LoLTheory is a desktop app for League of Legends that helps you make better decisions from champion select through the end of the game. It recommends which champion to pick, what to build, and which items to buy next — adapting in real time as the game develops.
I built LoLTheory as an ex-Master player with a data science and economics background because most tools recommend a static build regardless of what's happening in your game.
Champion Select — Who to Pick and Who to Ban
Win rates and scores for every champion, updating live as both teams pick. Score = win rate minus risk, where risk is the quantifiable chance a pick could hurt your team. If your mid laner first picks Xerath, the system can help you think about how to cover him.
Ban recommendations factor in your teammates' current hovers — accounting for both how well the enemy champion performs against your team and how likely the enemy is to actually pick it.
Shareable link: Generate a link from the app showing the best picks for your teammates. Share it in chat so they get real-time recommendations too.
Flex pick detection: The system factors in the chance enemies don't go to their expected lane, adjusting both win rate and score calculations. Mouse over enemy roles to see lane probability estimates.
Three recommendation modes handle player experience differently:
- Personalized — uses your ranked game history from the last 10 patches. Best for maximizing your next game's win rate.
- Fixed — assumes 25 games on every champion. Best for expanding your pool or reading the meta without personal bias.
- Classic — works from raw win rates without experience adjustments. The most responsive to meta shifts.
Enemy threat highlighting shows which picks pose the most risk to your team.
For the full breakdown on how these modes work, see Recommendation Modes: Personalized, Fixed, and Classic.
Build Recommendations — Your Whole Setup, Optimized Together
Runes, items, summoner spells, and skill order — recommended as a single package, not independently. This matters because rune choices affect which items are optimal and vice versa. Most tools pick runes from one data source and items from another.
Builds adapt to both team compositions. Your build against a full AD team looks different from your build against mixed damage.
Lock a first item or keystone rune and the rest re-optimizes around your constraint — useful when you have a specific playstyle or strategy in mind.
The enemy damage breakdown bar shows the physical/magic/true damage split from the enemy team, helping guide your defensive itemization.
In-Game Item Overlay — Builds That Change as the Game Does
An in-game overlay that updates your item recommendations every time an enemy buys an item. Top 3 recommendations, each labeled with why: "Good Against → Magic Damage," "Generally Good."
The model corrects for known biases in item win rate data — snowball items, defensive items, and player experience effects all distort raw win rates. There's a deep dive on how this works and why it matters.
The system currently analyzes 5 items + boots per build.
Free: 1 game analyzed per day, stackable up to 5. Premium: unlimited and ad-free.
How It's Different
Adapts during the game. Most tools give you a build at game start and stop there. LoLTheory's overlay updates as enemies buy items.
Corrects the data. The model adjusts for biases that make raw item win rates misleading — so recommendations reflect actual item effectiveness, not distorted stats.
Optimized as a package. Runes, items, summoners, and skill order are calculated together for your matchup, not pulled from separate "highest win rate" lists.
Try it in your next game. Champion select features are fully free with no limits. The in-game overlay includes a free tier. Download LoLTheory.
