Methodology — Research and Fact-Checking Process

Research Approach

RIICHI~SENSEI was developed with a commitment to factual accuracy and transparency about sources and certainty levels. The research process involved the following steps:

1. Japanese-Source Priority

Japanese-language sources were treated as primary authorities for all rule mechanics, scoring details, and strategic concepts. This reflects the fact that riichi mahjong originates in Japan, and the most extensive and authoritative body of knowledge about the game exists in Japanese. English-language sources were used as supplementary references and for verifying the accessibility of explanations.

2. Cross-Checking

Important factual claims—particularly scoring values, fu calculations, yaku conditions, and rule mechanics—were verified against multiple independent sources. Where sources disagreed, the disagreement was noted and the most commonly accepted interpretation was presented as the default, with alternatives flagged as ruleset variations.

3. Ruleset Sensitivity Identification

Every topic was evaluated for ruleset sensitivity. Rules that vary across platforms, tournaments, or regions were explicitly flagged with the ⚙ RULESET VARIANT marker. The default assumptions used throughout the academy are:

4. Strategy Debate Transparency

Where strategic advice is debated among strong players or analysts, the academy explicitly acknowledges the debate and presents the main viewpoints rather than asserting one position as universally correct. Topics with significant debate include: optimal riichi frequency, the value of flow/nagare concepts, GTO vs. exploit-based play, and specific push/fold thresholds.

5. Certainty Calibration

The academy distinguishes between: Rules (high certainty—verified against official sources), Standard practice (high certainty—widely agreed upon), Strategy heuristics (moderate certainty—generally correct but with exceptions), Debated theory (explicitly uncertain—competing viewpoints exist), and Platform-specific behavior (verified against platform documentation but subject to updates).

6. No Fabricated Citations

Source notes throughout the academy reference categories of sources rather than fabricating specific bibliography entries. When specific statistics are cited (such as average ura dora pickup rates), these are based on documented analyses of large game databases, primarily from Tenhou. We do not invent citations or pretend uncertain claims are certain.

Limitations

This academy is a static educational resource and does not update automatically as rules change or new research emerges. Platform-specific information (Tenhou rankings, Mahjong Soul features) may become outdated. Strategic advice represents the current consensus as understood at the time of creation but may evolve as the metagame develops. Some statistical claims are approximate and based on specific sample populations (primarily Tenhou ranked games).