Building a Map Pool: How to Actually Use Your Map Win Rates

Almost every player has one map dragging their account down, and almost nobody knows which one it is. Map win rate is the most actionable stat on your profile and the least examined.

Why map win rate is the most useful stat you own

Most stats tell you how you played. Map win rate tells you where to spend your practice time, which is a much rarer kind of information.

It is also unusually honest. Damage stats can be inflated by playstyle changes; win rate cannot. If you win 62% on Ascent and 34% on Icebox across a decent sample, that gap is real, and closing it is worth more rating than any amount of aim training.

The three reasons a map win rate is low

They need completely different responses, so diagnose before you act.

1. You do not know the map

Callouts, common angles, default setups, standard smoke placements. This is the fixable one and usually the cause on maps you rarely see. A player who cannot name the parts of a site cannot make a rotate call, cannot understand a teammate's call, and will lose rounds to information problems alone.

Fix: 20 minutes in a custom game with the map open. Walk every route from spawn to both sites. Learn the callouts out loud. That is genuinely most of it.

2. Your agent pool does not fit the map

Some agents fall off hard on certain layouts. A three-site map, a map with a long central sightline, a map with tight vertical routes — each rewards different kits. If your only two agents are both weak on a given map, you will lose there regardless of how well you play them.

Fix: add one agent that specifically covers your weak maps. This is the highest return-on-effort change available to most players.

3. You play a style the map punishes

Aggressive early peeking gets destroyed on open maps with long sightlines. Passive anchoring gets overrun on maps with many close entry routes. The map is not the problem; the mismatch is.

Fix: pick one behavioural change per map. "On this map I do not push out on defence." "On this map I always play the retake rather than the hold." One rule, applied consistently.

Building a pool that survives solo queue

The trap is thinking a map pool means "maps I like." A usable pool means:

The efficient way to build this is not one agent per map — it is picking agents with broad map coverage:

• One controller who works nearly everywhere. Omen and Brimstone travel well across almost any layout.
• One initiator with flexible utility. Sova on open maps, Skye or Fade on tighter ones.
• One duelist you actually entry with. Jett on maps with long angles, Raze on maps with tight chokes.
• One sentinel for the maps with a flank you cannot cover. Cypher and Killjoy both scale across most rotations.

Four agents chosen this way will cover the entire pool at a competent level. Twelve agents played occasionally will cover none of it. See the role benchmarks guide for what each of those roles should be delivering.

The map pool rotates — plan for it

Riot rotates maps in and out of the competitive queue regularly, which means the specific list you are playing today will change. Two practical consequences:

• Do not over-invest in one map. A player whose whole game is built on a single map loses a rank every time the rotation shifts.
• When a map returns, expect chaos. The first couple of weeks after a map rotates back in, nobody remembers the setups. Players who spend one custom game refreshing callouts get a genuine free run of easy games.

Should you dodge your worst map?

Honest answer: rarely, and never as a habit.

Dodging costs rating and queue time, and it guarantees the map stays bad forever. The exception is narrow — if you are one game from a promotion and you genuinely have no agent for the map, a single dodge is defensible. As a routine it is self-defeating, because the rotation will keep giving you that map and you will keep paying to avoid it.

The better play is to treat your worst map as a project for two weeks. Play it deliberately, on one agent, with one behavioural rule. Most players find the win rate moves faster than they expect, because the gap was knowledge, not skill.

Reading attack and defence splits

If your data separates attack and defence rounds, the split tells you more than the overall number. A map where you win defence comfortably and lose attack badly is an execute problem — your team cannot take space, which usually means smokes and flashes are not coordinated. The reverse is a hold problem: setups too passive, or rotations too slow. Each has a different fix, and the overall win rate hides both.

The short version

Get at least ten games per map before you trust the number. Diagnose whether a low win rate is knowledge, agent fit, or playstyle mismatch — they need different fixes. Build two playable agents per map by choosing four agents with broad coverage rather than twelve with narrow coverage. And treat your worst map as the cheapest rating on the table, because it is.

Written by the FragLy team. Benchmarks in this guide are rules of thumb drawn from competitive play, not official Riot figures — use your own tracked matches as the real reference point.