Minecraft: Why Your Anvil Says "Too Expensive"

The whole problem in one sequence

0, 1, 3, 7, 15, 31, 63.

That's the prior work penalty — the hidden counter on every item recording how many times it's been through an anvil. It's 2ⁿ − 1, where n is the number of previous anvil uses.

In Java Edition an anvil refuses any combine costing 40 levels or more.

The penalty alone reaches 63 on the sixth use.

So an item is finished after five anvil operations, no matter how cheap the enchantment is and no matter how many levels you've banked. You cannot out-level it. The anvil simply says no.

What a combine actually costs

Three things added together:

1. The level cost of the enchantments being applied
2. The prior work penalty of each of the two items — both of them
3. +1 if you're renaming it in the same operation

Both items contribute their own penalty, which is why combining two heavily-worked items is so brutal. You pay for both histories.

The result is worse than either input

Here's the mechanic that makes this compound so fast:

The output takes the higher of the two inputs' counts and adds one.

So combining an item at penalty 3 with a completely fresh book at penalty 0 gives you an item at penalty 7.

The fresh book didn't help. The anvil looked at the worse of the two, and doubled it.

This is why putting a well-used tool into an anvil for one small addition hurts twice — you pay its penalty now, and you double it for next time.

So the order you combine in is worth more than any other trick

Say you have four books to put on a tool.

One at a time: the tool's penalty is 0, then 1, then 3, then 7 — you pay it at every step.

Total: 11 levels of pure penalty.

Merged in pairs first: combine book+book (cost 0), book+book (cost 0), then the two double-books together (cost 2), then apply the finished book to the tool (cost 3).

Total: 5.

Identical result. Less than half the price. And the tool has been through the anvil exactly once, so it's still cheap to work on later.

With eight books it stops being an optimisation

• One at a time: 247 levels of penalty
• Merged as a balanced tree: 17

247 isn't merely expensive — it's impossible. You hit the 40-level wall partway through and the anvil stops cooperating entirely.

At that scale the combining order is the difference between the item existing and not existing.

The formulas, if you want them: adding k books one at a time costs (2ᵏ − 1) − k in penalties. Merging them as a balanced tree costs k × (log₂k − 1) + 1.

The rule

Build the book first. Touch the tool last.

Combine every enchantment you want into one super-book, pairing them up as evenly as you can rather than adding them to a growing pile. Then apply the finished book to the tool in a single operation.

With two books there's no advantage — the calculator will show a saving of 0, correctly, because there's no ordering to get right. From four upward it matters, and by eight it's decisive.

Repairing counts as an anvil use

This catches people constantly.

Patching a diamond pickaxe with diamonds in an anvil is an anvil operation like any other. It raises the counter and doubles the penalty for next time.

Which is why a tool you keep topping up eventually becomes unrepairable, even though you never enchanted it again after the first day.

Mending avoids this completely. Repairing through experience orbs isn't an anvil operation and never touches the counter.

That's the real reason Mending is the most valuable enchantment in the game — not that it repairs things, but that it repairs them without incrementing the number that's going to kill the item.

If you've already ruined something

A grindstone resets the penalty to 0 — and strips every enchantment off the item doing it.

You get a clean item back and nothing else.

Occasionally that's the right trade on a tool you've genuinely destroyed, especially if you can rebuild the enchantments from books you still have lying around. It's worth knowing the option exists before you throw the item in lava.

It is not a way to keep the enchantments.