A draft is genuinely losing, not just awkward, when no version of the game plan works — the practical test is whether you can even describe a late-game fight your team wins; if you can, however contrived, there's a win condition, and it's probably still salvageable through better decisions. If you can't, it usually isn't.

Comp archetypes and their win condition

Most comps fall loosely into a handful of shapes, each with its own idea of how the game gets won:

  • Poke — win by chipping health from range before a fight starts, so the actual fight is already decided.
  • Dive — win by committing hard onto a priority target and forcing a trade the enemy didn't want.
  • Teamfight — win by grouping and out-damaging the enemy in a straight five-on-five.
  • Split-push — win by pressuring a side lane and forcing the enemy to either concede map or fight undermanned.
  • Pick — win by isolating and killing a single enemy champion to create a numbers advantage elsewhere.

None of these is strictly better than the others. Each one is a plan for how the team turns individual picks into game state, and each has a matching set of things that go wrong for it.

No win condition versus an awkward one

A comp with an awkward win condition still has one — it's just harder to execute than it needed to be. A poke comp with one melee pick who has to walk through poke to participate is awkward, not hopeless; that one player plays around it by picking their engages carefully instead of joining every skirmish. A comp with no win condition at all has no version of the game plan that works: five immobile poke champions with zero engage and zero peel can't force a fight on favourable terms, can't disengage from an enemy dive comp, and can't out-scale a split-push team taking towers uncontested. The tell is whether you can describe, out loud, a sequence of events in a late-game fight where your team wins it. If you can, however contrived, the draft has a win condition. If you can't, it doesn't.

What's recoverable through play

An awkward comp is usually recoverable through decision-making — vision control, picking the right fights, patience. A comp with a genuine structural gap — no engage against a team that can freely disengage, no way to peel your one damage dealer against a team built entirely to dive them — isn't something good positioning fixes on its own; it needs the enemy to make a mistake, or it needs your team to win hard enough early that the late-game shape never gets tested. Knowing which situation you're in changes what "good play" even means for the rest of the game: patient scaling makes sense with an awkward comp, and forcing an early lead makes more sense with a structurally weak one, because you may not get to see the late game at all. See reading champion select for the specific things — frontline, engage, disengage — that tell you which of these situations you're actually in.

Most real comps are hybrids

Pure archetypes are rare — most drafts end up as some blend, a poke comp with one dive threat, a teamfight comp with a side-lane win condition as a backup plan. That's not a problem by itself; a backup plan is a strength. The problem is when a comp has two win conditions that actively work against each other. A slow-scaling poke comp with one hyper-aggressive early-game pick can end up neither poking effectively, because the aggressive pick keeps forcing early fights, nor snowballing effectively, because the rest of the comp isn't built to capitalise on an early lead. Recognising when a comp's pieces are pulling in different directions, rather than reinforcing the same plan, is a sharper read than sorting picks into archetypes alone.

The enemy has a win condition too

It's easy to only evaluate your own comp and forget the enemy's is subject to the same read. A losing draft is often less about your own comp's weakness in isolation and more about how it matches up against what the enemy is trying to do — a solid teamfight comp can still lose to a split-push comp it has no way to punish for taking towers. Read both sides' win conditions, not just your own.

What to watch for once the game starts

A draft read isn't a verdict, it's a starting hypothesis to test in the first several minutes. Whether a lane is actually as bad as the pick suggested, whether the jungler is pathing to support the comp's win condition, whether the enemy is actually playing around their own comp's strengths — all of that either confirms the pre-game read or complicates it. Treat the draft read as the plan you walk in with, not a fixed prediction you stop updating once minions spawn.

Where this fits with a win estimate

Should I Dodge's number is a rough signal from aggregate win rates, not a read on comp shape — it has no idea whether your team's win condition exists or is missing entirely. A comp of five individually strong champions with no coherent plan can post a perfectly reasonable estimate and still be a bad draft — partly because no single champion's own win rate can move that estimate very far on its own, which is exactly how five strong picks with nothing tying them together can still average out to an unremarkable number; see why a 43% champion isn't scored at 43%. Use the estimate alongside this kind of read, not instead of it. See how it's built for the full picture of what goes into the number.