Enchanting

Enchanting only happens with drop-only scrolls. Reading one bumps a piece of gear by +1. Past a guaranteed band, failure destroys the item — a voluntary gold/gear sink instead of a durability system.

There are two scrolls:

Scroll	Target	+N effect	Category	Price	World drop rate
Scroll of Enchant Weapon	Your equipped main-hand weapon	+N to hit and damage rolls	enchant_scroll	1,200c	1% (0.5% on kills at effective level 8 or below)
Scroll of Enchant Armor	A random piece of worn armor	+N to Guard	enchant_armor_scroll	1,200c	1% (0.5% on kills at effective level 8 or below)

Merchants don’t sell either scroll — the only source is a world drop, and you can sell one back if you find one you don’t want. Effective level is the monster’s level after dungeon-depth scaling, so a low-level monster on a deep floor doesn’t count as low-level; chests and barrels/crates have no level and drop at the full rate.

Success effect

A successful enchant now plays about 5 seconds of light effects (weapon or armor alike) — the server broadcasts the event to everyone on the same floor, including yourself. A weapon enchant traces golden filaments and a halo along the blade while your character holds it raised overhead; an armor enchant radiates golden rays outward from the chest. Moving, fighting, interacting, or mounting/dismounting immediately cuts the pose short and returns to normal — it’s purely a visual cue and doesn’t affect the enchant’s odds or outcome.

Whetstone Oil

Reading either scroll now also spends one bottle of Whetstone Oil — without one the read is refused outright and the scroll isn’t consumed.

Item	Detail
Item	whetstone_oil (category reagent, stackable, weight 0.2)
Buy	Sold by Rica for 1g (10,000c); sells back at 40% = 4,000c
Spent	1 bottle per read — whether it succeeds or destroys the item

The oil has no use on its own — its only purpose is enchanting. It’s purely an extra gold sink on the system and doesn’t touch any of the odds above.

Reading rules

Common to both:

• You can’t read while incapacitated (downed) — the scroll isn’t consumed.
• Locked equipment is excluded from the enchant target (protocol v69) — no weapon equipped, no armor worn, or every candidate locked, all get rejected the same way, without consuming the scroll. See Item locks.
• No Whetstone Oil also gets rejected, consuming neither the scroll nor the oil — this check comes after the target check, right before anything is spent.
• You can’t read while a trade window is open: the oil is drawn by item definition, not by instance, so it can’t be reconciled against quantities reserved in the trade.
• If there’s a target and oil to spend, the scroll and the oil are both consumed whether the roll succeeds or destroys the item.
• The odds roll happens server-side outside the lock; the result comes back with an updated inventory snapshot and Guard together.

Armor-only rules:

• The target is picked uniformly at random among currently worn and unlocked category=armor items — helmet, body armor, shield, gloves, pants, and boots all count.
• Weapons and accessories (rings, necklaces, belts) are never targeted. ring_of_protection grants Guard but is an accessory, so it can’t be enchanted.
• Only pieces worn and unlocked at the moment you read are candidates — taking a piece off or locking it removes it from the pool, which effectively lets you choose your target (see “What the random target actually means” below).

The success ladder

Weapons are guaranteed through +5; armor is guaranteed through +3. Past that, both follow the same curve (75% / 50% / 25% → halving from there → a 1% floor), but armor’s curve starts two steps earlier — armor has 6 slots, so giving it the same guaranteed band as a single weapon would multiply total Guard gain by 6.

Current → next	Weapon success	Armor success
+0 → +1	100%	100%
+1 → +2	100%	100%
+2 → +3	100%	100%
+3 → +4	100%	75%
+4 → +5	100%	50%
+5 → +6	75%	25%
+6 → +7	50%	12.5%
+7 → +8	25%	6.25%
+8 → +9	12.5%	3.12%
+9 → +10	6.25%	1.56%
+10 → +11	3.12%	1% (floor)
+11 → +12	1.56%	1%
+12 and up	1% (floor)	1%

The failure chance is whatever’s left over in that row, and destruction is the only failure outcome. The ladder stops thinning at 1% — there’s no theoretical cap, just an exploding cost.

Expected cost

Assuming a failure also loses the base item and resets to +0, here are the expected values (recurrence: E[item](N) = E[item](N-1) / p_N, E[scroll](N) = (E[scroll](N-1) + 1) / p_N):

Weapon (one main-hand weapon):

Target	Last step’s odds	Expected weapons	Expected scrolls
+5	100%	1	5
+6	75%	1.3	8
+7	50%	2.7	18
+8	25%	10.7	76
+9	12.5%	85.3	616
+10	6.25%	1,365	9,872

Armor (one slot):

Target	Last step’s odds	Expected armor pieces	Expected scrolls
+3	100%	1	3
+4	75%	1.3	5
+5	50%	2.7	13
+6	25%	10.7	55
+7	12.5%	85.3	445
+8	6.25%	1,365	7,141

In practical terms: pouring scrolls into a +0 piece of armor gets it destroyed after an average of 5.2 scrolls, landing around +4 on average (a weapon takes an average of 7.2 scrolls, landing around +6). Realistic caps are roughly armor +5 (37.5% reach rate) to +6 (9.4%), and weapon +7 (37.5%) to +8 (9.4%).

Effect on Guard

Armor +N is added straight to the wearer’s Guard while worn — and Guard is the hit threshold an attacker has to clear (see Combat), so each +1 drops every attack against that character by 5 percentage points (one d20 step). Past the attacker’s plain roll range, the exploding die takes over and the drop gets much steeper — Guard 40 is 0.25% against a low-level monster, for example.

• Maxing all 6 slots to the guaranteed band (+3) is +18 total Guard — but the actual cost to get there depends heavily on how you play it (see below).
• Weapon +N never touches Guard; armor +N never touches to-hit or damage.
• A destroyed piece disappears from your equipment immediately and Guard drops right away; the server pushes the updated snapshot and Guard together, so the client never has to recompute it.

What the random target actually means

The target pool is only whatever’s worn and unlocked at the moment you read, so taking off or locking everything except the piece you want enchanted effectively lets you choose your target. The cost to bring all 6 slots to +3 differs sharply by playstyle (20,000-run simulation):

Playstyle	Expected scrolls	Expected destroyed	All 6 slots survive
Leave one slot as the only candidate (rest off or locked) and read	18	0	100%
Read with all 6 worn and unlocked	25.7	3.1	2.8%

The random target only bites once you’re past the guaranteed band — once a piece is already +3, getting picked means eating a 25% destroy chance, so reading with all 6 as candidates loses an average of 3 pieces. Taking a finished piece off or locking it removes it from the pool and keeps its guaranteed band fully guaranteed.