Guide

Game pick-ban and map veto systems explained

Harbor Arena’s ranked tactical shooter launched with seven maps but matchmaking drew from a weighted pool that favored three layouts with the fastest load times. Telemetry showed 63% of Diamond+ series played on the same trio; one map (“Depot”) alone appeared in 41% of matches. Teams with narrow map mastery climbed by dodging unfamiliar layouts in queue dodges instead of adapting in-game. Player surveys flagged 38% of ranked frustration as “stale maps” or “no counter-pick.” After shipping a best-of-three pick-ban draft with a rotating competitive pool and a decider rule, repeat-map churn fell to 18%, queue dodges dropped 27%, and unique maps per series rose 31%.

Pick-ban and map veto systems are the pre-match draft layer where teams remove and select maps (and sometimes agents, weapons, or rules) before a match format begins. Done well, they reward preparation, reduce layout monotony, and pair with side-balance rules so neither team rides a permanent home-map advantage. This guide covers draft phase FSMs, pool curation, series vs single-map veto, timers and disconnect handling, telemetry, the Harbor Arena refactor, a technique decision table, pitfalls, and a designer checklist.

What pick-ban solves in competitive modes

Random map select is fine for casual playlists. Ranked and esports need structure:

Counter-strategy — teams ban an opponent’s comfort map and pick one that favors their comp or economy plan.
Pool hygiene — broken or overplayed maps rotate out without deleting content from casual queues.
Series narrative — Bo3/Bo5 drafts create tension: each map win shifts what remains on the board.
Balance feedback — ban rates expose maps players perceive as unfair before win-rate dashboards catch up.

Without veto, matchmaking optimizes for queue time and server load, not competitive depth. Players grind one layout, designers cannot tell whether a map is popular or merely default-selected, and side-balance patches land on maps nobody plays in ranked.

Draft phase finite-state machine

Model the veto as an explicit draft FSM with phases, acting team, and remaining map set. Common patterns:

Single-map veto (quick ranked)

Each team bans one map; system randomizes among survivors, or higher rank bans first and lower rank picks side.
Fast (<15 s total); good for Bo1 ladders with large pools.

Bo1 ban-ban-pick (CS-style)

Team A bans
Team B bans
Team A bans
Team B bans
Team A bans
Team B bans
Remaining map plays (or coin flip for side)

Bo3 pick-ban-pick-ban-pick

Team A bans
Team B bans
Team A picks Map 1
Team B picks Map 2
Team A bans
Team B bans
Remaining map is Map 3 decider

Persist draft_id, phase index, actor team, timer deadline, and chosen maps in server state so reconnecting captains resume cleanly. Never trust client-selected maps without server confirmation at each phase transition.

Map pools, rotation and decider rules

A competitive pool is the curated subset eligible for ranked veto. Rotate maps seasonally: add one, remove the lowest ban-rate outlier, keep at least seven for Bo3 variety.

Active pool flag — maps in pool vs casual-only; UI shows rotation end date.
Decider policy — when two maps remain after bans, either last map auto-plays or teams pick sides on a neutral decider with overtime rules from overtime systems.
Remake threshold — if a map is disabled mid-season, grandfather in-progress series with a substitute pick phase.
New map grace — week-one maps may be pick-only, not bannable, so data accumulates.

Track pick rate, ban rate, and win rate when picked separately. High ban + high win rate signals oppressive layout; high pick + low win rate signals comfort picking despite poor results.

Agent, weapon and mode sub-bans

Tactical shooters often layer agent veto after map selection: each team bans one operator before the round-based buy phase begins. Rules to document:

Whether bans are per-map or per-series
Duplicate bans (both teams ban the same agent → agent still unavailable)
Mirror bans in casual vs sequential bans in ranked
Weapon restriction modes (snipers banned on certain maps)

Keep sub-bans optional. Map veto is already cognitive load; add agent bans only when character count exceeds counter-pick depth and telemetry shows stale meta picks.

Timers, captains and disconnect policy

Draft stalls kill ranked throughput. Standardize:

Captain role — party leader or highest MMR; transferable with 5 s confirmation.
Phase timer — 20–30 s per ban/pick; 5 s warning ping; auto-random among legal options on expiry (never forfeit the series).
Disconnect mid-draft — pause once per team for 60 s; then auto-pick lowest ban-rate map for that team’s turn.
Spectator visibility — esports overlays show remaining maps and next actor; ranked UI minimizes leaks to opponents in lobby.

Log auto-picks separately. If auto-pick rate exceeds 8% for a rank band, timers may be too aggressive or UI too confusing.

Pairing veto with match format and servers

After draft completes, hand off to matchmaking or dedicated server allocation with the locked map ID. The veto service should emit a series_manifest JSON: maps ordered, starting sides per map, agent bans, and tiebreak rules if the series draws 1–1 in Bo3.

For bomb and zone modes, run side selection after map pick: higher seed picks side on Map 1, loser picks on Map 2, coin flip or higher seed on decider. Document interaction with halftime swaps so designers do not double-apply conflicting rules.

Harbor Arena refactor walkthrough

Harbor shipped ranked Bo3 with a seven-map active pool and this draft:

Higher MMR team bans one map
Lower MMR team bans one map
Higher picks Map 1 and starting side
Lower picks Map 2 and starting side
Each team bans one more map
Remaining map is decider; loser of Map 1 ban phase picks side

They added a draft lobby separate from matchmaking queue, 25 s timers with audio cues, and a post-draft agent ban (one per team per map). Seasonal rotation removed Depot from the pool after three patches failed to fix defender bias; ban rate had hit 67%. Outcomes: repeat top-three map rate 63% → 18%, queue dodge rate 14% → 10%, average maps played per series 1.8 → 2.6, and esports viewership on decider maps rose 19% (longer series tension). Designers gained a weekly ban-rate report that flagged one new map as ban-dominant two weeks before win-rate skew appeared.

Technique decision table

Scenario	Prefer	Avoid
Bo1 ranked with 10+ maps	Double ban then random	Full Bo3 draft every match
Bo3 esports series	Pick-ban-pick-ban-decider FSM	All random maps
Small map pool (4–5)	Single ban per team	Six ban phases leaving one map
New map launch	Pick-only grace week	Immediate ban-heavy pool
Captain disconnect	Auto-pick legal option	Series forfeit on draft timeout
Quick-play casual	Weighted random, no draft	Ranked veto rules in casual

Common pitfalls

Client-authoritative map choice — exploiters load unwinnable layouts; server must own draft state.
Pool too small for ban count — six bans on five maps deadlocks the FSM; validate at design time.
Hidden maps in draft UI — players ban wrong layout; show thumbnails, mode tags, and recent patch notes.
No auto-pick fallback — one AFK captain holds 10 players hostage.
Ignoring ban rate telemetry — 70% ban rate is a design ticket, not “players are wrong.”
Mismatched series and draft — Bo5 UI with Bo1 backend leaves two phantom maps.
Agent ban without map context — banning on Map 1 but forgetting to reset for Map 2 confuses teams.

Production checklist

Draft FSM documented with phase order for each match type.
Server authoritative; clients display pending selection only.
Competitive pool flagged; rotation schedule published in UI.
Per-phase timers with auto-pick on expiry.
Captain role assigned; transfer flow tested.
Disconnect pause once per team; logged auto-picks.
Series manifest handed to server allocator after draft.
Side-pick rules consistent with halftime swap policy.
Telemetry: pick rate, ban rate, win rate, auto-pick rate.
Esports spectator overlay shows remaining maps.
Seasonal rotation playbook for add/remove maps.

Key takeaways

Pick-ban turns map selection into a strategic layer — not just matchmaking convenience.
Model drafts as explicit FSMs with server-owned state and auto-pick fallbacks.
Curate and rotate pools — ban rates tell you when to patch or retire maps.
Pair veto with series format and side rules so Bo3 flows cleanly into gameplay.
Harbor Arena cut repeat-map churn from 63% to 18% with Bo3 pick-ban and seasonal rotation, not more random weights.