> ./projects --select "MTG Data"
Data Analytics applied to Magic: The Gathering
This line of work focuses on the use of data analysis and statistical modeling to study the Commander (EDH) format, with particular emphasis on identifying genuine card synergies, archetypal structure, and the effective relevance of cards within the metagame.
Rather than reproducing existing tools, the goal is to correct and formalize prior approaches that lacked well-defined metrics, transforming them into analytical instruments that are interpretable, reproducible, and methodologically sound.
> ./pyedhrec --status active
pyEDHREC
Metric-based reformulation of synergy analysis in Commander
The pyEDHREC project existed previously as an EDHREC-inspired implementation, but it suffered from fundamental issues: unstable results, absence of explicit metrics, and limited explanatory power.
The work presented here consists of a complete methodological renewal, introducing:
- Formal Metrics Modeling the probability of a card appearing conditional on a commander ($P(x|c)$).
- Objective Criteria Identifying informative card sets $I(c)$.
- Extremal Measures Using $p_{\max}$ to detect strong strategic dependencies.
[STATUS: Active and Functional]
[CONTRIBUTION: Metric formalization]
> ./clustering --status legacy
Clustering EDH
Unsupervised deck classification: limits and insights
As a natural extension of the previous work, unsupervised clustering techniques were explored to group Commander decks according to structural similarity.
While these methods produced reasonable groupings, the analysis showed that they do not provide significant new insight beyond existing and well-documented approaches. Consequently, the project was discontinued as a tool, but retained as a theoretical and methodological case study.
>> REFERENCE_RECOMMENDATION
For the state-of-the-art in EDH deck clustering, referencing a project that utilizes the full EDHREC dataset:
Lucky Paper's Commander Map ->
> ./outlook --future
This MTG Data Analytics line aims to:
- >> Develop explanatory, metric-driven tools rather than purely descriptive ones.
- >> Produce technical articles on the limitations of popular data-analysis methods in complex games.
- >> Contribute to a more rigorous understanding of the Commander metagame.
[CURRENTLY ANALYZING]:
We are currently applying these tools to analyze the CEDH (Competitive EDH)
metagame.
