Is Teamfight Manager 2 worth it? (Early Access verdict, v0.4.6)

What's actually in the game right now

Teamfight Manager 2 launched into Steam Early Access on 2026-05-25, developed by Korean indie studio Team Samoyed (Steam app id 3009300). The current patch is v0.4.7 — six patches shipped during the first launch week, through v0.4.6 on 2026-05-29. What you actually get for the EA price today:

The match engine

A real 5v5 MOBA auto-resolver. Top / Jungle / Mid / Bot / Support with role-specific lane passives (Top regen, Jungle execute + out-of-combat MS, Mid XP, Bot gold, Support kill-gold share). Two towers per lane (inner + outer), added in EA to slow snowball. Two new objectives: Serpen (early, permanent +3% AD / AP / armor / MR / MS per stack to the entire team, every 2:00) and Morgard reworked to empower minions rather than dump stats. Every champion has an ultimate that unlocks at level 5. Average match ~9 minutes.

The roster

60 champions, 5 classes (Melee / Ranged / Mage / Support / Assassin), 3 abilities + 1 ult each. 40 of those are returning from TFM1, but with skill reworks for the new 5-role system — name recognition is a hint about archetype, not a stat carry-over.

Tactical depth

Three explicit tactical axes (Objective Finish — Kill vs Fight Priority; Minion Wave Management — Wave vs Join Priority; Closing Out — Stable / Flexible / Aggressive) plus per-player item-category priorities in three slots from AD / AP / Attack Speed / Armor / MR / HP. These two screens are what new players most often skip, and they decide more matches than draft does.

League structure

6 regions × 2 divisions = 120 teams, with full promotion/relegation. Three international tournaments at the season's end: Champions Master Cup (Worlds analogue), Challengers Eastern Cup, and Challengers Western Cup.

Staff & training

Four staff archetypes: Head Coach / Coach / Scout / Analyst. The Coach is delegate-able — ban/pick, tactics, training, negotiations can all be handed off if you don't enjoy a system. Training itself is a weekly Training / Streaming / Rest ratio with Control / Decision / Mental sub-programs, on top of a per-player Stress system and branching event choices that move stress, condition, loyalty, and fan satisfaction.

Transfers

Scouting → contract negotiation → final approval, with a dedicated Scout Dispatch screen (region / age / position / free-agent toggle / max salary). Player retirement happens on a cycle, new players generate automatically, and retired players can return as staff with a probability.

Auto Patch System

The game collects per-season win-rate stats and auto-buffs and nerfs champions. Distinct from developer patches. This is the answer to the “static management sim” failure mode — the meta moves under you.

Modding & multiplayer

Steam Workshop + an official mod API + a Database Editor for custom teams, logos, rosters, and player stats. Real-team mod packs were live within 24 hours of launch — Gen G, T1, Cloud9, Fnatic, G2, KT, NRG, Pain, Loud, Furia, GAM and more. Online league mode, friendly multiplayer matches, and spectating ship at launch, cross-platform.

What's still rough

Steam reviews are Mixed (currently in the low-60% positive range) during launch week. Don't freeze the number — it's moving as patches land and the review count grows. The negative reviews are worth taking seriously, and they cluster around a few specific issues:

Match AI

The loudest complaint. Multiple Steam reviewers describe the AI as “Iron-level rather than pro-level.” Specific watchable examples from the LettucePlate launch stream include Plague Doctor self-ulting, the Pythoness firing her redirect ultimate into empty sky, and Chef (Support-only-tagged) failing to last-hit minions when accidentally slotted into a solo lane. These are not always-bad, but they're visible enough that the “your jungler will occasionally 1v5 the enemy team and you have to live with it” framing is fair.

Manager Training tutorial loop

Early reviews and creators reported being stuck inside the Manager Training tutorial — one widely-shared 0.3-hour review (“DJ Shadow”) flagged the issue right after launch. Patch 0.4.3 fixed the specific “selecting an already banned, picked, or locked champion disrupts the tutorial” path, and patch 0.4.4 added an immediate end option. For a first save, creators still recommend skipping Manager Training and learning directly in career mode.

UI clutter

A recurring negative-review thread is that TFM2 “lost the charm of the original.” The menus are denser than TFM1's; even fans of TFM2 concede the UI feels cluttered today. Worth noting if pixel-art charm was a big draw for you in TFM1.

Translation quirks

English is currently AI-translated, with community Weblate corrections layered on top. Team Samoyed disclosed this on the store page and in their dev blog. You will see missing spaces and the occasional typo (“fan” for “expectations” in one inbox event was widely screenshotted). Improving patch by patch, but visible today.

Best-player carry effect

In lower divisions, a single “smurf-level” player on a roster ends up deciding too many matches — the best player on each team over-decides outcomes in a way that can feel deterministic. Less of an issue at the top division where roster quality evens out.

Platform rough edges

The game wouldn't launch in certain Linux environments at release — fixed in patch 0.4.2 the next day. Patch 0.4.6 also fixed Mac scrolling and missing next-opponent UI in Mac environments.

Patch cadence & roadmap

Six patches shipped during the first launch week, from 0.4.1 through 0.4.6. The latest first-week patch covered transfer-market AI, match engine memory/frame-drop work, player positioning AI, draft screen performance, Facilities direct entry, and Mac UI fixes. That cadence matters for the “buy now or wait” question: this is a responsive Early Access, not a release-and-go-dark one.

From Devlog #30, the published roadmap covers: 40 more champions to reach 100; multiple map variants and new modes (deathmatch, area control, single-lane); locker-room scene depth; performance-based manager dismissals; deeper transfer drama; further match-engine refinements.

Team Samoyed also publishes their own AI improvement numbers rather than hand-waving them. Post-demo AI was reported as ~70% better than the original demo AI; post-playtest EA AI ~57% better than playtest AI. Concrete improvement targets, not vibes.

Who should buy now

• TFM1 returning players. The skill rework is a real caveat, but everything else — staff, transfers, training, modding, league structure — is a clear upgrade.
• Football Manager fans curious about esports and LoL fans curious about management. LettucePlate's framing of “Football Manager × League of Legends” is the most accurate one-liner about what this game is.
• MOBA spectators who'd rather watch a match unfold and influence it than micro a champion themselves.

Who should wait

• Anyone with low tolerance for Early Access AI quirks. The match engine is improving fast, but it's still EA — you'll see the AI do strange things.
• Anyone who avoids AI-translated games on principle. The translation is being polished but it's not native English.
• Anyone who actually wants to play the MOBA. TFM2's core is an auto-resolver; you do not control champions in-fight. That's the genre, not a bug.

Short: TFM1 vs TFM2

TFM1 is finished, around 83% positive on Steam, and was 50% off during TFM2's launch week. TFM2 is materially deeper but mid-development. If you have time for one and want polish, TFM1. If you have time for one and want depth, TFM2. The full per-system breakdown lives at TFM1 vs TFM2 — differences explained.

Price & promo

Steam lists an intro promo of 10% off through 2026-06-01; the actual price is region-localized, so check Steam for your currency. TFM1 ran at 50% off during the same launch window. Team Samoyed's EA estimate is 1–2 years, and Steam's standard refund window (under 2 hours played, within 14 days) applies.

There is only one place to buy it: the official Steam store page (app id 3009300). A “Teamfight Manager 2 key” just means a Steam activation code — there's no standalone launcher and no official third-party key store. You will see grey-market resellers listing it cheaper, but those keys can be region-locked, revoked after a charge-back, or undelivered, with no Steam refund protection. For a sub-$20 Early Access game from a small indie studio, buying on Steam is the safe call and supports continued development. There is no free version and no legitimate free key — treat any “free TFM2 key” offer as a scam.

FAQ