Minecraft Forums

I swear this is freaking me out i just spawned and i already saw like 10 pink sheep. Does it have to do with me playing with optifine? I created another world but that one didnt have any pink sheep. The seed is -5736348399150288665 I made a copy of the world and theyre still there!!! AM I THE LUCKIEST PERSON IN THE WORLD?!

There was a bug with worldgen where the game would forget what color the sheep was or something. You would end up with sheep all the same color, and that could be whatever color (in your case, pink). Not sure if/when this was fixed, however, but it sounds like the same thing.

Also, while rarer than white, grey, black, and brown, pink sheep aren't really all that rare.

Quote from DuhDerp»

There was a bug with worldgen where the game would forget what color the sheep was or something. You would end up with sheep all the same color, and that could be whatever color (in your case, pink). Not sure if/when this was fixed, however, but it sounds like the same thing.

Also, while rarer than white, grey, black, and brown, pink sheep aren't really all that rare.

Pink sheep have a 0.164% chance of spawning. That's like 1 out of ~610 sheep. Pretty rare to me.

https://bugs.mojang.com/browse/MC-2788

It is very unlikely to find more than 1-2 pink sheep in an area due to random chance, much less likely than indicated by the number of threads about finding many pink sheep, especially not 10 or more:

Chance of at least 1 pink sheep in a group of 100 sheep is 1 in 6.612873
Chance of at least 2 pink sheep in a group of 100 sheep is 1 in 83.59876
Chance of at least 3 pink sheep in a group of 100 sheep is 1 in 1583.7821
Chance of at least 4 pink sheep in a group of 100 sheep is 1 in 41493.777
Chance of at least 5 pink sheep in a group of 100 sheep is 1 in 909090.94

That was the result of counting the number of pink sheep in a group of 100 sheep (the number in 25 packs of 4 each and 250 chunks with a 10% chance of a pack per chunk) over 10 million runs, and in practice they will be rarer (per chunk) since not every pack is sheep, nor does every pack attempt succeed and some biomes do not spawn passive mobs or sheep (obviously, any world will have many more pink sheep overall but this is within a relatively small area, 250 chunks is about the size of a single biome).

The vanilla bug is due to improper usage of the world RNG (which is supposed to only be used for random events that happen in the world) for both structure generation, which sets the RNG to the same state within 32x32 chunk regions*, as well as sheep using the same RNG instead of their own (ever noticed that horses and other mobs with variants or village trades (as much as they can vary since 1.8) are never the same when you recreate a world? That's because each entity has their own RNG which is seeded according to the time the entity is created) - so every group of sheep within such large regions will have the same colors (possibly with some variation if not every individual spawn attempt succeeds; for example, attempts 1,2,3,4 all succeed for one pack while 1,2,3,5 succeed for another, with 5 having a different color due to 4 altering the RNG sequence).

*This is due to the way the structure generator (at least villages and temples) set the seed:

this.field_82665_g = 32;
int var5 = par1 / this.field_82665_g;
int var6 = par2 / this.field_82665_g;
Random var7 = this.worldObj.setRandomSeed(var5, var6, 10387312);

This is VERY easy to fix - it is just silly that Mojang has still not fixed it as of 1.12 when all you need to do is use the RNG instance in MapGenStructure instead of World.setRandomSeed(). Voila, no more mass spawning of pink sheep, and they are in ways even easier to find (if only one at a time, but you can breed them with another color to get more) since they vary per-chunk instead of per-region (1 region = 1024 chunks). They can even still be tied to the world seed (unique out of all mobs with variants) by calling setRandomSeed() just before spawning animals (or better yet, a RNG specific to the Entity(Sheep) class but I didn't want to modify it just for that; using a method specific to entities could be used to make all such mobs tied to the seed):

int regionX = chunkX / this.maxDistance;
int regionZ = chunkZ / this.maxDistance;
// Uses method and RNG specific to MapGenStructure, not World
this.setRandomSeed(regionX, regionZ, 10387312);

// Used to set the world RNG so sheep wool colors are based off the world seed
this.worldObj.setRandomSeed(chunkX, chunkZ);
SpawnerAnimals.performWorldGenSpawning(this.worldObj, biome, blockX, blockZ, 16, 16, this.rand);

In my most recent world with these modifications I found two pink sheep far apart from each other, and have seen quite a few in other worlds - always single - while in a vanilla world which is as large as all of my other worlds combined I've only found them once - with multiple pink sheep nearby, about as many as I've found in all other worlds combined, which makes sense due to how the bug works.

There may still be some correlation between nearby chunks due to the fact that their individual seeds are based on the world seed modified by their coordinates but it doesn't appear to be significant (e.g. 1234567890 and 1234567891 produce very different worlds despite a tiny change in the seed; the per-chunk seed changes much more since it is multiplying the coordinates by two values derived from the world seed, which are usually very large; 1770586251933415359 and -788814196833671301 for the seed -123775873255737467).

This happened to me when I was playing 1.10 or 1.11, almost half of the sheep were pink, sadly I lost the world file though.