Iron Ore generation was not increased in 21w13a

Can you prove it?

I analysed it:

Mean of 41.323 iron ores per chunk (including deepslate) with (almost) random sample of 31 chunks in 21w11a
Mean of 40.838 iron ores per chunk (including deepslate) with (almost) random sample of 31 chunks in 21w13a

Both are normal worlds although the seed is different.

90% Confidence interval for data in 21w13a was 38.489 to 43.188 iron ores per chunk (t-conf. interval). 41.323 is within this range, hence no evidence to suggest change in amount of iron ore generated at 10% significance level.

Whether they tweaked the generation is another thing, but the tweaks probably didn't have any overall impact on the frequency.

You can also look at the code. The iron ore features were let unchanged. The "old_iron_ore" feature was re-added, but it's not actually used in generation.

I didn't analyze it but it feels like this is still the case in 21w17a when using the data pack for the new generation. I've been mining on Y=32 and copper is 4-5 times more common than iron.

Edit: I have now analyzed it. I was in fact mining on the wrong layers. On Y=32 copper is more than twice as common as iron and the higher it gets the more extreme it gets, e.g. around Y=40 copper is 4-5 times as common.

The distribution seems to be accurate to the graphic from 21w10a: link

Here are the results of my analysis. I measured in blocks of 10 y coordinates in a 100x100 region with few caves. The numbers are the total numbers from each corresponding 100x10x100 area:

Coal
>60 1055
>50 1305
>40 959
>30 757
>20 688
>10 231
>0 41
.. 0

Iron
>60 0
>50 84
>40 135
>30 228
>20 303
>10 422
>0 300
>-10 201
>-20 96
>-30 12
>-40 56
>-50 84
>-60 90
>-70 9

Copper
>60 169
>50 637
>40 603
>30 637
>20 570
>10 424
>0 370
>-10 77
.. 0

Gold
>60 0
>50 0
>40 0
>30 3
>20 25
>10 96
>0 105
>-10 148
>-20 208
>-30 197
>-40 145
>-50 93
>-60 20
>-70 1

Lapis
>60 14
>50 29
>40 47
>30 66
>20 65
>10 106
>0 138
>-10 153
>-20 140
>-30 45
>-40 53
>-50 57
>-60 35
>-70 1

Redstone
>60 0
>50 0
>40 0
>30 0
>20 0
>10 0
>0 75
>-10 122
>-20 80
>-30 96
>-40 272
>-50 368
>-60 422
>-70 66

Diamond
>10 4
>0 11
>-10 31
>-20 23
>-30 53
>-40 61
>-50 44
>-60 50
>-70 4

Nice Data @unknown!

I don't think this issue is disputed. It's been marked as plausible, and assigned to Hendrik who we know is dealing with the world generation issues - of which they have been revising every snapshot. They also said that the change in world height had caused a lot of deep issues in the generation - so much so it contributed to splitting the releases in two. The fact the issue persists in a lot of the snapshots suggests to me, some deep rewriting of old code is being undertaken probably caused by world height changes.

Besides all you have to do is create a few worlds in 1.16 and then in the snapshots and enter spectator mode and fly around in the ravines. You'll notice visible iron is often absent entirely in ravines in the snapshots and you'll only find a few blocks in the deep dark.

However - What I'd like to add this thread is that this issue is also in 21w18a, the latest snapshot, but hasn't been marked as so yet.