Kiliminati Posted June 17, 2015 Share Posted June 17, 2015 Ok, could you go over this please? I plugged in the rumored rate and got these results using Vassar Stats. How is it that you are more likely to have exactly 1 Shiny Pokemon with the same timeframe and a lower percentage? Xela you're killing me ;_; For a rate of 1/8192: using binompdf (40,000, 1/8192, 1) we get a 3.6984% chance of finding exactly one shiny. For a rate of 1/32000: using binompdf (40,000, 1/32000, 1) we get a 35.813% chance of finding exactly one shiny. Now, I assume you're asking me why the probability is lower for the first rate than the second; it is literally because the first rate is greater than the second that this is the case. Due to the increased odds in the first scenario, it is much more likely to find something OTHER THAN exactly one shiny. Basically, it's implying that there's a good chance of you getting either two or more, not only one. The second scenario above also makes sense, as since the suggested rate is lower than the first, the odds of you finding only one shiny out of 40,000 is more likely. Link to comment
Quakkz Posted June 17, 2015 Share Posted June 17, 2015 nerds Bishav, screamfields, YagamiNoir and 16 others 19 Link to comment
doggydom Posted June 17, 2015 Share Posted June 17, 2015 So to get a shiny people have to encounter over 40 thousand pokemon or more. Thats crazy! Link to comment
XelaKebert Posted June 17, 2015 Share Posted June 17, 2015 Xela you're killing me ;_; For a rate of 1/8192: using binompdf (40,000, 1/8192, 1) we get a 3.6984% chance of finding exactly one shiny. For a rate of 1/32000: using binompdf (40,000, 1/32000, 1) we get a 35.813% chance of finding exactly one shiny. Now, I assume you're asking me why the probability is lower for the first rate than the second; it is literally because the first rate is greater than the second that this is the case. Due to the increased odds in the first scenario, it is much more likely to find something OTHER THAN exactly one shiny. Basically, it's implying that there's a good chance of you getting either two or more, not only one. The second scenario above also makes sense, as since the suggested rate is lower than the first, the odds of you finding only one shiny out of 40,000 is more likely. Ok, thank you for clearing this up. Kiliminati 1 Link to comment
Kiliminati Posted June 17, 2015 Share Posted June 17, 2015 Ok, thank you for clearing this up. No problem man :) Link to comment
XelaKebert Posted June 17, 2015 Share Posted June 17, 2015 No problem man :) I'm working out some statistics on my end using the stated 1/8192 odds across a base of 20,000 players. Kiliminati 1 Link to comment
Gilan Posted June 17, 2015 Share Posted June 17, 2015 actually he said he was told by a another member of staff who was higher in the totem pole. i dont recall exactly who but i still have a copy of the document. eh, that's not what I was told... But it was a long time ago and my source (whoever it was) probably wasn't reliable. I would like to see the doc. though Link to comment
KingBowser Posted June 17, 2015 Share Posted June 17, 2015 No way shiny counter can be right, spend so much time breeding and ev training. All i got to show for it is a shiny tangela, system is ridic. necund, Bearminator, frogbussoms and 3 others 6 Link to comment
Rigamorty Posted June 17, 2015 Share Posted June 17, 2015 So to get a shiny people have to encounter over 40 thousand pokemon or more. Thats crazy!no they don't. You could encounter 5 Pokemon and get a shiny. It's just all probability and luck Link to comment
XelaKebert Posted June 17, 2015 Share Posted June 17, 2015 Ok, so I wrote some code to simulate a study of 200,000 players reaching 40,000 encounters. The rate is the stated 1/8192 per encounter. Some things to note here. The number of players doesn't reflect the actual size of the playerbase here, I have no idea what the actual size is. The 40,000 encounters has been established as the standard for those testing these odds. This data would show the relative density of Shiny Pokemon across this sample size at the 1/8192 encounter rate. Note: You should take this with a grain of salt because even though the math lines up with probability there are some variables that do not get taken into account. This math assumes that all players in the sample have reached 40,000 encounters and none have quit by then. If I could effectively simulate a player quitting I would add that in the get more realistic looking data. SneaKyKhaLidA, noobihol and SirAlbert 3 Link to comment
Arimanius Posted June 17, 2015 Share Posted June 17, 2015 Well I feel betrayed now D: That's a lot of people with OT shinnies, something doesn't happen here that much Link to comment
XelaKebert Posted June 17, 2015 Share Posted June 17, 2015 (edited) Well I feel betrayed now D: That's a lot of people with OT shinnies, something doesn't happen here that much You're right, because that sample assumes all players had reached 40,000 encounters. I'm working on a way to simulate a player quitting at every 1000 encounters. The only problem is getting the numbers to actually work how they should. At this point I have it showing all 200,000 players quitting early, which just isn't possible. Will post back when I have the results. Edit: After lowering the numbers to debug the program is running exactly how it should. Here is the code: What I have done is that since I had it where all 200,000 players quit before the 40,000 encounter threshold. I had thought I was looking at a statistical improbability, but in reality it actually makes sense. Since we're looking at a really large sample size, the odds that all of them will quit early are quite large. Now for a more detailed explanation. k is the number of players = 200,000 i is the number of encounters per player = 40,000 j is the number of tests to run = 10 For each player the program will run through a loop of 40,000 encounters simulating the stated 1/8192 rate. At each increment of 1,000 encounters the system will do a random roll of 1 or 0 to determine if the current player will quit early. If the system rolls a 1 then it will break out of the encounter loop and move on to the next player and increment the number of players who quit early. It checks for every 1,000 encounters by checking the modulus, remainder, of dividing the number by 1000. If that number is 0 then 1,000 encounters have passed. At the end of each test the results are printed to the screen. These are the results: What I thought was a statistical improbability actually is statistically possible considering an extremely large sample size and a 50/50 chance of a player quitting at every 1,000 encounters. Is this 100% perfect? No, it still doesn't take everything into account but it is as realistic as can be simulated. Edited June 17, 2015 by XelaKebert noobihol and SirAlbert 2 Link to comment
Arimanius Posted June 17, 2015 Share Posted June 17, 2015 stuff Yeah you're right there are a lot of people who quit early or maybe alt runs can go on that too since in an alt run you can't reach to 40,000 encounters (not likely at least) but still that's a large number, 20k+ people with ot shinnies is a lot and I think it is not close to what happen here even though it is not 100% realistic this should be a close example of what happens in reality so it shouldn't exist such a high difference Link to comment
XelaKebert Posted June 17, 2015 Share Posted June 17, 2015 Yeah you're right there are a lot of people who quit early or maybe alt runs can go on that too since in an alt run you can't reach to 40,000 encounters (not likely at least) but still that's a large number, 20k+ people with ot shinnies is a lot and I think it is not close to what happen here even though it is not 100% realistic this should be a close example of what happens in reality so it shouldn't exist such a high difference Yes it is a lot, but you're thinking about everyone reaching the critical threshold where they are all but guaranteed to have a shiny. The thing is that very few people ever do, and the number of players with a shiny being small reflects those who obtained one either before or after the critical threshold, which I'd estimate is somewhere around 5,000 encounters. I would have to do some math to show where that would be at though. I find it interesting that you bring up alts as an exception to the sample size as well. Alts are not an exception to the sample size and would be counted as a separate player for the sake of the test since they have different IDs from the primary character of the person who created it. Those numbers are a lot more realistic than you think because the term "quitting early" is a blanket term used to cover players that start using repels, have quit the game, or only play in comp with no outside grinding. When a players "quits" it means that they've just stopped encountering wild Pokemon. To be 100% truthful, yes it is statistically possible that one player will reach 40,000 encounters in that sample size, but you're talking that the likelihood of that being less than 1% over the test period. To say that such a high disparity shouldn't exist is very closed minded because it does. This could also be used across the same sample size in handheld players in Gen 3. There are people who have played the handhelds for years and never once encountered a wild shiny outside of the Lake of Rage Gyarados (Which counts as a wild encounter, but for the sake of testing no one really counts since you're pretty much handed that one). Link to comment
RyukoNotsunaga Posted June 17, 2015 Share Posted June 17, 2015 (edited) Each time you encounter a Pokémon, that Pokémon has a 1/8192 chance of being shiny, and is chosen whether or not it is shiny by rng. It isn't a rate that adds up every time you encounter another Pokémon. So if you meet the 8,192nd Pokémon, it isn't guaranteed that it's going to be shiny. And the same follows through for the donator status of 1/7,373. Someone starting out can get a shiny starter, while those who've played for 3,000+ hours still haven't found one. It's all about luck. :3. Edited June 17, 2015 by RyukoNotsunaga WolfgangDamien 1 Link to comment
flavajabari Posted June 17, 2015 Share Posted June 17, 2015 Each time you encounter a Pokémon, that Pokémon has a 1/8192 chance of being shiny, and is chosen whether or not it is shiny by rng. It isn't a rate that adds up every time you encounter another Pokémon. So if you meet the 8,192nd Pokémon, it isn't guaranteed that it's going to be shiny. And the same follows through for the donator status of 1/7,373. Someone starting out can get a shiny starter, while those who've played for 3,000+ hours still haven't found one. It's all about luck. :3. thanks for pointing out what everyone alrdy said once more. ThinkNice, Gunthug, TheGloriousWalrus and 5 others 8 Link to comment
PandaJJ Posted June 17, 2015 Share Posted June 17, 2015 Each time you encounter a Pokémon, that Pokémon has a 1/8192 chance of being shiny, and is chosen whether or not it is shiny by rng. It isn't a rate that adds up every time you encounter another Pokémon. So if you meet the 8,192nd Pokémon, it isn't guaranteed that it's going to be shiny. And the same follows through for the donator status of 1/7,373. Someone starting out can get a shiny starter, while those who've played for 3,000+ hours still haven't found one. It's all about luck. :3. This is something that somebody who never heard about statistics would say. Link to comment
Gunthug Posted June 17, 2015 Share Posted June 17, 2015 'twas a very staffy comment PandaJJ 1 Link to comment
Inarin Posted June 17, 2015 Share Posted June 17, 2015 (edited) Anyways, on topic: Keep hunting a shiny. It may happen by the time you quit. Still looking for my OT. >.> Edited June 17, 2015 by Noad Link to comment
RyukoNotsunaga Posted June 17, 2015 Share Posted June 17, 2015 (edited) This is something that somebody who never heard about statistics would say. Statistics are used to say what is considered as the "norm" in something, using data collected from the subject that's being calculated. However, statistics aren't too viable in something that happens randomly. Sure, they may give a good insight on what to expect when you reach a certain point. But nothing is ever guaranteed. That's why it's called rng. Edited June 17, 2015 by Noad Link to comment
ThinkNice Posted June 17, 2015 Share Posted June 17, 2015 Statistics are used to say what is considered as the "norm" in something, using data collected from the subject that's being calculated. However, statistics aren't too viable in something that happens randomly. Sure, they may give a good insight on what to expect when you reach a certain point. But nothing is ever guaranteed. That's why it's called rng. omfg TheNakedRamen, jayfeatskydd, Kiliminati and 5 others 8 Link to comment
PandaJJ Posted June 17, 2015 Share Posted June 17, 2015 Statistics are used to say what is considered as the "norm" in something, using data collected from the subject that's being calculated. However, statistics aren't too viable in something that happens randomly. Sure, they may give a good insight on what to expect when you reach a certain point. But nothing is ever guaranteed. That's why it's called rng. Again, this is something that somebody who never heard about statistics would say. But maybe I should have said mathematical statistics to be precise. Statistics is the study of random behaviour. "Statistics aren't too viable in something that happens randomly" must be one of the most hilarious thing I've read in a while. Furthermore, this game is not even random, it is pseudo-random. Which means that if you repeat the same action with periodically, you are bound to experience every possible outcome. Kiliminati 1 Link to comment
noobihol Posted June 17, 2015 Author Share Posted June 17, 2015 (edited) Still waiting for a confirmation. Edit: Had 2500-3000 encounters more today with an active donator status. Edited June 17, 2015 by noobihol Link to comment
JoshLindsay10 Posted June 17, 2015 Share Posted June 17, 2015 (edited) Statistics are used to say what is considered as the "norm" in something, using data collected from the subject that's being calculated. However, statistics aren't too viable in something that happens randomly. Sure, they may give a good insight on what to expect when you reach a certain point. But nothing is ever guaranteed. That's why it's called rng. I'm actually off-duty from staff at the moment, due to having no access to a stable internet connection that I can access regularly. :3. Give this guy a promotion when he gets good internet Still waiting for a confirmation. Edit: Had 2500-3000 encounters more today with an active donator status. Clearly not going to happen dude, they've been here because posts have been removed etc they just dont want to comment Edited June 17, 2015 by JoshLindsay10 noobihol 1 Link to comment
RyukoNotsunaga Posted June 17, 2015 Share Posted June 17, 2015 Again, this is something that somebody who never heard about statistics would say. But maybe I should have said mathematical statistics to be precise. Statistics is the study of random behaviour. "Statistics aren't too viable in something that happens randomly" must be one of the most hilarious thing I've read in a while. Furthermore, this game is not even random, it is pseudo-random. Which means that if you repeat the same action with periodically, you are bound to experience every possible outcome. Woops. Sorry. It seems that my understanding of statistics was incorrect. I'll definitely have to research further into statistics, as well as mathematical statistics to get a better understanding on them. I did look a little into the definition of statistics on "Google," and now see that the "Statistics aren't too viable in something that happens randomly" is, in fact, contradictory to itself. XD. PandaJJ 1 Link to comment
Recommended Posts