Wow #1 · Ore per chunk
~24
Mean diamond ore blocks per chunk (median 24) across 1,024 chunks.
Wow #2 · Habit vs peak layers
~2.4×
Ore in Y=−54…−50 ÷ ore in Y=−59 plus −58 in this count.
A Java playtest: ore that actually exists in generated chunks. Not a wiki reprint of the generation rules.
Digging advice often collapses to “mine at −59.” In these worlds, a short strip held more total ore than the two named peak layers alone.
1. What did I count?
A chunk is a 16×16 block column of the world. For each seed I started a vanilla server, force-loaded blocks 0…255 on X and Z (chunks 0…15), saved the world, and scanned the region files.
Seeds: 42, 8012026, −13371337, 20260814.
2. Where did the ore sit?
In this sample the single busiest layer was Y=−53 (644 blocks), not −59. Y=−59 still ranked #3. Y=−58 ranked #10.
Diamond ore blocks by Y (measured)
Real count from 1,024 Java chunks. Y=−64 to 16. Not an illustration.
| Band | Ore blocks | Of this count |
|---|---|---|
| Wiki peak pair (Y=−59 and −58) | 1,172 | 4.7% |
| Common dig strip (Y=−54 to −50) | 2,840 | 11.4% |
| Deep band (Y=−64 to −48) | 8,175 | 32.9% |
| All ore in Y=−64…16 | 24,821 | 100% |
The habit strip is five layers. The wiki pair is two. Per-layer, −59 is still elite in this sample. As a dig plan, a short strip gathered more ore than pinning only the two peak numbers.
3. What about ore on cave walls?
About 2.9% of counted ore blocks had at least one open-air, water, or lava neighbor at save time. That is a cave-exposure readout for these chunks. It is not the same as the wiki’s “chance to skip next to air” during generation.
Deep deepslate still pays. In my 1,024-chunk Java test, mining a short strip around the deep peak beat treating −59/−58 as the whole answer. For the wiki rule of thumb, see the diamond Y FAQ.
