Teamfight Manager 2 Guide: every system from draft to worlds
Updated 2026-05-28 · based on v0.4.4 · current patch v0.4.7 · 20 min read
On this page
Quick Answer
Teamfight Manager 2 is a manager sim with an auto-battling MOBA inside it. The match plays itself; you decide the roster, the draft, the tactics, the items, the training, and the transfers. Every system in this guide links to a deeper page if you want to drill in — start here for the full map.
Go deeper
Every system below has a dedicated page — jump straight to the deep dive, or read the orientation first.
What TFM2 actually is
Teamfight Manager 2 is the sequel to Team Samoyed's esports manager sim. It launched in Steam Early Access on 2026-05-25, with the 1.0 target roughly one to two years out. Two things matter for orientation:
- You are not piloting the MOBA. The match engine auto-resolves. Your leverage sits around the match — draft, tactics presets, per-player items, training, scouting, contracts.
- The meta moves under you. The Auto Patch System adjusts champion balance between seasons based on actual win-rate data inside your save.
If you're still on the fence about buying it, the EA-verdict guide breaks down what shipped, the open complaints, and who should wait. If you played TFM1 and want to know what changed, the TFM1 vs TFM2 comparison has the full diff.
The match
A single match is 5v5, five roles, roughly nine minutes. Destroy the enemy Nexus to win. The systems that define how a fight unfolds:
- Two towers per lane (outer + inner). Inner towers were added in EA so a single tower loss doesn't instantly snowball the map.
- Role passives. Top regenerates HP out of combat; Jungle gets out-of-combat move speed and a smite-style execute on epic monsters at low HP; Mid gets bonus XP; Bot gets +20% gold and −30% XP; Support shares last-hit gold with the nearest ally.
- Serpen (early epic, bottom side). Each kill grants the entire team a small permanent stat buff (AD / AP / armor / MR / movement speed) per stack. Stacks every two minutes — the math compounds.
- Morgard (late epic). No longer a stat buff — it empowers minions for about a minute. It's a siege objective, not a stat dump.
- Ultimates unlock at level 5. Champions feel generic before that — don't judge a draft on the first two minutes.
- Match flow is fixed: the first 3 minutes are laning, then the operations phase begins about a minute before each Morgard spawn, and the pattern repeats. Your tactics decide what your team does inside each phase.
The full mechanics-and-mistakes breakdown lives in the beginner's guide.
Champions
60 champions at EA launch, five classes (Melee / Ranged / Mage / Support / Assassin), three abilities + 1 ult each. 40 returned from TFM1 with skill reworks — same names, different math, so don't import your old assumptions.
Position tags matter. Some champions are flexible by design; others (especially Support-tagged ones) have weak last-hit behaviour in solo lanes and shouldn't be slotted out of role. The role tier list is the fastest read on which champions stand out per role this patch; the full database lives at /champions.
Draft and Ban-Pick
Draft is standard MOBA-style bans-then-picks with role assignment. The wrappers around it are where the strategy lives:
- Fearless Ban/Pick (new-game option) — a champion used in one game of a series can't be re-picked in the rest.
- Hide Attributes (new-game option) — gradual information reveal across the draft, for players who find full-info drafts too cold.
- Ban/Pick AI difficulty — separately configurable for opponent teams. Easy vs Normal nets about a 4:6 win rate from draft skill alone; Easy vs Hard is roughly 2:8.
- Personal Tier List overlay during draft. If you have an Analyst on staff, delegate tier classification to them and tiers auto-populate based on their judgment stat.
- Side selection on loss — the losing team picks side for the next set.
- Pre-Match Analyst Report — the Analyst writes up the opponent's key player, their lane priority, dive habits, snowball vs late-game tendencies, target-ban candidates, and conditional win-rate stats (e.g. “100% when securing the first Serpen, 0% without”). It is the single most-underused screen in the game.
Patch 0.4.7 also added presentation effects to the pick/ban screen so it's clearer whose turn it is — previous patches had been criticised for unreadable turn order.
Tactics
The tactics screen is where most new players leave value on the table. The core levers:
| Axis | Options | When it matters |
|---|---|---|
| Operations formation | Group Together / 1-4 Split / 1-3-1 Split | Decides where players sit a minute before each Morgard spawn — defines whether you contest, split, or trade. |
| Objective Finish | Kill Priority vs Fight Priority | Whether your team prioritises securing the last hit on the objective or winning the fight around it. |
| Wave Management | Wave Priority vs Join Priority | Mid-game cohesion. Wave Priority can keep solo laners on creeps when the rest of the team needs them. |
| Closing Out | Stable / Flexible / Aggressive | Late game: do you stall, react, or force. |
| Tower siege | Poke / Dive / Defend pressure lane / Force fight | When you have a tower lead and want to convert it. |
| Per-player item categories | Pick 3 of AD / AP / Attack Speed / Armor / MR / HP, or "Let player decide" | Highest leverage. Wrong items make a strong champion functionally useless even with a gold lead. |
| Per-player stance / target / dive | Aggressive / Balanced / Defensive · Backline / Frontline · Dive on/off | Assassins need aggressive + backline + dive; control mages and CC supports need defensive + protect backline. |
| Fight distance | Narrow (death-ball) vs Wide (spread) | Dive comps want narrow; AoE / mage comps want wide to avoid group CC. |
The beginner's guide walks each of these in order; this section is the reference.
Roster, attributes and training
Players have a stack of attributes that move under training and age. The ones surfaced in the in-game scout / training menus include positioning, judgment, monster kills (CS), skill hit, skill dodge, control speed, calling, roaming, aggression, focus, mental, and ego, plus per-player stress and a language tag.
- Attribute priority. Behaviour stats (judgment, map awareness) outperform raw mechanics in long matches — train those first, then skill accuracy and skill dodging, then atk/def numbers.
- Stress rises with high training intensity, match pressure, and transfer drama. It capsults performance. Manage with rest periods and the “lift through match sharpness” action.
- Language matters. A Chinese-only or Korean-only player on an English-speaking team underperforms until they train the language.
- Aging. Stats decay in the late 20s. Players in their 30s lose attributes weekly. Rotate veterans onto the bench, train juniors to take over.
- Weekly Training / Streaming / Rest ratio. Streaming generates revenue; rest reduces stress; training moves stats. Default ratios are too aggressive — bump rest in your first hour.
- Retirement and inheritance. Players retire on a cycle and the league generates new ones automatically. Retired players can return as staff with a probability — that's how the league keeps its character across long careers.
Staff
Four staff archetypes, all delegate-able if you don't want to micromanage a particular system:
- Head Coach — overall direction. The manager (you) can self-assign this role on lower-division saves.
- Coach — training and game prep. If you hand tactics to the Coach, they'll draft and configure based on opponent analysis.
- Scout — player discovery via Scout Dispatch with detailed filters (max age, position, per-stat minimums on control, judgment, mental, calling, roaming, aggression, ego).
- Analyst — writes the pre-match report and applies tier classifications during draft.
Top-tier league saves start with all four; lower-division saves start lean and you hire the rest from your budget.
Transfers
The transfer window is a discrete phase between seasons. Inside it:
- Free agents are flagged separately and are usually the cheap roster fillers.
- Offer composition — base salary, contract length in years, optional per-win bonus.
- Counter-offers happen — a player can demand more after your first offer.
- Resell clauses when selling players. Common pattern: 50% of any future sale flows back to your club.
- Final-approval gate before a signing commits — you see the full deal before it's locked.
- Trade interest from other teams arrives in the inbox during the window; if a star player has multiple suitors a bidding war is visible in the news ticker.
Season structure
The competitive year:
- Six regions (Korea, China, EU, NA, Japan, South America) × two divisions each, with promotion and relegation. 120 teams total.
- Spring split → playoffs inside each region.
- Three international tournaments at season's end: the Champions Master Cup (Worlds analogue) plus Challengers Eastern Cup and Challengers Western Cup. A region gets roughly three Master Cup slots and six Western Cup slots, though this varies by region strength.
- Transfer windows are scheduled between splits.
Patches and the Auto Patch System
Two patch sources move the meta under you:
- Team Samoyed's real-world patches — at EA launch the team shipped six patches in the first launch week, fixing stability, multiplayer reliability, draft tutorial friction, item prices, transfer-market AI, and Mac UI issues. See all tracked patches.
- In-game Auto Patch System — per-champion win-rate data is collected each season and applied as buffs / nerfs automatically inside your save. Verified via internal simulation before shipping. View patch history under Game Info → Patch History.
Practical impact: a champion that's “always banned” in your save's first season may have been pulled back to A-tier by season three because the AI nerfed it. Don't over-commit to a permanent pocket pick.
Mods and Workshop
Steam Workshop and a Database Editor shipped with EA. Within 24 hours of launch the community filled in real-team gaps (LCK, LPL, LCS, LEC, CBLOL) so if you want T1, Gen G, C9, Fnatic, G2, KT, NRG, Pain, Loud, Furia or GAM rosters, they're a Steam subscribe away.
The five mistakes new managers make
- Ignoring the tactics screen. The Steam “I had no agency” reviews are almost always written by players who never opened the per-player items / formations UI.
- Sharing bot-lane farm. Support next to Attack earns less gold while clearing minions — funnel everything to the Attack carry.
- Treating every champion as flexible. Position tags matter. Support-tagged champions in solo lanes have weak last-hit AI and underperform.
- Over-training. Stress capsults stat growth. Default training presets are too aggressive — give your roster real rest in pre-season.
- Skipping the Analyst report. It tells you the opponent's key player, lane priorities, conditional win-rate stats, and a recommended target ban. Reading it once per match changes drafts.
Full mistake list with examples in the beginner's guide.
FAQ
Advertisement
Related Guides
Teamfight Manager 2 beginner's guide: your first season
You're the manager, not the MOBA pilot. New-game cheat sheet, the tactics screen, per-player item categories, lane passives, Serpen and smite, and what NOT to do.
Is Teamfight Manager 2 worth it? (Early Access verdict)
Honest verdict on TFM2 in Early Access — match-AI complaints, first-week patch cadence, who should buy now vs wait for 1.0. Based on v0.4.7.
Teamfight Manager 2 vs 1: what's actually different
Every verified system change — two towers per lane, Serpen + smite, ultimates at level 5, reworked champions, real staff, transfers, modding, Auto Patch System.
Sources