Ben DraftDraft Engine

Ben Draft Engine: Chess Engine for League of Legends

The first drafting tool that brings chess engine algorithms to League of Legends draft strategy.

Ben Draft Engine uses minimax search algorithms to think multiple picks ahead in champion select. No more guessing. Get the best draft recommendations that consider team synergy, counter-matchups, and future possibilities.

🖥️ Desktop app includes auto pick locking

Chess Programming

Deep Strategy

Thinks 4+ moves ahead using chess engine algorithms. Plans entire draft sequences, not just the next pick.

Real-time Analysis

Instant recommendations that adapt as the draft evolves. Considers team synergy and matchup advantages dynamically.

Auto Integration

Desktop app locks picks automatically and shows overlay recommendations during ranked games.

Neural Networks

NNUE evaluation trained on millions of games. Understands complex positional factors.

Chess Engine Technology

Ben Draft Engine brings chess engine algorithms and technical innovations to League draft strategy

🧠 Minimax + Principal Variation Search

Uses minimax search with PVS - the same advanced pruning technique in Stockfish. Null-window searches dramatically reduce computation time while maintaining accuracy.

🔄 Iterative Deepening + Alpha-Beta

Progressive search from 1-ply to 6+ plies deep with aggressive pruning. Early results guide deeper analysis while eliminating worse move branches.

⚡ 16-bit Move Encoding

Instead of using typical strings maps or objects, each move is encoded in Custom compact representation inside a uint16. Allows exploring millions of draft positions without memory bloat. For example ”Ahri Mid Blue” becomes decimal 84.

🎯 Neural Network Evaluation

NNUE-style position evaluation trained on millions of pro games. Understands synergies and matchups better than any tier list.

⚙️ Custom Bitmask Engine

Lightning-fast bitwise operations for champion availability and role assignments. C++23 optimizations deliver chess engine-level performance.

🔬 League Specific Optimization

Evaluation function at each node analyzed takes into account the current team comp synergy and opponnent matchups

“If Stockfish has mastered chess, why not apply the same algorithms to League of Legends draft?”

- Me

🎯 The Quest to Solve Draft

My ultimate goal: solve League of Legends draft completely. Calculate the perfect first pick by simulating the entire 10-pick sequence.

🧮 The Scale Challenge

Chess (average position):~35 moves
Chess (max theoritical position):=218 moves
LoL Draft (worst case):≤1,710 moves
171 champions × 5 roles × 2 teams = 1,710 theoretical moves per position

⏱️ Current Performance

Depth 4 (default setting):~15 seconds
Depth 6 (max allowed):~2/3 hours
Depth 8 (max achieved):=3.59 days
Depth 9:~4 months
Full 10-pick solve:~9 years

My deepest Calculation so far (Patch 15.6)

Depth 8 analysis - here's what the engine calculated as the best first pick champions:

Bard
Support
Corki
ADC
Qiyana
Mid
Nami
Support
Lucian
ADC
💭

Theoritical future plan

In the future I want to experiment even more with neural networks. Currently the engine is built like stockfish, brute forcing positions with a neural net at the eval function. I want to redo this project using a monte carlo tree search with a neural net trained to predict win rates of draft positions, similair to Leela Chess Zero.

Why This Beats Traditional Draft Tools

🎯 vs Tier Lists

Tier lists show static win rates vs a single champion. Ben Draft calculates dynamic strategy based on your specific draft situation and opponent responses.

🧠 vs Simple AI Tools

Other tools use basic heuristics. Ben Draft uses proven chess engine algorithms with mathematical optimization and neural network evaluation.

⚡ vs Manual Analysis

Humans can't calculate millions of positions in seconds. The engine explores draft trees faster and deeper than any human could.

Common Questions

Why are scores negative and positive?

Blue team is attempting to minimize, whereas red team is attempting to maximize. A negative score means blue team is favored to win. This is similair to chess engines where a positive score means white is favored and negative means black is favored.

How is this different from tier lists like op.gg or u.gg?

Tier lists show static win rates. Ben Draft Engine uses chess engine algorithms to calculate optimal strategy for your specific draft situation, thinking 4+ moves ahead like Stockfish.

What makes this technically different from other AI tools?

We use minimax search with Principal Variation Search and iterative deepening - proven chess engine algorithms. Other tools use simple heuristics or basic neural networks without strategic depth.

Do I need to download anything?

The web version works immediately. The desktop app adds auto pick locking for seamless integration with your ranked games.

Is this allowed by Riot Games?

Yes! This tool only provides strategic recommendations during champion select - it doesn't automate gameplay or provide unfair advantages during matches.

Higher depth = better recommendations?

Yes - higher depth means the engine thinks more moves ahead, leading to more strategic picks. However, higher depth also increases computation time exponentially. Depth 4 is a good balance of speed and strategy for most users. Depth 4 is usually under 20 seconds, but depth 6 can take hours.

Sometimes it takes a longer than usual to calculate - why?

The engine is running on my local hardware. It's too expensive to run a large server for free users. If lots of users are requesting analysis at once you'll be placed into a queue and only start when it's your turn. This allows fair access for everyone.

Is this on up to date patch data?

Yes, I update the champion data and train the neural network every patch. It weights the 10 latest patches of data to avoid overfitting and provide stable recommendations. It may take a couple days after a new patch to retrain and redeploy.

Can I feature request or suggest improvements?

You can contect me on github: cowleyb or linkedin: Ben Cowley. My league account is Wubs#euw

Start using the best Drafting Tool?

Join me and my friend group using chess engine algorithms to win champ select and climb the ranked ladder.