Farming Ephemera from Lich / Sisters and 20% fallacy

[(PSN)magusat999]

After wearing my hands and my patience farming a fire based Sister... Too many times to count, I put on my tinfoil muffhat on and started thinking about these odds... Im not anywhere mathmatical, so please be aware if you choose to start down that road. Its just my thought in the subject.

I call it a fallacy because its not a "true" 20%, imho. 20% is supposed to equate to one in five. But, often when some entity states a percentage, they are not stating the truth of what it really means (or should mean). In this case, 20% isnt a true 20% but a re-occuring odd that resents and forgets after each attempt. The way a "pure" 20% shoukd work is 1 out of 5 attempts should give the prize. The odds are supposed to have come from calculating real world cases, until a reliable odd can be determined and announced.

A true 20% could work like this - 5 dice with ascending amount if sudes with a winning symbol. Each die has one more, until the last, which has wins on all sides! You get one cast per die. The 20% is spread among the dies, not "per die", where your real odds are -80%! At the end, you win, because otherwise its not 20% anymore. By comparison , what we are being told is 20% is actually -80%, with infinite chances to lose. We never get past that first die, and you can keep throwing it an infinite amount of times, never winning. Thats not 20%, thats GACHA. To call that 20% is deceptive and disingenuous - not that deception is on purpose, quite likely they actually believe its really 20%.

[(PSN)EntityPendragon]

Soooo...what 20% are we talking about here. Bit much math for me to understand. More what this 20% is about. Are we talking about stabs idk. I got stupid juice rn

[rapt0rman]

What you're describing is 100% chance with 2-5 extra steps, you do have 20% chance on the first attempt, but the similarity stops there.

1 in 5 chances per attempt is inherently not a guarantee you will get it after 5 attempts nor 500.

Don't get me wrong though, I'll happily accept any pity systems they add.

 

To add a bit more to this since I was curious of what the actual terms were: Discrete probability distribution is for finite/countable outcomes, and Continuous probability distribution is for infinite. They're both applications of probability, but with very different uses.

Edited September 27, 2023 by rapt0rman

[CrownOfShadows]

Yes this is a problem with ALL of DE's RNG. I feel your pain, having farmed all of the sister ephemeras the hard way. I've asked multiple times for the ephemeras to simply show like the weapons do, so you know if you're wasting your time or not. It's enough work going through the process as it is,s tacking RNG on top of that is so bad.

Thank god they've finally given players a modicum of pity here and there, like with the archon shards, but pity systems should exist everywhere RNG does, from relics to ephemera to part drops. RNG is simply miserable. It only works in your favor when you don't care, which is a terrible system. You should reward players for caring and chasing things, not punish them.

[Numerounius]

It is a 20% chance with no fallacy/"true" term to adhere to it, here is how it works:

Chances you are to never get the ephemera after X tries: (1-N)^X

Where N is drop rate chance and X is attempts. For example, after 5 attempts you have about a 33% chance of not getting the epherema which is a very reasonable amount and goes way down the more attempts poured into it. Yes there is a chance that you will be in the 0.1% for some of the drop rate items in Warframe, which you can find out based on the above equation instead of just assuming, but you aren't going to be that massively unlucky on everything in this game. This can be done for any drop chance item in Warframe. The equation is basic statistics and doesn't require extensive mathematical knowledge to prove the above equation works.

Just be grateful that the drop rates are provided, if you want to claim that this is Gacha and predatory go tell that to a game that doesn't have their droprates officially posted and requires real-world currency to obtain the drop rate item.

[taiiat]

Chances are Chances. nothing more to it.
20% is always 20%, it's always a """true""" 20%.

guaranteed rates by turning Chance into a 'run requirement' would get People stuff faster on average, but ofcourse then People would be complaining there's nothing for them to get, to play the game for.

[quxier]

3 hours ago, rapt0rman said:

What you're describing is 100% chance with 2-5 extra steps, you do have 20% chance on the first attempt, but the similarity stops there.

1 in 5 chances per attempt is inherently not a guarantee you will get it after 5 attempts nor 500.

Don't get me wrong though, I'll happily accept any pity systems they add.

 

To add a bit more to this since I was curious of what the actual terms were: Discrete probability distribution is for finite/countable outcomes, and Continuous probability distribution is for infinite. They're both applications of probability, but with very different uses.

If you have to use "fancy" name I would use combinatorial analysis (couldn't find simple explanation) more than probability distribution (even discrete one). Probability distribution says about probability that something occurs (e.g. 20%) Combinatorics (another term, meaning the same) is about data (e.g. numbers) and how to sort/pick them.

But to be honest I'm not sure if there is exact term in math for "N% + X*number-of-runs". WoW used "progressive percentages".

2 hours ago, taiiat said:

 

guaranteed rates by turning Chance into a 'run requirement' would get People stuff faster on average, but ofcourse then People would be complaining there's nothing for them to get, to play the game for.

That's not true. Now you want to get as fast as possible because you don't know how many runs you need to do. Hence faster runs = potentially quicker way to obtain reward.

4 hours ago, (PSN)magusat999 said:

A true 20% could work like this - 5 dice with ascending amount if sides with a winning symbol. Each die has one more, until the last, which has wins on all sides! You get one cast per die. The 20% is spread among the dies, not "per die", where your real odds are -80%! At the end, you win, because otherwise its not 20% anymore. By comparison , what we are being told is 20% is actually -80%, with infinite chances to lose. We never get past that first die, and you can keep throwing it an infinite amount of times, never winning. Thats not 20%, thats GACHA. To call that 20% is deceptive and disingenuous - not that deception is on purpose, quite likely they actually believe its really 20%.

That's why I hate base N% chance without any modifier (e.g. like Archon shards that adds 10-20% up to guaranteed 100%) usage in games, at least in important thing. 50% for some player will mean 1% for other. If it works then it's fine. Even 2x more runs required (20% chance > 5 runs > 10 runs max) would be fine. Then you add randomness on top of randomness + some longer grind. I despise Riven system. It thrives on players "unknowledge" & wasting time. With Slivers: Eximus/Sliver chance > Iron lady chance (e.g. melee vs zaw) > do some challenge. After that time you either get some good (e.g. for weapon you use or it's good) or just waste of time. And you have to wait 1 week for that.

[taiiat]

17 minutes ago, quxier said:

That's not true. Now you want to get as fast as possible because you don't know how many runs you need to do. Hence faster runs = potentially quicker way to obtain reward.

i don't know what you're trying to say.
but in practice, it'd probably be 20% actual Chance, vs '20 runs to get', ofcourse, it wouldn't be '5 runs to get'.

but uh, is that any better? not really. now you're guaranteed a much higher minimum time, with not much reduction to the maximum. this just forces other Players, like me, to play more than they want to play. 
sounds like just making things more tedious than they already are. similarly, if it was 20 runs to get something, means that the first 19 times you play it, minimum possible effort to qualify because there's no reason to be there for more than what's required. only the 20th one would actually be meaningful, the rest were just padding.
ala XP Bars, and XP bars are generally one of the worst parts about Video Games.

[rapt0rman]

1 hour ago, quxier said:

If you have to use "fancy" name I would use combinatorial analysis (couldn't find simple explanation) more than probability distribution (even discrete one). Probability distribution says about probability that something occurs (e.g. 20%) Combinatorics (another term, meaning the same) is about data (e.g. numbers) and how to sort/pick them.

But to be honest I'm not sure if there is exact term in math for "N% + X*number-of-runs". WoW used "progressive percentages".

Yeah... The more I read up on it the less sure I was, and then got distracted after not finding a satisfying answer.

[(PSN)slightconfuzzled]

I'd like to focus on your disclaimer of sorts first for a moment. Your line about not being mathematical. For a few reasons, first being I really appreciate and admire that. Not a lot of people can be comfortable with talking about not being the best at something, people often think that its some sort of massive flaw and so better to just pretend and fake confidence and understanding. Then also, well, many people aren't really that good with maths or statistics. Myself included. I had bad maths teachers that put me off the subject, and whilst I grew to appreciate it later, I still struggled with quite a few "basic" concepts, and whilst I had good intuition in some areas, which helped my understanding progress, my intuition in others, created struggle.

For example, it took me a while to understand that 1, 2, 3, 4, 5, 6, 7 had the same chance to win a Lottery as any other "random" number on non sequential order. To my understanding "it didn't make sense", but then I had to sort of remember, that to reality, the actual numbers on the little balls didn't matter, it was us as humans, assigning those designations to the numbers. To the big jumble cylinder throwing the little balls around, they are all the same. Many intelligent, articulate, reasonable adults though, still struggle with this idea though, as I have seen a few times with viral videos. There is nothing wrong with struggling with such concepts though, often its more about the education system, and also many don't learn, because they are unwilling to admit or acknowledge they still have room to learn, since like I said, many people often prefer the fake confidence or fake understanding idea. For me personally, I constantly remind myself that statistics can be deceptive, so even if something seems simple with statistics... I should be humble, careful and willing to take on more information and explanations from those who may be more advanced in understanding them. Oh and side note, I don't recommend anyone partake in lotteries, for a few reasons, and even though 1, 2, 3, 4, 5, 6, 7 is as good as any other number, its actually bad for a different reason, since a lot of people pick that number, so your winnings would potentially be diluted, but then again the chances of winning anyways...)

I also know that this forum does have quite a few people who are pretty knowledgeable and good with statistics as well, I am always willing and favourable to being corrected or pointed in the right direction over such subjects. Its nothing so personal in the end. 

So with all that being said, and acknowledging what you said... I don't think you should use the term fallacy. Rather and I don't mean this in a mean way, but its more like your understanding is flawed. Take a coin and flip it 10 times. Do it as many times as you are willing to entertain this experiment, how many times did you get 5 heads and 5 tails? How many times did they alternate? Are you noticing that occasionally you get "unusual" streaks of either heads, or tails, recurring? Like 4 heads in a row, or 4 tails? A lot of what we are doing is just observing... As in coins don't have memories, and there is no magic external force trying to adjust or correct the amount or ratio involved. If that makes sense? 

Now I do think I understand what you are trying to do, when introducing ideas like a ""pure 20%" and a "true 20%" just the terminology is a bit... it seems like you are using those terms based on your perception and understanding? Thing is, statistics don't exist for video game rewards, our understanding of it is much larger and important. The frustrations you are experiencing are because of how it is, not some impurity, misunderstanding or misapplication. 

That frustration is also valid and worth highlighting and considering. Though in that sense, we actually already have terminology for what you are asking for. Its a pity system. Or sympathy system. Since RNG implementation can be harsh in games, since statistics can be deceiving, some people will implement pity/sympathy systems, to account for bad luck and severe bad luck, to encourage tangible progress and offset frustrations that can arise with people, especially unlucky people. Its not always implemented per say, because there are various pros and cons to both design philosophies. Going with just RNG, and incorporating pity systems alongside RNG. 

How the pity/sympathy system is implemented can differ, your suggestion for example, but other ways too. 

Cheers! 

[quxier]

13 hours ago, taiiat said:

13 hours ago, quxier said:

That's not true. Now you want to get as fast as possible because you don't know how many runs you need to do. Hence faster runs = potentially quicker way to obtain reward.

i don't know what you're trying to say.
but in practice, it'd probably be 20% actual Chance, vs '20 runs to get', ofcourse, it wouldn't be '5 runs to get'.

but uh, is that any better? not really. now you're guaranteed a much higher minimum time, with not much reduction to the maximum. this just forces other Players, like me, to play more than they want to play. 
sounds like just making things more tedious than they already are. similarly, if it was 20 runs to get something, means that the first 19 times you play it, minimum possible effort to qualify because there's no reason to be there for more than what's required. only the 20th one would actually be meaningful, the rest were just padding.
ala XP Bars, and XP bars are generally one of the worst parts about Video Games.

Is it better? Not for everyone.

If you are lucky (get an item once every 5 runs) then there might be some slight shift in minimum runs required.

If you are NOT lucky (get an item after e.g. 20 runs) then it would be better.

However that's not the point of it. It's not about getting better or worse. It's about informed decision(s) without wasting players' time (at least not too much). You see an item, time required (e.g. number of missions, difficulty etc) and see that you need max N runs. XP bars are much better than base chances (%).

 

And as for ideal number of runs (5) vs 20 (or some high number). Look at Archon shards. You get like 5 runs afair.

 

[(PSN)magusat999]

18 hours ago, rapt0rman said:

What you're describing is 100% chance with 2-5 extra steps, you do have 20% chance on the first attempt, but the similarity stops there.

1 in 5 chances per attempt is inherently not a guarantee you will get it after 5 attempts nor 500.

Don't get me wrong though, I'll happily accept any pity systems they add.

 

To add a bit more to this since I was curious of what the actual terms were: Discrete probability distribution is for finite/countable outcomes, and Continuous probability distribution is for infinite. They're both applications of probability, but with very different uses.

This is why I used the term "true 20%" - because if you say 20%, your saying 1 in 5 - are you not? Now you brought up a good point - one in five "per attempt" or one in five "per 100% sets of tries"? If we are talking per attempt, where each time you fight the Sister, your odds are reset and the game forgets your earlier failed attempts - is that really 20% chance at getting the drop? If you have tried more than 5 times and have not received the drop - can that be called 20%. If you put that into an equation, it surely doesn't equal 20%, so how can that be called 20%. Here's my attempt at an equation of such: (100.00 / 5.00) ≠ (100.00 / 28.00) On both sides of the equation, the divisors (5.00 and 28.00) are the attempts at getting the drop. 100 is the dividend, obviously - the total percentage the odds are based on. So the question is - how is 5 attempts and 28 attempts both 20 percent? I dont see why people argue that 20% can be any number of attempts - because it's 20% per run. Is that really 20%?

[taiiat]

2 hours ago, quxier said:

However that's not the point of it. It's not about getting better or worse. It's about informed decision(s) without wasting players' time (at least not too much). You see an item, time required (e.g. number of missions, difficulty etc) and see that you need max N runs. XP bars are much better than base chances (%).

And as for ideal number of runs (5) vs 20 (or some high number). Look at Archon shards. You get like 5 runs afair.

make it a Shop that can be ignored using Currency that can be ignored, make it something invisible that doesn't get in the way of existing Chances, and then do whatever you want.

or Trade with someone else, most of the time - that's the intended design after all, you get a bunch of stuff even if it's not what you wanted and so you can Trade with someone else to get what you're looking for, and the game profits off of that - part of the Business Model after all

 

51 minutes ago, (PSN)magusat999 said:

one in five "per attempt" or one in five "per 100% sets of tries"? If we are talking per attempt, where each time you fight the Sister, your odds are reset and the game forgets your earlier failed attempts - is that really 20% chance at getting the drop? If you have tried more than 5 times and have not received the drop - can that be called 20%.

these are the same thing.
whether it's a single procedure or 20,000, the Ratio is always the same.

the only way to scale a Chance based data set is to increase the procedure counts - and the larger you make a dataset, the closer the observed Ratio will be to the absolute Ratio.

[rapt0rman]

3 hours ago, (PSN)magusat999 said:

This is why I used the term "true 20%" - because if you say 20%, your saying 1 in 5 - are you not? Now you brought up a good point - one in five "per attempt" or one in five "per 100% sets of tries"? If we are talking per attempt, where each time you fight the Sister, your odds are reset and the game forgets your earlier failed attempts - is that really 20% chance at getting the drop? If you have tried more than 5 times and have not received the drop - can that be called 20%. If you put that into an equation, it surely doesn't equal 20%, so how can that be called 20%. Here's my attempt at an equation of such: (100.00 / 5.00) ≠ (100.00 / 28.00) On both sides of the equation, the divisors (5.00 and 28.00) are the attempts at getting the drop. 100 is the dividend, obviously - the total percentage the odds are based on. So the question is - how is 5 attempts and 28 attempts both 20 percent? I dont see why people argue that 20% can be any number of attempts - because it's 20% per run. Is that really 20%?

But 20% per run is the only thing it ever claimed to be. You can calculate a new probability based off that after x amount of runs, (and yes, that number definitely won't be 20%) but the math will still rely on that 20%.

Ya know... real, physical dice do exist right? Yes, average probability changes every time you roll the dice, but dice have never "cared" about what your previous roll was, each roll is always a new roll with a 16.667% (1 in 6) chance.

If you take away from the pool after the first 1 in 5 chance... then it just isn't 20% any more. That initial 20% figure is now gone, because 1 in 4 chances is 25%. There isn't really anything "true 20%" about it. If we're talking "true 20%" then you wouldn't even get a second try, let alone 5 (or infinite) attempts. You would get one single 1 in 5 chance, and if you blew it, that would be it.

The main thing is that the system you described already exists for tauforged archon shards, and doesn't require redefining established mathematics to achieve it :P

Edited September 27, 2023 by rapt0rman