About LoL Recommender

Updated for Patch 26.10

What is LoL Recommender?

LoL Recommender is a champion recommendation engine I built to help players discover new League of Legends champions that actually match how they like to play. You can get recommendations using your Riot ID, a champion you enjoy, or gameplay tags.

Recommendations are powered by live ranked data from the Riot API.
Champion similarity is based on gameplay identity, role behavior, and tag overlap.
LoL Recommender supports both League of Legends and Wild Rift champion suggestions.

Why I Created LoL Recommender

Originally, LoL Recommender started as a small project for my friends who kept asking:

“Who else plays like this champion?”
“What should I try next if I like this playstyle?”

Most tools only showed basic stats or meta builds. I wanted something that focused on how champions feel to play so I built a system that combines data with custom gameplay tags to discover meaningful “this actually feels similar” champion recommendations.

Champion Data & Builds

Recommendations focus on ranked Summoner’s Rift data (Plat+ and above).
Each champion page shows build suggestions, runes, Summoner Spells, and skill order tuned for Patch 26.10.
You can filter recommendations by role (Top, Jungle, Mid, Carry, Support) for more accurate matches.
Wild Rift recommendations follow a similar philosophy with WR-specific builds and tags.

How Does the Recommendation System Work?

Under the hood, LoL Recommender uses a custom similarity model that mixes weighted gameplay tags, role awareness, and match data. The goal is to recommend champions that play like your favorites, not just share a lane or damage type.

Weighted Gameplay Tags

Each champion is assigned descriptive gameplay tags (like Roaming, Power Farmer, Late Game Monster).
Tags capture strengths, weaknesses, and overall champion feel.
These tags are weighted and used as a core signal for similarity matching.

Champion Similarity Modeling

Champions are compared using a multi-factor similarity score.
The model focuses on how champions behave in real games, not just raw stats.
This helps reveal champions that share patterns like “roams early and snowballs” or “scales hard into late game.”

Role & Lane Awareness

The system understands common roles and lane patterns for each champion.
Results prioritize practical, on-role suggestions instead of off-meta picks.
You can further refine suggestions with the position filter on champion, multi-champion, and tag pages.

Gameplay Flow & Power Spikes

Power spikes, pacing, and match flow are taken into account.
The goal is to match the experience of playing a champion.

Types of Recommendations

Single Champion Recommendations

You pick a champion you like.
The system compares it against the full roster based on tags, role, and behavior.
Only meaningful matches appear as “similar champions” suggestions.

Multi-Champion (Blended) Recommendations

You can select multiple champions you enjoy playing.
LoL Recommender blends their shared traits into a combined profile.
Results favor champions that overlap across your picks, not just one of them.

Riot ID Recommendations

Your Riot ID match history is analyzed to find repeated patterns in how you play.
The system builds a profile from your most-played champions over recent games.
Recommendations are then tailored to that real in-game playstyle, not just what you say you like.

Match Score Calculation

Each recommended champion gets an internal similarity score.
The score is normalized and displayed on the site as a percentage (0–100).
Higher percentages indicate a closer match to your selected champion(s) or Riot ID profile.