Every layer needs a job.

Separate the rule from the theory

Confirmed: official guidance treats march size and troop-specific buffs as important, and its ranged guide prioritises attack, then HP, then defence. Not confirmed: Evony does not publish its hidden rounds, target-selection order or one correct layer size. Specialist guides call those mechanics battle-tested theory, and player reports disagree on whether small layers should contain 10, 100 or 1,000 troops.

Price the support before filling the march

Main block = march capacity − total support layers

Worked capacity example: a 2,000,000 march uses T12–T14 ranged as its main block. That leaves 53 other type-tier slots across T1–T14. At 1,000 troops per slot, support uses 53,000 capacity and the main block receives 1,947,000. At 100 per slot, support uses 5,300 and the main block receives 1,994,700. The leaner preset returns 47,700 spaces, or 2.45% more core troops than the first preset. It is a test candidate, not a universal answer.

Round numbers are tidy. Battle reports have not promised to be.

Name the job of every group

• Core damage: the troop family and tiers supported by the general, gear, refines and buffs. This receives most of the march.
• Same-family support: lower or adjacent tiers that preserve damage across the fight. Their useful size depends on the target and must be tested.
• Cross-family layers: small groups intended, in player-tested theory, to absorb targeting actions or alter exchanges. This is not an official formula.
• Context layer: troops added for a rally, reinforced building or specific defence. If the context changes, the preset may need to change too.

Make the report earn the slot

Copy the preset. Lock the general pair, gear, dragon or spiritual beast, march size, buffs, target type, reinforcement and battlefield context. Change only the layer size. Compare at least 3 similar reports.

Exchange ratio = enemy troop power lost ÷ your troop power lost

Worked report example: preset A loses 200M troop power and removes 280M, ratio 1.40. Preset B loses 190M and removes 300M, ratio 1.58. That is about 12.9% better than 1.40, but only comparable reports can support the choice. One attractive report is a clue, not a constitution.

Use four stopping rules

1. Keep the leaner preset only if the exchange ratio and surviving core remain stable or improve across 3 comparable reports.
2. Make separate presets for solo keep hits, rallies and reinforced buildings when their reports behave differently.
3. Do not copy defensive layer counts into an attack march. A city and a march have different capacity, purpose and loss risk.
4. Do not use monster reports to validate PvP layers. The target, combat rules and useful measurements are different.

A published percentage is a starting hypothesis, not a warranty. New troop tiers, buffs, debuffs, reinforcements and opponent composition can change the result. Save the old preset until the new one has earned the job.

Sources: Official ranged-troop guidance; Specialist battle-mechanics and layer testing; Specialist attack-layer ranges; Player discussion comparing thin layers.

Evidence reviewed 2026-09-28. Official Evony material confirms march size and troop-specific buffs but does not publish the hidden-round or targeting algorithm. Layer explanations and counts are specialist theory and player testing. Recheck after game updates.

Budget the support. Test one change. Let the report hire the layer.

— Kat
The preset has been asked to justify its headcount.