A Dofus Retro Ecaflip configuration in MoonBot must never choose a spell by its best roll. It compares distributions, survivable minimum, target, final cell and next round. Roulette, Smell, Luck and Feline Spirit each require an explicit adverse branch.
Begin with the Retro Ecaflip guide and confirm effects in the client. Automation may breach game terms and lead to sanctions.
Define the configuration goal
Choose speed, net XP, resources, safety or maximum bread cost. Add failure and duration constraints.
Attach median duration, first kill, worst quartile or realised kamas. Without a metric, buffs cannot be compared with attacks.
Version the profile
Store date, server, path, level, characteristics, gear, spells, ranks, AP/MP, range, critical rate and area. Loaded state must match the character.
Change one rule per experiment and retain the previous version. One item may alter several thresholds without a logic fault.
Build combat states
Use opening, spacing, buff, focus, conditional burst, survival, healing, contact, retreat, finishing and emergency. Each state defines entry, legal actions and exit.
Forbid buff and healing loops. A maximum round count guarantees transition to damage or retreat.
Capture a snapshot
Read positions, health, AP/MP, range, resistance, effects, cooldowns, critical rates, summons and reachable cells. Recalculate after movement or global effects.
Never reuse start-of-turn candidates after Leap, Roulette or Smell. Costs and paths may be stale.
Model damage rolls
For each spell store minimum, maximum, element, critical, possible healing, self-damage and probability. Calculate the distribution against target resistance.
Add the hostile action possible after a low roll. Correct scoring uses the full consequence rather than the isolated hit.
Create a health budget
Reserve health for next-round threat, self-damage and contact. Subtract the margin before risky actions are allowed.
When an adverse branch exceeds the budget, use stable damage, retreat or protection. Maximum output cannot justify probable death.
Create AP/MP budgets
Calculate buff, movement, attack and exit inside the complete sequence. Reserve retreat MP before ranking damage.
Prepare a rotation under Smell losses or hostile removal. Every round needs a legal action with tested minimum resources.
Select the opening
Compare immediate attack, movement, buff and summon by distance, predicted duration and first kill. Small packs often favour damage.
Wheel, Perception or Clover enter when enough attacks benefit. Summoning Claw enters for a named hostile action or future damage.
Score target priority
Weight next-round threat, healing, summons, movement, range, resistance, health and finishing cost. Add committed damage.
Lock focus until a material event. Roll variation should not scatter attacks.
Configure regular attacks
Earth calculates net Heads or Tails damage after possible healing; Fire calculates Topkaj lines and range. Playful Claw validates every affected cell.
Rank by finishing minimum, AP and final cell. Regular attacks are the reference against which risk is compared.
Configure Feline Spirit
Generate hostile-damage and self-damage branches under the active mechanic. Add the enemy response after each.
Permit only where kill probability is sufficient or self-damage remains under the health margin.
Configure Bluff
Calculate Water and Air separately against target resistance. Do not reuse an average from other families.
Reject Bluff where one branch makes no progress and leaves exposed contact. Hybrid profiles load both actual characteristics.
Configure Roulette
Encode every effect, beneficiaries, enemies helped, duration and focus impact. Calculate collective value in the current snapshot.
Allow only where the worst effect stays inside limits and expected value beats an attack. Roulette is never required to win.
Configure Smell
Store AP/MP distributions for the used rank and all affected fighters. Simulate essential actions by initiative.
Reject where a loss blocks healing, placement or retreat. Recalibrate after every rank change.
Configure Ecaflip’s Luck
Model recovery and amplified incoming damage, including critical effects. Select a target already under threat rather than a healthy partner.
Create a negative-branch response through protection, recovery, range removal or killing the attacker. Without it, reject the spell.
Configure critical hits
Load displayed rate by spell and weapon, then calculate weighted damage and utility. Never convert the 1/2 cap into certainty.
An effect requiring a critical receives a normal-hit fallback. The engine never retries indefinitely for a critical.
Configure healing
For Heads or Tails, All or Nothing and Felintion, calculate net healing, target, area, delay and offensive cost. Prevent uncompensated enemy recovery.
Trigger only where an allied action or caster survival is saved. Otherwise continue focus.
Configure Summoning Claw
Generate cells blocking no line, Leap or retreat. Log kitten AI by family before assigning a role.
Keep a branch when target or movement differs. The next round must not require one exact summon action.
Create finishing mode
When attack minimum is lethal, disable buffs, Roulette, Smell, Luck, summon and slow effects. Choose the smallest safe expenditure.
Handle Heads or Tails healing or a low roll. Select the next target only after confirmed death.
Create fallback branches
Handle missing line, short range, absent target, occupied cell, low AP/MP, lock, unknown resistance and cooldown. Every rejection leads elsewhere.
The final fallback deals safe damage, protects or retreats. No loop should reselect an impossible candidate.
Write an explainable log
Store state, snapshot, candidates, distributions, rejection rules, action, roll, final cell and result. Use stable identifiers.
During incidents find the first divergence. Final death may begin with an empty buff three rounds earlier.
Test in stages
Begin on a fixed board, then ten supervised fights and thirty comparable encounters. Introduce variable effects one at a time.
Test no-critical streaks, Smell losses, adverse Roulette, self-damage and unexpected summon AI. Rare branches belong in validation.
Validate the configuration
Confirm goal, version, states, snapshot, rolls, health, AP/MP, opening, target, attacks, risk, global effects, healing, summon, finishing, fallbacks, logs and tests.
The profile is ready when every choice is explainable, the worst roll retains an action and thirty battles respect limits.
Botting breaches the game's terms of service and carries a sanction risk. No tool can guarantee that an account will never be banned.
