📊 World Heavyweight Championship Bouts (n = 179): Outcome Mix

Source (public PDF): https://pdfs.semanticscholar.org/7453/303ac36e1c6903e8d514248e5185c9b5bca7.pdf

🧭 Stance Distribution (Top 340 Professional Boxers)

Source (public PDF): https://czasopisma.uni.lodz.pl/ar/article/download/11285/10924

🏆 Average Win Ratio (Same Dataset)

Source (public PDF): https://czasopisma.uni.lodz.pl/ar/article/download/11285/10924

🧭 Stance Distribution (MMA sample)

Source (public PDF): https://pdfs.semanticscholar.org/e704/277c8b2f79b2fd1b0a7821f49b095d359b7e.pdf

🏆 Win Percentage (Same MMA source)

Source (public PDF): https://pdfs.semanticscholar.org/e704/277c8b2f79b2fd1b0a7821f49b095d359b7e.pdf

🔺 A Simple Two-Sport “Top-Level” Pyramid (Verifiable points only)

🏟️ The Core Problem: “Home” Is Not a Standard Field

Most combat sports records do not encode a clean “home vs away” designation. You can approximate it, but every approximation has tradeoffs:

• Country-of-event vs fighter nationality: requires nationality data that’s not consistently structured.
• City-of-event vs fighter training base: requires private gym/camp data.
• Judging bias vs crowd: you need scorecards + close-fight labeling.

🧤 Why This Isn’t Publicly Verifiable at Aggregate Level

• Boxing: most public boxing datasets don’t tag “KO caused by left hook vs right cross” in a structured way.
• MMA: even when finish is listed (KO/TKO), the exact limb/strike that ended it often lives in narrative text or video review, not a universal field.

✅ Stance (verifiable, from sources above)

• Boxing (top 340 sample): win ratios are basically similar (88% vs 87%).
• MMA (sample): southpaw shows a small edge (64.0% vs 62.6%).

⚠️ Style and combinations (public-data gap)

“Most winning style” and “best combinations” require a dataset that tags:

• style (pressure boxer, out-boxer, counterpuncher; or in MMA: wrestling-heavy, BJJ-heavy, striker, etc.)
• combination sequences (e.g., 1–2–3, jab→cross→low-kick)
• outcomes attributable to those sequences

Those labels are typically subjective and/or proprietary unless the analysis is done in a controlled academic study.