Skip to main content

RoboCup Soccer Simulation (2D and 3D) — Competing Without a Single Screw

RoboCup Soccer Simulation exists for anyone who wants to compete at a world event without buying a single component. There is no robot here — only code, a simulation server and eleven autonomous agents.

Two different versions

2D Simulation

The pitch is viewed from above in two dimensions, with body physics abstracted away. Each player is an independent program receiving limited, noisy information from the server — meaning no player ever sees the full field. That deliberate limitation makes 2D one of the best research platforms for multi-agent systems, focused entirely on tactics, communication, prediction and decisions under uncertainty.

3D Simulation

Here real physics enters: a humanoid model in a 3D world with gravity and friction. On top of tactics, you must program the robot's gait. 3D bridges the software world and the real challenges of the Humanoid League.

Why this league matters in Iran: sanctions, part costs and hardware import problems simply do not apply. Iranian teams have a strong international record in simulation, and it remains the most accessible route to a world stage.

Skills it builds

  • Professional programming — usually C++ or Python at real project scale.
  • Machine and reinforcement learning — many top teams train player behaviour rather than hand-coding it.
  • Decisions under uncertainty — working with incomplete, noisy data.
  • Software teamwork — repository management, module ownership, automated testing.

These are precisely the skills the software and AI industries value, which is why graduates of this league often move straight into work.

How to start

  1. Learn the language: Python or C++ to a solid level.
  2. Run a base team: most teams start from an open-source base and build on it.
  3. Small changes, measured: alter one behaviour and evaluate it over hundreds of automated matches.
  4. Enter domestic eventsIranOpen runs these leagues annually.

For students not ready for that level, CoSpace teaches the same logic at a simpler scale.

Frequently asked questions

What do you need to enter?

A computer and programming skill. No robotics hardware at all, which makes it the cheapest route to world-level competition.

Which language?

Most top teams use C++; Python is common for getting started and for machine-learning work.

How do 2D and 3D differ?

In 2D, body physics is abstracted and the focus is tactics and decision-making; in 3D you must also program a humanoid gait.

Want to reach a world competition with code alone?

Novin Zehn competition teams train from zero to the world stage — in Tehran and live online.

Register / free consultation AI & machine learning course

Related articles