Scars of Honor tier list — why there isn't one yet, and what the alpha showed

A tier list ranks classes against a stable set of rules. Scars of Honor does not have one yet, and the developers say so themselves.

• The game is unreleased. No release date and no Early Access date have been announced — early access is only ever discussed as an estimate. The release-date record tracks every dated statement about timing; none of them is a commitment.
• There has been one public test. The technical alpha ran 30 April – 11 May 2026, and the studio framed it as a technical test — a stress test for crash rates, networking and server stability, not a balance test (Technical testing goals and feedback focus, 2026-04-27). What was playable is on the playtest page.
• Four of ten classes were playable. Mage, Paladin, Druid and Ranger. Warrior, Assassin, Necromancer, Mystic, Pirate and Priest were not in the build (playtest guide, 2026-04-27). Six classes cannot be ranked, because nobody outside the studio has played them.
• Talent trees exist for the same four. Only those have shipped data, which is why the class catalogue can show a talent page for Mage, Paladin, Druid and Ranger and not for the rest.
• Balance values are explicitly provisional. At the April Paladin showcase the developers said the class's abilities, values and visuals were all still being tested, and demonstrated talents that were bugged in the build (Paladin balance iteration and combat rules, 2026-04-12). The Mage tuning shown that month was described as placeholder, not target balance (Talent trees, build variety, and combat feel, 2026-04-05).

Ranking classes against placeholder numbers, in a build its own developers called unfinished, produces a document that expires the first time a value changes.

The test left two kinds of record: aggregate numbers the studio published, and eleven days of player reports.

The published numbers. The official recap counted participants from 165 countries, roughly 2.5 million NPCs killed, more than 100,000 PvP kills, and a Lord of Shadows world boss defeated repeatedly at an average interval of about two minutes thirty-one seconds. Mage was named the most popular of the four available classes, with Paladin close behind; the Sacred Order was the most-selected faction and Human the most-played race (Technical test results and player behavior, 2026-06-02).

Those are headcounts, not power measurements. A June thread noted the Mage figure was distorted because anyone who wanted to play Mage had to pick Human (2026-06-11 digest). Popularity is a fact about players, not about class strength, and tier lists built from participation data routinely confuse the two.

The recurring criticism was about systems, not classes. The studio received more than 12,000 feedback reports covering, in its own words, "everything from combat and movement to animations, UI, controls, talents, and overall gameplay", and committed to making combat "more direct, more exciting, more precise, more responsive" (Beast Burst, 2026-07-18). Mein-MMO's account was blunter, listing "minimal hit feedback, skill trees filled with passive bonuses without noticeable changes" among the reasons players came away disappointed (Mein-MMO, 2026-07-13).

That is the most important result of the alpha, and it is not a ranking: combat did not feel finished, and talents did not feel impactful. Both are verdicts on systems every class shares. Neither separates one class from another.

Everything loud about a specific class traced back to a bug. Players watched a Paladin's health climb from 40k to 240k in open-world PvP, caused by a crit interaction that stacked when it plainly should not have. Druid drew the heaviest "overpowered" complaints — first for attack range that appeared to ignore line of sight, which chat diagnosed as the attack not behaving like a proper projectile, and later for AoE farming that outpaced questing. Ranger was accused of one-shots and absurd range on some days, and called weak, bugged or half-implemented by its own players on others. Mid-test, Paladin players reported that large animation-driven abilities stalled so badly under latency that the class felt half-disabled. Underneath all of it: cooldowns not displaying, talents needing a full reset to work, abilities vanishing from action bars.

A tier list built on those eleven days would rank bugs, not classes — and it would rank a build that has already changed. The Mage was reworked afterwards, and players who tried the revamped version in June called it notably better (2026-06-10 digest).

What can be stated without inventing anything is design intent: the role each class is built for, where it fights from, and how much of it has actually shipped.

Resources are barely documented, which on its own should stop anyone writing a numbers-based ranking. Mana is the only resource named in public material: Mage is described as using health, mana and a separate barrier resource (Mage talents and builds, 2026-04-26), and Paladin talent nodes as modifying mana cost among other things (Paladin talent tree redesign, 2026-04-16). No alternative resource — rage, energy or anything else — has been detailed for any class. Anyone presenting one is guessing.

Each class is also planned around several directions inside one large tree rather than a fixed specialisation: Warrior's pathfinder, vanguard and berserker; Assassin's bounty hunter, spellbane and blade dancer; Ranger's sharpshooter, trapper and artificer (Class roster and archetype concepts, 2026-02-15). Trees shown so far have run from roughly 150 nodes on the redesigned Paladin toward 300 on earlier versions of other classes. That is a further reason one letter grade per class would mislead even after launch: the unit of balance here is the build, not the class.

Five conditions, all checkable from outside the studio. This page will be updated against them.

1. All ten talent trees ship. Six classes have no tree and no playable build, so any ranking today covers 40% of the roster and guesses the rest.
2. A deliberate balance pass lands. The current values are placeholders by the developers' own description. A ranking needs numbers somebody intends to keep.
3. The combat and talent rework arrives. The studio has committed to reworking how combat feels, and the Mage rework already landed after the alpha. Ranking classes before that work ships means ranking a system being replaced underneath you.
4. A live population plays past the starter zone. The alpha ran in a single region with a target level cap of around 20 (Paladin balance iteration and technical test context, 2026-04-12). Strength at level 20 in the first zone predicts very little about the level cap.
5. Endgame content exists to measure against. Dungeons, raids and rated PvP are what turn opinions into rankings, because they hand every class the same fixed problem. None of that has been tested publicly.

When those land, this page can carry a ranking built on data. Until then it is the record of what was observed, which is the only version of this question anyone can answer honestly today. The next round of class data will come out of FirstWatch, the studio's invite-based iterative testing programme, so it may arrive under different visibility rules than the open Steam alpha did.

To reason about class strength in the meantime, use the tools that show the data rather than a ranking of it: the per-class pages under /classes, the talent planner for the four trees that exist, and the equipment planner for what gear actually does.