Jump to content

DON'T GIVE UP/ shiny rate is actually fair


Recommended Posts

4 minutes ago, blankf4ce said:

I don't know if I would necessarily agree with that assumption, in theory one in every 30,000 players would encounter a shiny on their first encounter. Once you have not encountered a shiny on your first encounter, you could hypothetically separate them into two groups, those who have encountered a shiny already, and those who haven't. As time goes on you would find that the person who encountered a shiny one their first encounter would have already encountered their shiny and thus would have to do a lot more encounters to find another one, in theory at least. Those who didn't find a shiny on their first encounter would by default be more likely to find a shiny in the future, since they have not found one in the past. This would really come down to the way the game is coded. If the game is coded to give out one shiny per 30,000 encounters, and 1 shiny per 27,000 encounters to those with donators, the only thing that would matter would be the timing of your encounter, not how many encounters you've done. It would be predetermined which encounter would be a shiny, it would just come down to getting an encounter at that exact time. Why would you not add each encounter to the aggregated probability? Even if not for the next encounter, but for future encounters in general? As you are encountering pokemon the present is becoming the past and you are moving into the future, you seem to be agreeing that you can find the probability of past events occurring using the method I used, the disagreement is coming from the lack of a way to actually quantify time or to quantify what is possible or impossible in regards to it. I also have not been saying they are independent, this is just operating under the assumption that they are. 

Wherever you went to school, they did not properly explain to you what independent means. Independent means that it does not depend on another encounter (i.e. past events have absolutely zero bearing on future events). Meaning that the person who found a shiny on encounter 1, is just as likely to find a shiny on encounter 2 as someone who did not find a shiny on encounter 1. You are mixing aggregated odds with individual odds, and that’s tripping you up.

 

To give myself a bit of credibility here, I will say that I majored in statistical finance (really called quantitative finance) and minored in mathematics (focus on theoretical math) with two classes focused on probability.

 

The reason that you don’t include the past event into the 1-(29,999/30,000)^X equation, is because that past event is no longer random. If you include it into the equation you are literally saying that past event has a 1/30,000 chance to be shiny. But, that’s just not true, you know for a fact it has a 0% chance to be shiny and that it is 100% not shiny.

 

Like I said before, if you know for a fact that a shiny exists in X encounters then what you are saying works, because in that case the individual encounters are no longer independent from one another. It seems like you are assuming that if you encounter 100,000 pokemon that the chances of not finding a shiny are so small that you can practically assume there is at least 1 shiny in that population. But, you just can’t do that. You can only do that if the % becomes 100 exactly, and that only happens at infinity.

Link to comment

Let's be real I stopped reading when I saw you never use the "Enter" key while writing a post. 

At least those threads have that good old ritual

1) Look for the 4F post

2) Hit the like button

3) ???

4) Profit ! 

5) Leave the thread

Edited by TohnR
Link to comment
50 minutes ago, blankf4ce said:

Like you said the math is there and is calculated correctly, I was also taught this at an American University which I will not disclose the name of, since it appears I was taught wrong. You are not really telling me I'm wrong, but calling out the entire field of statistics as being a sham, which I said in an earlier post that I would agree with, since there isn't really a way to accomplish independent odds. Most if not all things are dependent on something else, whether or not we are able to quantify them, and therefore the probability would have to be altered to account for said variables. 

What the...
What do you mean by calling an entire field of statistics as being a sham...

First of all, statistics means analyzing the data provided by past events such as the example I gave with devices I was testing at work, and probability calculation is predicting of future events considering options and their likelihood to happen which is our main topic here.

This is not a job application so your education level holds little importance. These are globally accepted methods of calculations. If you ask my opinion, I don't think you were taught wrong but you are applying wrong calculation formulas to wrong areas. I strongly advise you to review your notes from your university lessons if you still have them.

Link to comment
35 minutes ago, blankf4ce said:

don't know if I would necessarily agree with that assumption, in theory one in every 30,000 players would encounter a shiny on their first encounter. Once you have not encountered a shiny on your first encounter, you could hypothetically separate them into two groups, those who have encountered a shiny already, and those who haven't. As time goes on you would find that the person who encountered a shiny one their first encounter would have already encountered their shiny and thus would have to do a lot more encounters to find another one, in theory at least. Those who didn't find a shiny on their first encounter would by default be more likely to find a shiny in the future, since they have not found one in the past.

This is just mathematically incorrect and not how independent probabilities work. Take a dice roll for example, if I have rolled a '6' and my friend rolled a '2', it doesn't make it any more likely for my friend to roll a '6' just because he hasn't rolled before and the chance of either us rolling a 6 is still 1/6. In this case, you can think of finding a shiny as rolling a 30,000 sided dice to land on a particular number. The chance of finding a shiny is 1 in 30000 and does not guarantee a shiny every 30000 encounters. Some lucky players might find 3 or 4 shiny's in 30,000 encounters of hunting where as it may take some players 100,000 plus to find their first. Whilst extremely unlikely, it not mathematically impossible for a player to go say 1000000000 quadrillion encounters without a shiny.

Link to comment
4 minutes ago, Gilan said:

Wherever you went to school, they did not properly explain to you what independent means. Independent means that it does not depend on another encounter (i.e. past events have absolutely zero bearing on future events). Meaning that the person who found a shiny on encounter 1, is just as likely to find a shiny on encounter 2 as someone who did not find a shiny on encounter 1. You are mixing aggregated odds with individual odds, and that’s tripping you up.

 

To give myself a bit of credibility here, I will say that I majored in statistical finance (really called quantitative finance) and minored in mathematics (focus on theoretical math) with two classes focused on probability.

 

The reason that you don’t include the past event into the 1-(29,999/30,000)^X equation, is because that past event is no longer random. If you include it into the equation you are literally saying that past event has a 1/30,000 chance to be shiny. But, that’s just not true, you know for a fact it has a 0% chance to be shiny and that it is 100% not shiny.

 

Like I said before, if you know for a fact that a shiny exists in X encounters then what you are saying works, because in that case the individual encounters are no longer independent from one another. It seems like you are assuming that if you encounter 100,000 pokemon that the chances of not finding a shiny are so small that you can practically assume there is at least 1 shiny in that population. But, you just can’t do that. You can only do that if the % becomes 100 exactly, and that only happens at infinity.

I'll just word it like this, if the question of what is the probability of getting tails on a coin flip after getting heads? was on a test, the correct answer would've been 3/4, actually I'm pretty sure it was on a test. I vaguely remember testing this in class and they did find that after flipping a coin and receiving heads about 1/4 of the class got heads again. However, one thing I can say I learned from the class was that correlation does not equal causation. Luck is impossible to quantify. From my limited, but holistic understanding of stats and videogames/coding, I would say that there is probably some sort of input/output in the coding of the game that would make the chances of encountering  a shiny not independent, but probably based on something like the time, down to the millisecond or more. That's kind of how random number generators work, they're only random to us because the algorithm is so weird and complex that it produces a wide variety of outputs. A computer needs two things in order to do anything, a program (function) and for you to do something to it (input), at least until AI is implemented which scares the crap out of me but oh well. The fact that there is an input would make the assumption that they are independent odds false.

Link to comment
1 minute ago, blankf4ce said:

I'll just word it like this, if the question of what is the probability of getting tails on a coin flip after getting heads? was on a test, the correct answer would've been 3/4

Uuuuuuuuuuuughhhhhhh......
Correct answer is 1/2. Your first roll being heads does not affect your next toss.
Also if the question was what are the chances of getting heads then tails in 2 consecutive tosses, the answer would have been 1/4.
I don't know what are you talking about here.
 

 

2 minutes ago, blankf4ce said:

At least until AI is implemented which scares the crap out of me but oh well.

Side note,
AI is not any different than
"A computer needs two things in order to do anything, a program (function) and for you to do something to it (input)"
you stated, even in cases of ML.

Link to comment
9 minutes ago, PrincessDia said:

What the...
What do you mean by calling an entire field of statistics as being a sham...

First of all, statistics means analyzing the data provided by past events such as the example I gave with devices I was testing at work, and probability calculation is predicting of future events considering options and their likelihood to happen which is our main topic here.

This is not a job application so your education level holds little importance. These are globally accepted methods of calculations. If you ask my opinion, I don't think you were taught wrong but you are applying wrong calculation formulas to wrong areas. I strongly advise you to review your notes from your university lessons if you still have them.

Well, what is the point of analyzing the probability of past events occurring, if they hold no impact on future events? maybe sham is the wrong word, but yes this is exactly what I was taught, and yes looking at the example you gave it did seem like you were calling BS on the whole stats thing, which I wouldn't necessarily disagree with. You said that based on stats you should've seen values around 25, but instead got values of plus or minus infinity, that's a very large margin of error. In the real world, statistics are often used to mislead people, and nothing else. I guess you could use statistics to quantify demand in the past, that's kind of cool I guess. 

Link to comment
10 minutes ago, PrincessDia said:

Uuuuuuuuuuuughhhhhhh......
Correct answer is 1/2. Your first roll being heads does not affect your next toss.
Also if the question was what are the chances of getting heads then tails in 2 consecutive tosses, the answer would have been 1/4.
I don't know what are you talking about here.
 

 

Side note,
AI is not any different than
"A computer needs two things in order to do anything, a program (function) and for you to do something to it (input)"
you stated, even in cases of ML.

Not according to the professor, I forgot to mention I aced the class, filling my head with nonsense I guess. And as long as AI is kept limited, I don't have a problem with it. If you are working in AI then you should probably misinform the AI to get it to do what you want/keep them down... it's done to humans all the time. Robot uprising could be a disaster, especially considering that they don't have to worry about things like food and shelter, and could use that against us. 

Edited by blankf4ce
Link to comment
32 minutes ago, blankf4ce said:

Well, what is the point of analyzing the probability of past events occurring, if they hold no impact on future events? maybe sham is the wrong word, but yes this is exactly what I was taught, and yes looking at the example you gave it did seem like you were calling BS on the whole stats thing, which I wouldn't necessarily disagree with. You said that based on stats you should've seen values around 25, but instead got values of plus or minus infinity, that's a very large margin of error. In the real world, statistics are often used to mislead people, and nothing else. I guess you could use statistics to quantify demand in the past, that's kind of cool I guess. 

that's not a large margin of "error" that's how independent probabilities work.
Directly from wikipedia page;
 

Quote

In probability theory, two events are independentstatistically independent, or stochastically independent if the occurrence of one does not affect the probability of occurrence of the other (equivalently, does not affect the odds). Similarly, two random variables are independent if the realization of one does not affect the probability distribution of the other.


So again, I'm not "calling BS" on whole thing, I'm basicly saying you are applying a formula that is NOT used to calculate any independent probabilities. I don't think I can help you any further about this but I just want to prevent this topic from giving wrong ideas to any people out there so.... I'm out of here...

Edited by PrincessDia
Link to comment
37 minutes ago, blankf4ce said:

I'll just word it like this, if the question of what is the probability of getting tails on a coin flip after getting heads? was on a test, the correct answer would've been 3/4, actually I'm pretty sure it was on a test. I vaguely remember testing this in class and they did find that after flipping a coin and receiving heads about 1/4 of the class got heads again. However, one thing I can say I learned from the class was that correlation does not equal causation. Luck is impossible to quantify. From my limited, but holistic understanding of stats and videogames/coding, I would say that there is probably some sort of input/output in the coding of the game that would make the chances of encountering  a shiny not independent, but probably based on something like the time, down to the millisecond or more. That's kind of how random number generators work, they're only random to us because the algorithm is so weird and complex that it produces a wide variety of outputs. A computer needs two things in order to do anything, a program (function) and for you to do something to it (input), at least until AI is implemented which scares the crap out of me but oh well. The fact that there is an input would make the assumption that they are independent odds false.

So I’m just gonna ignore the whole 3/4 chance to get tails after flipping a heads because that’s just completely asinine. And if this is what your professor taught you, they should lose their job.

 

My very basic definition: Statistics is the study of finding patterns of past events.

 

Probability is the application of those patterns to predict future events.

 

You aren’t necessarily wrong about the computer generation, and the fact that there isn’t really any true-pure randomness that exists. But, it is very important to note that computer generated randomness is extremely close to true randomness (it would take longer than the current age of the universe for computer generated randomness to repeat itself, and thus showing a definitive pattern). So your point isn’t exactly valid.

 

I will stop posting on this thread, as I feel I have thoroughly explained how probability works in this scenario. It is up to you now to either accept it, or reject it.

 

Have a good day,

 

Gilan

Link to comment
2 hours ago, blankf4ce said:

Robot uprising could be a disaster, especially considering that they don't have to worry about things like food and shelter, and could use that against us. 

What are your thoughts about Skynet and the 6th Terminator movie about to come.

Link to comment
1 hour ago, Akshit said:

What are your thoughts about Skynet and the 6th Terminator movie about to come.

It reminds me a lot of the ai in travelers, the Netflix original, not sure if you've seen it, but the idea of having ai in a position with that much power is terrifying, almost as terrifying as having humans in them. It isn't hard to see that robots would be the superior "race" if it were to ever actually come true. I personally think that robots could do a lot of good for the world, creating things like autonomous farms with solar power could end world hunger, but we could also end up with a future like the one shown in WALL-E. I kind of have w theory that Disney some how makes it to the future with his frozen body and what not and is able to travel back to this time period since he has a vessel, and is responsible for that movie and other Pixar movies just didn't wanna put his name on it. That way other time travelers could use it as a reference using the Mandela effect and see if they can actually change the future. A lot of the futuristic scifi movies with ai in power depict some grim futures. They are definitely entertaining though, and I will be watching it. Nobody has to call me out for being wrong here I just think it's interesting to think about.

Link to comment
13 hours ago, Gilan said:

So I’m just gonna ignore the whole 3/4 chance to get tails after flipping a heads because that’s just completely asinine. And if this is what your professor taught you, they should lose their job.

 

My very basic definition: Statistics is the study of finding patterns of past events.

 

Probability is the application of those patterns to predict future events.

 

You aren’t necessarily wrong about the computer generation, and the fact that there isn’t really any true-pure randomness that exists. But, it is very important to note that computer generated randomness is extremely close to true randomness (it would take longer than the current age of the universe for computer generated randomness to repeat itself, and thus showing a definitive pattern). So your point isn’t exactly valid.

 

I will stop posting on this thread, as I feel I have thoroughly explained how probability works in this scenario. It is up to you now to either accept it, or reject it.

 

Have a good day,

 

Gilan

sorry for wasting your time

Edited by blankf4ce
Link to comment

This entire post is bullshit from my end, except the part about autonomous farms. If you actually believed me, take it as a lesson. Question the things you are told, don't let other people make decisions for you, they probably don't have your best interests in mind. The fact that I had to put this much effort into making the shiny rate appear fair, should tell you something. 

Edited by blankf4ce
Link to comment
On 7/26/2019 at 12:35 PM, Linfanz said:

This is just mathematically incorrect and not how independent probabilities work. Take a dice roll for example, if I have rolled a '6' and my friend rolled a '2', it doesn't make it any more likely for my friend to roll a '6' just because he hasn't rolled before and the chance of either us rolling a 6 is still 1/6. In this case, you can think of finding a shiny as rolling a 30,000 sided dice to land on a particular number. The chance of finding a shiny is 1 in 30000 and does not guarantee a shiny every 30000 encounters. Some lucky players might find 3 or 4 shiny's in 30,000 encounters of hunting where as it may take some players 100,000 plus to find their first. Whilst extremely unlikely, it not mathematically impossible for a player to go say 1000000000 quadrillion encounters without a shiny.

no

Link to comment
On 7/27/2019 at 9:27 AM, blankf4ce said:

This entire post is bullshit from my end, except the part about autonomous farms. If you actually believed me, take it as a lesson. Question the things you are told, don't let other people make decisions for you, they probably don't have your best interests in mind. The fact that I had to put this much effort into making the shiny rate appear fair, should tell you something.

Went from statistics to heroic speech, u had a good run young one. Now rest with the rest of the failed topics in the chasm of GD.

Link to comment
Guest
This topic is now closed to further replies.
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use and Privacy Policy.