The best pick isn't the best champion.
It's the best champion for you.

PocketPick is a personal draft assistant for League of Legends. During champion select it estimates your expected win rate for every champion you play — from your ranked history, the live matchup situation, and your team's needs — and shows its working for every number.

Join the closed beta

One question, answered honestly

Every stats site can tell you a champion's global win rate. That number is about a million other players. PocketPick answers the question that actually matters at pick time: “Which of MY champions gives ME the best chance of winning THIS game?” — and it never hides how it got there. Every recommendation expands into a plain-language breakdown you can interrogate.

Darius
54.1%
Ornn
52.6%
Fiora
49.9%

Darius 54.1% = your record 52.3% with him this season +1.6% lane matchup vs their likely top laners +0.2% team fit — tap any part to see the full working. (Illustrative view of the live product.)

Product screenshots go here — replace with captures of the live app before submitting the site.

How it works

You + meta

Your own ranked record, weighted honestly

Your per-champion, per-role solo-queue history (read via Riot's Match-V5 API) blended with rank-appropriate meta data using an empirically calibrated Bayesian weight — small samples lean on the meta, big samples lean on you.

Lane

The live matchup, including counter-pick risk

Real matchup data against the opponent you're likely to face — revealed or predicted — including the risk of being counter-picked if you commit early, at your rank, on the current patch.

Comp

What your team actually needs

Frontline, damage balance, engage, peel — measured from real data, priced as an effect on your expected win rate, and explained in plain words rather than a mystery score.

Built to respect the game — and your data

Status

PocketPick is in closed beta ahead of a public release. Core recommendations will remain free; a paid tier will later unlock the deeper personalised analysis built on our own models and computation. Want in early? Email us.