10+10+10+10 does not equal 34. 

So when you claimed that this formula was wrong because it was additive, you were wrong because it's not additive. Nobody is arguing with you that it should be additive, everybody is trying to teach you to read the post so you can see for yourself that it's not additive.

If you're still sure that you're right, please go back to @16Bitman's second post (the long post you said you didn't read) in the thread where all the math is laid out in detail. Point me to the place where the probabilities are being added. If you're right that his formula is additive, this should be very easy.

[Sovyul]

lol

[SheikaVoid]

All this because someone didn't like their Prime part drops...

Also, if anyone's interested in some neat math that's somewhat related, I recommend looking into Kolmogorov's Law, fascinating stuff, and you're sure to learn some beautiful things about math!

[(NSW)Rafa-el]

En 3/6/2021 a las 16:15, White_Matter dijo:

4 radiant relics mean I should have 34% chance to have at least 1 rare item. In 7 runs I didn't get a single rare item and 70% of the drops were common items.

Every run reset the percentage 

[SheikaVoid]

3 minutes ago, (NSW)Rafa-el said:

Every run reset the percentage 

RNG isn't run-dependent, the roll is the same percentile regardless.

[TeaHands]

Of course all this math answers a lot of questions involving this games RNG. But did you know? The best way to up your chances of getting the prime part you need is by switching controllers (or keyboards) every run to reset your luck. Also make note of the position of your feet, that helps too.

[Buff00n]

Oh good a math thread!

I feel like people are making this more complicated than it needs to be.

The OP ran 7 radshares and came up empty.  That means they witnessed a total of 4*7 = 28 radiant relics fail to get the rare drop.

Each relic is an independent drop.  With each one, a 10% chance of the rare drop = 90% chance of failure.  The chance of getting a 90% result 28 times in a row => 0.9^28 => about 5%, or 1 in 20.

In a game with millions of players, a 1/20 occurrence happens thousands of times, and this time OP is one of them.

My personal records are 13 radshares for a Galantine Prime part and 20 for an Aklex Prime part back in the day, runs of bad luck that have a 1/150 chance and 1/3000 chance, respectively.  It's going to happen to somebody, and sometimes that somebody is you.

[-CdG-Zilchy]

TLDR - OP opens a pitiful 7 radiant relics, doesn't get the drop he wants, gets mad and throws a tantrum causing him to create a rant post which is then set upon by mathematicians, all the while OP is ignoring the fact he just got 7 prime parts he can sell to buy the part he didn't get... 

^ this is how you balance relic RNG OP, learn it.

[(PSN)thowed]

Did 3 rad shares on Gara release day and got 3 BPs.  Not saying this is common because it rarely happens to me.  It does however prove the rng nature.

[16Bitman]

9 hours ago, -CdG-Zilchy said:

TLDR - OP opens a pitiful 7 radiant relics

He did 7 radshares, so that's 28 relics.

The chance of not getting a rare was 5.23%, which is pretty low, but still very believeable and not worth creating a thread and calling the official drop tables fraudulent.

 

Maybe I misunderstood and you actually think 7 radshares is pitiful, then you can just ignore this post :)

[SneakyErvin]

On 2021-06-03 at 4:15 PM, White_Matter said:

4 radiant relics mean I should have 34% chance to have at least 1 rare item. In 7 runs I didn't get a single rare item and 70% of the drops were common items.

I love that kind of math. It even cracks me up, a person that is genrally really bad at math.

Chance is always static, probability changes, the two are not interchangable.

edit: Iirc you need to run 70+/-20 run or well crack that amount of relics to be nearly guaranteed atleast 1 drop if the chance is 10%.

[-CdG-Zilchy]

8 minutes ago, 16Bitman said:

He did 7 radshares, so that's 28 relics.

The chance of not getting a rare was 5.23%, which is pretty low, but still very believeable and not worth creating a thread and calling the official drop tables fraudulent.

 

Maybe I misunderstood and you actually think 7 radshares is pitiful, then you can just ignore this post :)

Yes to the second part, I meant that only doing 7 radshares is rookie numbers.

[LillyRaccune]

Wasn't there a scandal a few years ago about the drop rate of Forma blueprints? Could that have anything to do with this situation?