Shiny Charm and Masuda Method

[Wichu]

Mod edit (to be removed by Wichu if he so wishes):

The spreadsheet of this thread has been inactive since September 2014. Thread starter Wichu has not been logged on since August 2015. PLEASE STOP PLEADING EGGS UNTIL WICHU POSTS AGAIN.

Sincerely, the mods of OI.


Original OP starts here.

A bit of background:

The Shiny Charm is a key item earned by completing the National Pokédex. In 5th generation games, its effect was to increase the chance of wild Pokémon and bred Pokémon being shiny. In particular, it raised the chance of an egg from two different-language parents being shiny from 6/8192 to 8/8192.

The Shiny Charm returns in X/Y, but its English-language description now specifically mentions 'wild Pokémon'. According to Serebii, the Japanese description says no such thing. I want to determine whether the Shiny Charm still works for Masuda Method breeding in X/Y.

First, we need to take into account that the shiny rate in X/Y is double that of previous games. I don't know whether this applies to Masuda Method breeding too, but from my experience it seems to be a safe assumption. Therefore, I will assume the chance of hatching a Masuda Method shiny in X/Y is 6/4096 (=3/2048).

Now, to test whether this is affected by the Shiny Charm, we just need to hatch loads of eggs. The more eggs, the better! Of course, it's impossible to prove it with 100% certainty, so I'll settle for being 90% sure the Shiny Charm improves the odds.

Ideally, I want to get 50,000 eggs hatched in total. This is a large enough sample that we can be 90% sure both ways (if there are 84 or fewer shinies, then we can be 90% sure that the shiny chance is not increased; on the other hand, if there are 85 or more, we can be 90% sure the shiny chance is increased).

Now, I can't hatch 50,000 eggs by myself, so I'm looking for volunteers to help me out. The only requirements are that you have the Shiny Charm in X/Y and two different-language parents to breed. If you want to volunteer, post here, and give an estimate of how many eggs you are planning to hatch. It is important that you report back once you're done, even if you didn't hatch any shinies. I can't accept any data if you don't post before you hatch the eggs (i.e. I don't care how many eggs it took to hatch shinies you already have), since those results are inherently biased (someone who hasn't hatched a shiny wouldn't post that, skewing the data).

I'm planning on hatching a couple of thousand eggs myself - my initial target is 1,000.

Hatching Pledge Spreadsheet: https://docs.google.com/spreadsheet/ccc?key=0AhIb4pLnCYU4dEFDZVhSQzNEUDJ2VnZZVV9FN255YVE

 

[Mulan15262]

How do you get the shiny charm?

If it is something I can get the shiny charm without too much effort, I'm willing to hatch 73 eggs, and maybe more later. Otherwise, I can't help.

 

[thelandlordsmm]

I'm currently working on attempting to complete the Pokedex to get the shiny charm (almost at 600 caught). Once I get there, I'll let you know, and I can help.

 

[CBManBearPig]

Yeah I would help but I haven't completed my pokedex yet and probably won't be able to without some help because I sold all my old games and won't be able to transfer legendaries. But it I do finish it, I'll help.

 

[goomy pls]

I suggest adding this rule to the first post:

To contribute data to this research, you must first create a post stating the target number of eggs you wish to hatch. Keep track of the number of eggs you hatch until you reach your target. Once you hit your target, post your results. At this point, you may create a new target.

Data will only be accepted if you fully meet your posted target. Partial results will NOT be accepted. For example if you post a target of 100, then quit after 30 eggs, your data will be discarded.

It is recommended to create multiple, smaller targets (60-120) rather than one large target (500-1000). A target must be at least 30, to avoid spamming with too many posts.

This is to avoid confirmation bias. You mention a more lenient version of this rule in the first post

f you want to volunteer, post here, and give an estimate of how many eggs you are planning to hatch. It is important that you report back once you're done, even if you didn't hatch any shinies.

but it is absolutely crucial that this is strictly enforced to collect accurate, unbiased data.

 

[Wichu]

I suggest adding this rule to the first post:

To contribute data to this research, you must first create a post stating the target number of eggs you wish to hatch. Keep track of the number of eggs you hatch until you reach your target. Once you hit your target, post your results. At this point, you may create a new target.

Data will only be accepted if you fully meet your posted target. Partial results will NOT be accepted. For example if you post a target of 100, then quit after 30 eggs, your data will be discarded.

It is recommended to create multiple, smaller targets (60-120) rather than one large target (500-1000). A target must be at least 30, to avoid spamming with too many posts.

This is to avoid confirmation bias. You mention a more lenient version of this rule in the first post but it is absolutely crucial that this is strictly enforced to collect accurate, unbiased data.

Intuitively, stopping short of your target shouldn't affect the results (as long as you say where you stopped), since the probability of any one egg being shiny is independent of all others. It shouldn't matter if you stop when you hatch a shiny, as long as the total number of eggs hatched is large enough. Am I wrong?

 

[GameMaster0000]

I can try this.

Shiny Charm Obtained
Has foreign Ditto
Bank can help me on process that can send all fail shiny to bank for count.

 

[The Scientist]

Intuitively, stopping short of your target shouldn't affect the results (as long as you say where you stopped), since the probability of any one egg being shiny is independent of all others. It shouldn't matter if you stop when you hatch a shiny, as long as the total number of eggs hatched is large enough. Am I wrong?

Yes, this logic is wrong. Let's say the true shiny rate is 1/1000. A player, who let's say does not know the true shiny rate, decides to hatch 1000 eggs. More often than not, he will get a shiny out of 1000 eggs. Sometimes it happens after 50 eggs, or 500 eggs, or 950 eggs. If he hatches a shiny after 50, 500, or 950 eggs, and then hatches the rest without any shiny, the rate would be 1/1000. Sometimes he won't hatch any shinies in 1000 eggs, but he should get a shiny more often than not. So, if he stops after hatching the shiny, the rates would be 1/50, 1/500, and 1/950. If the player stops after hatching a shiny, the perceived rate would be much lower than the true rate. This is what Goomy pls meant when he spoke of biased vs. unbiased data. Stopping after hatching a shiny would create a biased dataset, and continuing to 1000 eggs (or however many you originally decide) would create an unbiased dataset.

If you assume the shiny charm + Masuda method works as it did in generation V, the rate would be 8/4096 in generation VI, since it was 8/8192 in generation V. So that's 1/512 this generation, but we won't know that for sure until we do the research or get a hacker to confirm it.

Shiny charm with my USA Mudkip and ITA Ditto.

500 eggs- no shiny!

 

[THE_411]

I just got the shiny charm.

I'll start hatching using Masuda and report every 100 eggs.

 

[Wichu]

Yes, this logic is wrong. Let's say the true shiny rate is 1/1000. A player, who let's say does not know the true shiny rate, decides to hatch 1000 eggs. More often than not, he will get a shiny out of 1000 eggs. Sometimes it happens after 50 eggs, or 500 eggs, or 950 eggs. If he hatches a shiny after 50, 500, or 950 eggs, and then hatches the rest without any shiny, the rate would be 1/1000. Sometimes he won't hatch any shinies in 1000 eggs, but he should get a shiny more often than not. So, if he stops after hatching the shiny, the rates would be 1/50, 1/500, and 1/950. If the player stops after hatching a shiny, the perceived rate would be much lower than the true rate. This is what Goomy pls meant when he spoke of biased vs. unbiased data. Stopping after hatching a shiny would create a biased dataset, and continuing to 1000 eggs (or however many you originally decide) would create an unbiased dataset.

OK, I think I understand. Argh, I can't statistics :/

EDIT: Hang on, what's the difference between doing that and repeatedly setting a target of 1 egg?

 

[The Scientist]

OK, I think I understand. Argh, I can't statistics :/

EDIT: Hang on, what's the difference between doing that and repeatedly setting a target of 1 egg?

I'm not sure I understand the question. As long as you don't stop after obtaining a shiny, your results should not be biased. There would be no difference between doing 1000 trials of one egg and 1 trial of 1000 eggs.

 

[goomy pls]

Intuitively, stopping short of your target shouldn't affect the results (as long as you say where you stopped), since the probability of any one egg being shiny is independent of all others. It shouldn't matter if you stop when you hatch a shiny, as long as the total number of eggs hatched is large enough. Am I wrong?

It all comes down to this: people are more likely to quit after finding a shiny than not finding a shiny. For most people, their mindset is not "I want as many shinies as possible so I'll keep hatching eggs," but rather something like "I want one shiny female Eevee." It's just human nature to have specific targets rather than doing the same task over and over without knowing what you want from it. They would see the first one that meets their target as more valuable that any subsequent ones, because the target has already been met and there's no incentive to keep going.

The purpose of this rule is to encourage people to keep contributing even after they've met their target, because otherwise they'd quit early and skew the data. At the same time, you cannot ask people to set their targets too high, because that would encourage early quitting.

OK, I think I understand. Argh, I can't statistics :/

EDIT: Hang on, what's the difference between doing that and repeatedly setting a target of 1 egg?

There is no difference. It's important to realize that repeatedly setting a target of 1 egg WILL create more bias in the data. Ideally you'd set the targets very high, but that would conflict with the limits of human patience.

example data set, take it with a grain of salt:

Hide

Person A: Shiny hatched after 40 eggs
Person B: Shiny hatched after 301 eggs
Person C: Shiny hatched after 1010 eggs, second Shiny at 1050 eggs

Assuming everyone quits immediately after one shiny and never returns,
Person A would report a shiny rate of 1/40
Person B would report a shiny rate of 1/301
Person C would report a shiny rate of 1/1010

If each person pledges 100-egg targets,
Person A would report a shiny rate of 1/100
Person B would report a shiny rate of 1/400
Person C would report a shiny rate of 1/525

Again, the data is fabricated so don't jump to any conclusions. I'm just trying to show that having fixed targets higher than 1 will help to smooth out the outliers.

To have truly unbiased data, everyone would have to continue contributing for an infinite amount of time.



Anyway, I'll probably help out once pokebank comes out and I get the shiny charm.


Edit: one other thing. So far, there has been no confirmation on whether it's the LANGUAGE or the REGION of the parents that matter for 6th gen MM. I think it would be wise to have everyone specify whether their parents are Different Language and Region, Different Language and same Region, or Different Language from Pokebank.

Note that there actually is a way to conclusively prove whether it's Language or Region that matters, but that requires the RNG Breeding method.

After hatching a Masuda shiny using the rng method, reset and swap the parents for two pokemon of different languages but same region. If the pokemon is no longer shiny, the Masuda method is based on region. If the pokemon remains shiny, reset and swap the parents for two local pokemon. If the pokemon is no longer shiny, then the Masuda method is based on language.

 

[13mjvr]

I'm currently hatching for a shiny Chimchar with HA. However, I'm using a SPA ditto plus my ENG chimchar. I dont know for sure if that combination works for MM. I have the shiny charm. I can start counting eggs if you say it's safe to assume that the SPA & ENG combination works. If not, I'm going to hatch for a female chimchar with more than 4 IVs and breed it with my japanese Mawile and start counting.

 

[Pikachuun]

I'm currently hatching for a shiny Chimchar with HA. However, I'm using a SPA ditto plus my ENG chimchar. I dont know for sure if that combination works for MM.

It should, because of the different locales of Ditto and Chimchar.

 

[Delyth-draws]

I'd happily take part. Just got the shiny charm this morning and plan to breed a japanese tynamo with a english tynamo. Going aim for 100 eggs.

 

[THE_411]

Posting some results.

Breeding a Japanese mon with an English (Australian) mon.

First 100 hatched - no shiny
Second 100 hatched - no shiny

 

[Wichu]

It all comes down to this: people are more likely to quit after finding a shiny than not finding a shiny. For most people, their mindset is not "I want as many shinies as possible so I'll keep hatching eggs," but rather something like "I want one shiny female Eevee." It's just human nature to have specific targets rather than doing the same task over and over without knowing what you want from it. They would see the first one that meets their target as more valuable that any subsequent ones, because the target has already been met and there's no incentive to keep going.

The purpose of this rule is to encourage people to keep contributing even after they've met their target, because otherwise they'd quit early and skew the data. At the same time, you cannot ask people to set their targets too high, because that would encourage early quitting.


There is no difference. It's important to realize that repeatedly setting a target of 1 egg WILL create more bias in the data. Ideally you'd set the targets very high, but that would conflict with the limits of human patience.

example data set, take it with a grain of salt:

Hide

Person A: Shiny hatched after 40 eggs
Person B: Shiny hatched after 301 eggs
Person C: Shiny hatched after 1010 eggs, second Shiny at 1050 eggs

Assuming everyone quits immediately after one shiny and never returns,
Person A would report a shiny rate of 1/40
Person B would report a shiny rate of 1/301
Person C would report a shiny rate of 1/1010

If each person pledges 100-egg targets,
Person A would report a shiny rate of 1/100
Person B would report a shiny rate of 1/400
Person C would report a shiny rate of 1/525

Again, the data is fabricated so don't jump to any conclusions. I'm just trying to show that having fixed targets higher than 1 will help to smooth out the outliers.

To have truly unbiased data, everyone would have to continue contributing for an infinite amount of time.

I thought that might be important at first, but I'm pretty sure it doesn't matter in this case.

As each egg has an equal and independent chance of being shiny, the number of total eggs hatched is the only variable that affects the number of shinies. It doesn't matter who hatches the egg. Even if people stop when they get a shiny, as long as we get 50,000 eggs in total, the result should be valid (the difficult part is cutting it off at exactly 50,000, since that's where the bias comes in).

Imagine writing a program to simulate 500 people each with a target of 100 eggs. You'd do 500 runs of 100 independent random numbers, which is identical to one run of 50,000 random numbers. Now, instead of this, simulate someone stopping when they get a shiny and being replaced by someone else, and repeat until 50,000 eggs. This simplifies to exactly the same thing.

 

[The Scientist]

As each egg has an equal and independent chance of being shiny, the number of total eggs hatched is the only variable that affects the number of shinies. It doesn't matter who hatches the egg. Even if people stop when they get a shiny, as long as we get 50,000 eggs in total, the result should be valid (the difficult part is cutting it off at exactly 50,000, since that's where the bias comes in).

Here's the thing- either you need have people pledge how many eggs they are going to do, or you need to tell people to keep hatching until they get a shiny, no matter what. But it needs to be consistent!

Take a look at this post on Reddit- http://a0.reddit.com/r/pokemon/comments/1ol2tw/research_increased_shiny_chance_in_friend_safari/

Be sure to view the Google Docs. As you can see, many people did approx. 100 random encounters in the Friend Safari. They finished their 100 encounters whether or not they found a shiny. Imagine if the people that found a shiny stopped after they found the shiny? This would leave random people having stopped at 30 encounters, 60 encounters, etc. The hypothesized rate would be much lower than what it actually is.

On the other hand, if everyone did random encounters until they found a shiny, this would provide similar data, and it would be OK since it was consistent. The key is consistency!

Furthermore, regardless of whether or not you believe there's bias, it's still beneficial for people to pledge how many eggs they are going to hatch. That way people are hatching eggs specifically for this thread. It would prevent lots of random people coming in and saying things like "I hatched a shiny after 70 eggs."

 

[thelandlordsmm]

Finally got the shiny charm. I'm going to start by hatching 100 eggs. I'll report the results when finished.

 

[Thirdbird]

Edit: one other thing. So far, there has been no confirmation on whether it's the LANGUAGE or the REGION of the parents that matter for 6th gen MM. I think it would be wise to have everyone specify whether their parents are Different Language and Region, Different Language and same Region, or Different Language from Pokebank.

Note that there actually is a way to conclusively prove whether it's Language or Region that matters, but that requires the RNG Breeding method.

I was under the impression that it was the region, not the language, that mattered. Take for example Dratini. Its English name is the same as its German, Italian, and Spanish names.

 

[iZac]

(i am supposed to post here not make my own thread? )
So I would love to help got my Shiny charm today.
my first goal will be for like 80 eggs / until I get a shiny Beldum

 

[iZac]

Breeding. 1:

Ditto: SPA
lvl: 56
Nature: Limber

Luxray: USA
lvl:33
Nature: Lonely

Goal to get 2 boxes of eggs and then start hatching.

 

[Pajerski]

I hate to be ants at a picnic here, but it clearly states wild pokemon. Usually descriptions such as these need to be taken literally, Masuda odds have already been cut in half, I don't see them lowering it any further than they already have.

They usually only change descriptions like this if it's necessary to do so.

 

[Tsal]

Heard rumours saying shiny charm does not stack with madusa for x/y

 

[The Scientist]

I hate to be ants at a picnic here, but it clearly states wild pokemon. Usually descriptions such as these need to be taken literally, Masuda odds have already been cut in half, I don't see them lowering it any further than they already have.

They usually only change descriptions like this if it's necessary to do so.

The Japanese translation is the same in XY as it was in BW2. We don't know why the English translators changed the description and the Japanese translators did not. Jumping to conclusions will not get us anywhere, but perhaps testing the shiny rate will.