How many bookshelves give which level, exactly where those bookshelves have to stand in Java and in Bedrock, and what a level 30 enchant can and cannot put on each material. Every table below is a fixed reference; the widget works the same numbers for your own setup.
The number on each of the three slots is a level requirement, and the range it can land in depends only on how many bookshelves the table can see. This is the complete chart for 0 to 15 bookshelves. Fifteen is the ceiling: with 15 the bottom slot always reads 30, and no amount of extra shelving pushes any slot higher.
| Bookshelves | Top slot | Middle slot | Bottom slot |
|---|---|---|---|
| 0 | 1–2 | 1–6 | 1–8 |
| 1 | 1–3 | 1–7 | 2–9 |
| 2 | 1–3 | 2–8 | 4–11 |
| 3 | 1–4 | 2–9 | 6–12 |
| 4 | 1–4 | 3–10 | 8–14 |
| 5 | 1–5 | 3–11 | 10–15 |
| 6 | 1–5 | 3–12 | 12–17 |
| 7 | 1–6 | 3–13 | 14–18 |
| 8 | 1–6 | 4–14 | 16–20 |
| 9 | 1–7 | 4–15 | 18–21 |
| 10 | 2–7 | 5–16 | 20–23 |
| 11 | 2–8 | 5–17 | 22–24 |
| 12 | 2–8 | 5–18 | 24–26 |
| 13 | 2–9 | 5–19 | 26–27 |
| 14 | 2–9 | 6–20 | 28–29 |
| 15 | 2–10 | 6–21 | 30 |
The three formulas behind the chart, as the wiki gives them, with b the bookshelf count capped at 15 and randInt(n) a whole number from 0 to n inclusive:
The cap is applied before all three slot formulas here. The wiki states that the bookshelf count is capped at 15 and writes the cap inside the base formula; its bottom-slot line shows a bare bookshelf count. Read with the cap, the formulas reproduce all 48 cells of the chart above, which is how this page checks them.
The level on a slot is what you must have, not what you pay. The wiki's enchanting table page is explicit that the charge equals the lapis lazuli count: the bottom slot at level 30 takes 3 levels and 3 lapis lazuli.
A bookshelf counts when it stands exactly two blocks from the table along one horizontal axis and no more than two blocks along the other, level with the table or one block higher. That is the outer ring of a 5×5 square centred on the table: 16 positions per layer, two layers, 32 positions in all. Only 15 are ever needed. Each ring holds three kinds of position, and the two editions disagree about which gap block they inspect.
| Position class | Offset from table | Count (both layers) | Java checks | Bedrock checks |
|---|---|---|---|---|
| Straight | (2, 0) | 4 × 2 = 8 | The block between table and bookshelf, at the bookshelf's height | The block between table and bookshelf, at table height and the block above it |
| Offset | (2, ±1) | 8 × 2 = 16 | The block right next to the table on that side, at the bookshelf's height: offset (1, 0) | The column right next to the bookshelf on the table side, at table height and the block above it: offset (1, ±1) |
| Diagonal (corner) | (2, ±2) | 4 × 2 = 8 | The block diagonally next to the table, at the bookshelf's height: offset (1, ±1) | The column diagonally between them, at table height and the block above it: offset (1, ±1) |
| What may fill the gap | — | 32 positions | Any block in the tag #enchantment_power_transmitter; the wiki names air, snow and short grass | Air only |
Two practical consequences. In Java a gap block is tested once, at the height of the bookshelf it serves, so an upper-layer bookshelf needs the upper gap block clear and ignores the lower one. In Bedrock both heights are tested for every bookshelf, and the replaceable-block allowance does not exist: snow cover or grass in the gap, harmless in Java, switches a Bedrock bookshelf off. A torch or any other ordinary block in the gap breaks the link in both editions; the wiki singles out torches as an example.
Glyph particles are not proof. The wiki notes that the floating glyphs can travel to the table from a blocked bookshelf, because the particles follow their own rules. Trust the level on the bottom slot, not the particles: if 15 bookshelves are placed and the bottom slot does not read 30, one of them is not counting.
The level on the slot is not the number the game uses to pick enchantments. It first adds a bonus that depends on
the item's enchantability, then nudges the total by up to 15% either way. The wiki calls the result
eCost; this page calls it the modified level. Each enchantment level owns a band of modified levels,
and a level can only appear if the modified level can land inside its band.
Only ⌊E / 4⌋ matters, so the eleven enchantability values on the wiki collapse into six tiers:
| Tier | Enchantability | ⌊E / 4⌋ | Materials and items | Modified level at 30 |
|---|---|---|---|---|
| 1 | 25 | 6 | Golden armor | 26–49 |
| 2 | 22 | 5 | Golden weapons and tools | 26–47 |
| 3 | 15, 14, 13, 12 | 3 | Leather armor, netherite armor, netherite weapons and tools, wooden weapons and tools, mace (15) · iron weapons and tools (14) · copper weapons and tools (13) · chainmail armor (12) | 26–43 |
| 4 | 10, 9, 8 | 2 | Diamond armor, diamond weapons and tools (10) · iron armor, turtle shell (9) · copper armor (8) | 26–40 |
| 5 | 5 | 1 | Stone weapons and tools | 26–38 |
| 6 | 1 | 0 | Book, bow, crossbow, fishing rod, trident | 26–36 |
The window arithmetic is short. At level 30 the smallest pre-nudge total is 31 and 31 × 0.85 rounds to 26 for every material. The largest is 31 + 2 × ⌊E/4⌋, times 1.15: 43 × 1.15 gives 49 for golden armor, 35 × 1.15 gives 40 for diamond, and 31 × 1.15 gives 36 for a bow or a book.
Each cell is the highest level whose band overlaps that tier's window. A dash means no item in that tier takes the enchantment in Java. Tier 6 includes books, which accept every enchantment in the list, so its column doubles as the book ceiling. Bands are the wiki's effective ranges.
| Enchantment (max) | Top band | Tier 1 · 26–49 | Tier 2 · 26–47 | Tier 3 · 26–43 | Tier 4 · 26–40 | Tier 5 · 26–38 | Tier 6 · 26–36 |
|---|---|---|---|---|---|---|---|
| Protection (IV) | 34–45 | IV | — | IV | IV | — | IV |
| Fire Protection (IV) | 34–42 | IV | — | IV | IV | — | IV |
| Feather Falling (IV) | 23–29 | IV | — | IV | IV | — | IV |
| Blast Protection (IV) | 29–37 | IV | — | IV | IV | — | IV |
| Projectile Protection (IV) | 21–27 | IV | — | IV | IV | — | IV |
| Respiration (III) | 30–60 | III | — | III | III | — | III |
| Aqua Affinity (I) | 1–41 | I | — | I | I | — | I |
| Thorns (III) | 50–100 | II | — | II | II | — | II |
| Depth Strider (III) | 30–45 | III | — | III | III | — | III |
| Sharpness (V) | 45–65 | — | V | IV | IV | IV | IV |
| Smite (V) | 37–57 | — | V | V | V | V | IV |
| Bane of Arthropods (V) | 37–57 | — | V | V | V | V | IV |
| Knockback (II) | 25–75 | — | II | II | II | II | II |
| Fire Aspect (II) | 20–80 | — | II | II | II | II | II |
| Looting (III) | 33–83 | — | III | III | III | III | III |
| Sweeping Edge (III) | 23–38 | — | III | III | III | III | III |
| Lunge (III) | 21–41 | — | III | III | III | III | III |
| Efficiency (V) | 41–91 | — | V | V | IV | IV | IV |
| Silk Touch (I) | 15–65 | — | I | I | I | I | I |
| Unbreaking (III) | 21–71 | III | III | III | III | III | III |
| Fortune (III) | 33–83 | — | III | III | III | III | III |
| Power (V) | 41–56 | — | — | — | — | — | IV |
| Punch (II) | 32–57 | — | — | — | — | — | II |
| Flame (I) | 20–50 | — | — | — | — | — | I |
| Infinity (I) | 20–50 | — | — | — | — | — | I |
| Luck of the Sea (III) | 33–83 | — | — | — | — | — | III |
| Lure (III) | 33–83 | — | — | — | — | — | III |
| Loyalty (III) | 26–50 | — | — | — | — | — | III |
| Impaling (V) | 33–53 | — | — | — | — | — | V |
| Riptide (III) | 31–50 | — | — | — | — | — | III |
| Channeling (I) | 25–50 | — | — | — | — | — | I |
| Multishot (I) | 20–50 | — | — | — | — | — | I |
| Piercing (IV) | 31–50 | — | — | — | — | — | IV |
| Quick Charge (III) | III: no band listed · II: 32–50 | — | — | — | — | — | II |
| Density (V) | 37–57 | — | — | V | — | — | IV |
| Breach (IV) | 42–92 | — | — | IV | — | — | III |
The rows people ask about most:
One source discrepancy, stated plainly. The Protection row in the wiki's receivable-enchantments table reads 5–12, 13–20, 21–28, 29–37, which is character for character the Blast Protection row. The wiki's own Protection page defines the cost as a minimum of 1 plus 11 per level and a maximum of 12 plus 11 per level, and its Blast Protection page gives 5 plus 8 and 13 plus 8, which is what produces that row. This page uses the definition: Protection I 1–11, II 12–22, III 23–33, IV 34–45. Both pages were read again on 20 September 2026 and the row had not been corrected.
There are three different reasons, and they have different fixes.
Seven enchantments are never offered by a table in either edition, whatever the level or material: Mending, Frost Walker, Curse of Binding, Curse of Vanishing, Soul Speed, Swift Sneak and Wind Burst. The wiki's selection rule is that only non-treasure enchantments are candidates. Channeling is not on this list: it is an ordinary trident enchantment with a weight of 1, which makes it uncommon, not unavailable (17.0% on a trident at level 30 in Java).
That is the ceiling table above: Sharpness V off gold, Efficiency V on diamond, Power V and Thorns III anywhere.
This is the one that surprises people. Bands have a top as well as a bottom. A level 30 slot never produces a modified level below 26, so an enchantment whose highest band ends in the twenties is squeezed out, and when the modified level overshoots the band the enchantment is not a candidate at any level.
| Enchantment | Top band | Window at level 30 | Chance at level 30 | Best slot level | Chance at best level |
|---|---|---|---|---|---|
| Feather Falling IV | 23–29 | 26–40 | 2.7% | 23 | 34.3% |
| Projectile Protection IV | 21–27 | 26–40 | 0.1% | 21 | 16.9% |
So boots meant for Feather Falling IV are better enchanted from a slot reading 23 than from one reading 30, and a slot reading 21 is the best shot at Projectile Protection IV. Those figures are the maximum over every slot level from 1 to 30, not a hand-picked example.
Whatever the table will not give you has to come from books instead, and once you have the books the Minecraft Enchantment Calculator works out the cheapest anvil order to combine them without hitting “Too Expensive”.
| Topic | Java Edition | Bedrock Edition |
|---|---|---|
| Gap blocks tested per bookshelf | One, at the bookshelf's height | Two: at table height and the block above |
| Gap tested for an offset bookshelf | The block right next to the table | The column right next to the bookshelf |
| What may fill the gap | Anything in #enchantment_power_transmitter (air, snow, short grass named) | Air only |
| Blocks that can stand in for bookshelves | Anything added to the tag #enchantment_power_provider | Not described |
| Thorns from the table | Chestplate only | Helmet, chestplate, leggings and boots |
| Axe pool | Efficiency, Unbreaking, Fortune, Silk Touch | The same, plus Sharpness, Smite and Bane of Arthropods |
| Mace pool | Density, Breach, Unbreaking | The same, plus Fire Aspect, Smite and Bane of Arthropods |
| Sweeping Edge | In the sword pool | Not offered (the wiki marks it Java only) |
| Percentages on this page | Exact, labelled Java | None given |
The last row is deliberate. The wiki describes one selection algorithm and does not split it by edition, but it never states that Bedrock's numbers are identical either, so this page gives Bedrock its pools, ceilings and placement rules and keeps every percentage under a Java label.
Once the modified level is fixed, the game builds a candidate list: every non-treasure enchantment the item takes, each at the highest level whose band contains the modified level. Then it runs a four-step loop.
At level 30 on diamond the first continuation succeeds between 54% and 82% of the time depending on the modified level, which is why a diamond pickaxe leaves the table with one enchantment 32.0% of the time, two 44.4% and three 23.6%. A book that ends the loop with several enchantments loses one of them at random.
All of that randomness runs from a seed stored with the player, not with the table or the world. The seed only changes when that player enchants something, so taking the item out and putting it back shows the same three offers, and so does walking to a different table with the same bookshelf count. Spending one level on the top slot of a spare book or tool is the usual way to get a fresh set.
The percentages on this page are computed by walking every branch of that loop and every possible modified level with its exact probability. Nothing is simulated or sampled.
Sources. Minecraft Wiki (minecraft.wiki), retrieved 20 September 2026: “Enchanting table mechanics” — sections Bookshelf placement (Java Edition, Bedrock Edition, 15 bookshelf arrangements), Experience cost, Enchantment selection (Enchantment cost, Candidates, Selection), Enchanting seed, Receivable enchantments; “Protection”, “Blast Protection” and “Thorns” — section Enchantment definition, and Thorns primary items; “Enchanting” — Treasure enchantments; “Enchanting Table” — Usage. Formulas, constants and table values are reproduced as data; the wording and all percentages are this page's own. Last verified: September 2026.