Kat’s verdictCompare a complete before-and-after build, not two isolated pieces. Count relevant fixed attributes, relevant refines, lost and gained set attributes, and the condition on every line. Craft only when the candidate wins the job you actually give the march.

Name the march job first

Attack, defense, reinforcement, rally leadership and monster hunting are different jobs. A line that applies only while attacking does not improve a defending wall. Enemy debuffs are not interchangeable with your own troop buffs. Write one job and one troop type at the top of the comparison. Otherwise every percentage will volunteer for duty.

Audit every threshold that changes

Civilization Equipment has individual equipped attributes and set attributes that activate at matching-piece counts. Current specialist tables document 2, 4 and 6-piece thresholds, while official equipment pages confirm that each piece can carry several conditional buffs and debuffs. A swap can improve one slot while crossing a threshold in either direction. Check the live Set Attribute screen because new sets and adjustments continue to arrive.

Use one swap ledger

For the chosen job, copy only the lines that can apply. Record the old piece, the new piece, both sets of refines, and every set attribute that turns on or off. Keep attack, defense, HP, march size, rally capacity and enemy debuffs in separate rows. A single grand total hides tradeoffs that battle reports will later introduce with less tact.

Net change for one row = new piece + new relevant refines + gained set attributes − old piece − old relevant refines − lost set attributes

Worked example: the attractive sidegrade

Suppose a ranged attacking march is considering one replacement. These are example values, not a claim about a named set. The method is the point.

The larger gold number has submitted only the first page of its application.

Price the rebuild, not just the craft

Official Evony rules say advanced refining unlocks only after all four refining attributes on a Civilization Equipment piece reach Gold. The extra specific attribute then consumes more Gold and Refining Stones. A candidate with weak refines is therefore unfinished, while a finished old piece already contains sunk materials. Sunk cost should not decide the winner, but the remaining cost should decide whether today is the day to replace it.

Run a five-step replacement test

  1. Screenshot the current six pieces, their refines, the active set attributes and the march statistics.
  2. Build the ledger before selecting or crafting a scroll. Include every threshold that the swap breaks or creates.
  3. Use the preview or temporary equip screen when available. Confirm that the displayed march size, rally capacity and applicable buffs move as expected.
  4. Compare several reports against the same target type and similar conditions. One report is weather; a small series is evidence.
  5. Keep the old piece locked until the candidate has the required refines and the reports support the intended trade.

Use three stopping rules

Sources: Official Munsö equipment attributes; Official advanced refining and equipment upgrade rules; One Chilled Gamer Civilization Equipment and set tables; Evony Guide Wiki mixed and unified build comparisons; Player discussion recommending full-build spreadsheets and report tests.

Evidence reviewed 2026-10-03. Official sources confirm conditional piece attributes and the advanced-refining unlock. Current set thresholds and mixed-versus-unified comparisons come from independent specialist tables; the ledger and repeated-report test are practical inferences from those mechanics and player testing. Exact live values, ownership buffs, Star Atlas effects and newly released sets can vary, so the in-game Equipment and Set Attribute screens control the final decision.

A matching set is not automatically best. A mixed set is not automatically clever. The complete ledger gets the vote.