The Meaning of Life, the Universe, and Everything.
Join Date:
4/27/2012
Posts:
48
Location:
Forests, snowy forests
Member Details
hmmm , are you really checking what biome is at precisely 0,0 ?
Personally i would be looking at at least 300x300 blocks around the 0,0. Sure there could be multiple biomes inside that zone, but i would count only the one thats the largest.
Interesting data, and great seeds you found
Rollback Post to RevisionRollBack
I don't like giving too much info about the seeds i share. I feel like its better to let others explore and find out the chest contents,dungeon places and biomes. If someone is giving me a gift i would rather not know what it is until i open it for my self.
There is another potential source of error which may result in the amount of oceans being understated - the game is coded to force a continent to generate around the origin, a legacy from the pre-1.7 days which is still in the code as of the latest versions (1.12 is the newest version for which I could find code online but 1.13 did not make any changes to the land/ocean map):
This code is exactly the same as in 1.6.4 and it adds initial landmasses at a rate of 1 in 10 "areas", with the part outside the loops at the bottom placing a landmass at the origin (this does not guarantee coverage there since they randomly "grow" in subsequent stages; 1.7+ added code that fills in much of the oceans between the 1.6 landmasses):
public int[] getInts(int areaX, int areaY, int areaWidth, int areaHeight)
{
int[] aint = IntCache.getIntCache(areaWidth * areaHeight);
for (int i = 0; i < areaHeight; ++i)
{
for (int j = 0; j < areaWidth; ++j)
{
this.initChunkSeed((long)(areaX + j), (long)(areaY + i));
aint[j + i * areaWidth] = this.nextInt(10) == 0 ? 1 : 0;
}
}
if (areaX > -areaWidth && areaX <= 0 && areaY > -areaHeight && areaY <= 0)
{
aint[-areaX + -areaY * areaWidth] = 1;
}
return aint;
}
For example, this is a comparison of the biome frequency in 1.6.4 for 1000 random seeds within 1024 blocks of the origin and another 1000 seeds centered at least 16384 blocks away (each seed with a different random location) to eliminate bias from the "origin continent" (this was done with my own code which uses a copy of the game's biome generator to generate a biome map much like AMIDST, then it reads the biomes at every point). The amount of ocean is nearly three times higher in the latter case, it also appears that Ice Plains ("cold" climate zones, the only such areas prior to 1.7) are less likely closer to the origin (the way 1.7 works is that it favors the edges of landmasses for cold/hot with the interiors more likely to be cool/warm, so this can also introduce a bias since the area around 0,0 is more likely to be the middle of a landmass):
(I have no idea why but I simply could not get this to be formatted properly as text by using code tags; spaces would always be removed - it is just insane how broken the editor still is after years and years, no wonder these forums are nearly dead)
Granted, the difference would not be as extreme in newer versions due to a subsequent generation stage which fills in oceans with land but it should still be considered if you want to be as accurate as possible.
One day I was trying to find out precisely the rarity of Minecraft biomes, I thought of different ways to do it and finally decided to use one with Amidst and Python in search of more accurate and updated results, it is still in a very early phase but there are some results interesting for themselves and others in relation to your results. This is my study Minecraft Java 1.16.1 Biomes extension using Amidst and Python.
hmmm , are you really checking what biome is at precisely 0,0 ?
Personally i would be looking at at least 300x300 blocks around the 0,0. Sure there could be multiple biomes inside that zone, but i would count only the one thats the largest.
Interesting data, and great seeds you found
I don't like giving too much info about the seeds i share. I feel like its better to let others explore and find out the chest contents,dungeon places and biomes. If someone is giving me a gift i would rather not know what it is until i open it for my self.
Maybe this is of some help...
1 in 74 overworld biomes: 1.351% (proportion)
confidence level: 99.999% (there's a 0.001% chance that a certain biome rarity is not the measured value +/- the margin of error)
margin of error: how much the measured value can deviate from the real world gen value.
Needed sample size:
I'm not 100% sure, but since you list to one hundreth of a percent precision, I think you need 26 million samples for 99.999% confidence.
There is another potential source of error which may result in the amount of oceans being understated - the game is coded to force a continent to generate around the origin, a legacy from the pre-1.7 days which is still in the code as of the latest versions (1.12 is the newest version for which I could find code online but 1.13 did not make any changes to the land/ocean map):
public int[] getInts(int areaX, int areaY, int areaWidth, int areaHeight) { int[] aint = IntCache.getIntCache(areaWidth * areaHeight); for (int i = 0; i < areaHeight; ++i) { for (int j = 0; j < areaWidth; ++j) { this.initChunkSeed((long)(areaX + j), (long)(areaY + i)); aint[j + i * areaWidth] = this.nextInt(10) == 0 ? 1 : 0; } } if (areaX > -areaWidth && areaX <= 0 && areaY > -areaHeight && areaY <= 0) { aint[-areaX + -areaY * areaWidth] = 1; } return aint; }https://github.com/ObelouixServer/Minecraft-1.12/blob/master/src/minecraft/net/minecraft/world/gen/layer/GenLayerIsland.java
For example, this is a comparison of the biome frequency in 1.6.4 for 1000 random seeds within 1024 blocks of the origin and another 1000 seeds centered at least 16384 blocks away (each seed with a different random location) to eliminate bias from the "origin continent" (this was done with my own code which uses a copy of the game's biome generator to generate a biome map much like AMIDST, then it reads the biomes at every point). The amount of ocean is nearly three times higher in the latter case, it also appears that Ice Plains ("cold" climate zones, the only such areas prior to 1.7) are less likely closer to the origin (the way 1.7 works is that it favors the edges of landmasses for cold/hot with the interiors more likely to be cool/warm, so this can also introduce a bias since the area around 0,0 is more likely to be the middle of a landmass):
(I have no idea why but I simply could not get this to be formatted properly as text by using code tags; spaces would always be removed - it is just insane how broken the editor still is after years and years, no wonder these forums are nearly dead)
Granted, the difference would not be as extreme in newer versions due to a subsequent generation stage which fills in oceans with land but it should still be considered if you want to be as accurate as possible.
TheMasterCaver's First World - possibly the most caved-out world in Minecraft history - includes world download.
TheMasterCaver's World - my own version of Minecraft largely based on my views of how the game should have evolved since 1.6.4.
Why do I still play in 1.6.4?
.
One day I was trying to find out precisely the rarity of Minecraft biomes, I thought of different ways to do it and finally decided to use one with Amidst and Python in search of more accurate and updated results, it is still in a very early phase but there are some results interesting for themselves and others in relation to your results. This is my study Minecraft Java 1.16.1 Biomes extension using Amidst and Python.