Hiroshi Harada — May 2, 2026
This repository contains a collection of Python scripts for simulating and visualizing the Shiftless 3-adic Collatz (Collatz-3) model. This is an information-preserving model that eliminates the "information loss caused by right shifts" inherent in the standard 3-adic Collatz map.
The Shiftless model aims to visualize the following:
- The geometric structure of carry avalanches
- The interference moiré between Term A (pure expansion) and Term B (complement injection)
- Fractal structures (gaskets) unique to 3-adic spaces
- TITLE_EN (English title page for the research report)
- REPORT_JP / REPORT_EN
Research reports detailing the theoretical background, the linear decomposition of the Shiftless Collatz-3 trajectory, and in-depth explanations for each Figure.
Executing each script will automatically generate the corresponding Figure (PNG format).
Verifies equivalence with the standard 3-adic Collatz model.
Extracts the non‑multiple‑of‑3 part (uk) of the trajectory to demonstrate perfect alignment with the standard model.
Outputs: figure1.png, figure1.csv
Visualizes the compact trajectory of the Shiftless Collatz-3 model.
Captures the intense carry avalanches triggered by NZT=2.
Output: figure2.png
Visualizes the full trajectory by decomposing a single step of the Shiftless model into two distinct phases:
- Reach (Left shift + self-superposition in ternary notation)
- Fill (Local distortion via NZT injection)
Output: figure3.png
A time-synchronized comparison of three distinct dynamics:
- Pure Expansion: 4n
- Fixed Addition: 4n + 1
- Shiftless: 4n + ck * 3vk
Visualizes the structural differences side-by-side.
Output: figure4.png
Performs a perfect linear decomposition of the trajectory into the following components to visualize their spatial interference (moiré):
- Term A (Pure Expansion)
- Term B (Complement Interference)
Output: figure5.png
- Python 3.x
- NumPy
- Matplotlib
Run each script from your terminal or command prompt:
python code_01.py
python code_02.py
python code_03.py
python code_04.py
python code_05.py- Research Document: CC BY 4.0
- Python Source Code: MIT License
- © 2026 Hiroshi Harada