How to evaluate member performance in Guild Wars in MU Online
Build a fair performance evaluation system for guild members in Castle Siege and Guild Wars in MU Online, with combat, presence, and teamwork metrics.
Guild Wars and Castle Siege are the competitive peak of MU Online, and the outcome rarely depends only on who has the strongest gear — it depends on coordination, presence, and each member executing their role well within the team. Guilds that grow sustainably have a clear process for evaluating eac
Guild Wars and Castle Siege are the competitive peak of MU Online, and the outcome rarely depends only on who has the strongest gear — it depends on coordination, presence, and each member executing their role well within the team. Guilds that grow sustainably have a clear process for evaluating each member's war performance, deciding who makes the main roster, who stays as reserve, and who needs attention before the next dispute. Without this process, roster decisions become a source of internal conflict and, in the worst case, cause good players to leave out of a sense of unfairness. This tutorial presents objective metrics by class role, along with a ready-to-use post-war evaluation model.
Why evaluate war performance
Competitive guilds fight scheduled Castle Sieges and guild wars weekly, and each dispute has limited slots in the main roster — usually between 40 and 100 players depending on the server. Without criteria, the roster becomes internal politics: whoever is friends with the leader gets in, whoever isn't stays out even if they perform better. An objective evaluation system solves two problems at once: it improves competitive results (putting whoever actually performs on the front line) and reduces social friction (because the decision is justifiable with data, not personal preference).
Direct combat metrics
The foundation of any war evaluation is combat numbers, pulled from the server's war log or from direct observation during the dispute.
| Metric | How to measure | Relevance |
|---|---|---|
| Confirmed kills | War log or manual count by an observer | High for damage classes (Dark Knight, Blade Master, Magic Gladiator) |
| Deaths | War log | Should be read together with survival time, not in isolation |
| Assists (damage applied before an ally's kill) | Estimated by observation, hard to log automatically | Medium — shows contribution beyond the final kill |
| Average time alive per life | Time between spawn/entry and death | High — a player who dies in 5 seconds every life isn't holding territory |
| Damage absorbed (tanking) | Observation of how many attacks the player "pulled" onto themselves | High for Dark Knight and Magic Gladiator in front-line roles |
A player with 15 kills and 20 deaths may be performing better than one with 8 kills and 2 deaths, if the former is holding the front line and opening space for the team — that's why the reading needs to be combined with the player's declared role.
Metrics by class role
Not every class should be evaluated by the same criteria. A common mistake is measuring support players by K/D and wrongly concluding they "didn't help."
| Class/role | Primary metric | Secondary metric |
|---|---|---|
| Dark Knight / Blade Master (front line) | Time holding the objective/tanking | Kills and damage absorbed |
| Dark Wizard / Soul Master (ranged damage) | Kills and area damage | Positioning (not dying surrounded) |
| Elf / Muse Elf (support) | Team buff uptime and healing done | Own survival (a dead support heals no one) |
| Magic Gladiator (hybrid) | Flexibility — switches between damage and tank as needed | Kills + time alive |
| Dark Lord (command/control) | Use of control skills (fear, summon) at the right moment | Push leadership, not just individual combat |
Objective and presence metrics
Wars aren't won by kills alone — they're won by completing the objective (destroying the gate, holding the crown, controlling the tower). Evaluate separately:
- Objective push: does the player actively participate in attacking/defending the structure, or go chasing easy kills far from the objective?
- Presence/punctuality: did they arrive on time for formation and stay until the end, or arrive late and leave early?
- Tactical positioning: do they follow the formation agreed on by the leader, or act alone ignoring commands?
- Call communication: do they report enemy position and needed reinforcements, or stay silent the whole fight?
Post-war scorecard model
A scorecard applied right after each relevant war, scored 1 to 5 per dimension:
| Dimension | Weight | High-score example | Low-score example |
|---|---|---|---|
| Combat (role-specific) | 35% | Fulfilled the class role notably | Was out of position, no role contribution |
| Objective/push | 25% | Actively participated in attack/defense | Went chasing isolated kills, far from the objective |
| Presence and punctuality | 15% | Arrived on time, stayed until the end | Was late or left early without notice |
| Communication and tactical discipline | 15% | Followed formation and reported over call | Repeatedly ignored the leader's commands |
| Team behavior | 10% | Collaborative, doesn't publicly blame teammates | Public argument after a loss, blaming allies |
Consolidate the score across the last 4 wars (not just the most recent one) before deciding on a roster change — this avoids overreacting to a single bad war, which could have been connection issues or lag rather than actual performance.
How to structure the roster by result
With the scorecard consolidated, define roster tiers that are clear and communicated in advance:
| Consolidated average score range | Roster position |
|---|---|
| 4.0 to 5.0 | Fixed main roster |
| 3.0 to 3.9 | Rotating main roster (alternates with others in the same range) |
| 2.0 to 2.9 | Reserve, with an agreed improvement plan |
| Below 2.0 | Out of the war roster until re-evaluation |
Communicate the tier and the reason for the placement directly to the member, not only publicly on a ranking — this reduces the sense of public humiliation and opens space for the member to ask how to improve.
Individual feedback after evaluation
Just as with staff evaluation, the number alone doesn't change behavior — structured feedback does. After consolidating 4 wars, have a quick conversation (even by text) with each member outside the main tier, explaining specifically what weighed on the score (e.g., "died too early in 3 of the 4 wars by entering the objective alone before the leader's signal") and a practical plan for the next war.
Special cases: newcomers and justified absence
A player who just joined the guild shouldn't be evaluated by the same standard as someone with 6 months of war experience — give them an adjustment period of 2 to 3 wars before factoring their score into roster decisions. Likewise, an absence announced in advance (a personal commitment communicated ahead of time) shouldn't count the same as an unannounced absence — treat the two situations explicitly differently in the guild's policy, documented in the rules channel.
Common errors and fixes
| Symptom | Likely cause | Fix |
|---|---|---|
| Support players unfairly rated as "weak" | K/D metric applied equally to all classes | Use function-specific metrics (healing, buffs, tanking) |
| Roster decision looks like favoritism | No scorecard or published criteria | Publish roster tiers and the 4-war consolidated score |
| Public conflict after a loss | Critical feedback given in a public channel | Take critical feedback to a private conversation |
| A good member leaves the guild after one bad war | Decision based on 1 war, not history | Always consolidate the last 4 wars before any change |
| Inconsistent kill/death data between observers | No reliable war log | Use the server log or a war bot instead of manual counting |
War evaluation checklist
- Server war log (or dedicated observer) set up.
- Metrics by class role defined and published for the guild.
- Post-war scorecard applied after each relevant dispute.
- Roster tiers by consolidated score (4 wars) defined.
- Individual feedback scheduled for members outside the main roster.
- Justified vs. unannounced absence policy documented.
- Adjustment period for newcomers defined and respected.
With war evaluation structured, the natural next step is looking at the management of the resources that sustain that war — items, jewels, and Zen accumulated by the guild — which connects directly to the MU Online server creation tutorial and to the overall organization of your server's economy.
Frequently asked questions
Is kill/death ratio enough to evaluate a player in war?
No. K/D alone favors whoever farms easy kills and ignores players fulfilling support, objective-control, or leader-protection roles. Always combine K/D with objective and presence metrics.
How do I evaluate support classes that don't have high kill counts?
Use function-specific metrics: healing done, buffs maintained, survival time of the protected team, and participation in objective pushes. Comparing a support Elf to a Dark Knight by the same kill criterion is unfair and discourages class diversity.
How often should I review the war roster?
Review after every relevant war (weekly Castle Siege, at minimum) and do a monthly consolidation. Individual wars have high variance — only the cumulative results of several wars show a reliable pattern.
What do I do about a member who misses wars without notice?
Define a clear absence policy (notice with X hours in advance) and apply it consistently to everyone, including veterans. Recurring unannounced absence should lower priority for items and fixed slots in the main roster.
Is it worth using a spreadsheet or a bot for these metrics?
For small guilds, a shared spreadsheet is enough. For large guilds or federations, it's worth integrating a Discord bot that pulls war logs from the server automatically, reducing typos and disputes over numbers.