Match Analytics Lab
Independent AI-Powered Video Analysis
All statistics independently derived from publicly available broadcast footage. No league-provided data. No third-party stats providers.
AI-ANALYZED·Proprietary Video Analysis Pipeline
We built a proprietary analysis pipeline that processes pro pickleball match broadcasts and extracts granular shot-level data using advanced AI video understanding. Every statistic on this page is independently derived — no third-party stats providers, no manual data entry, no league-provided feeds.
Data source: Publicly available pro match broadcasts
Matches Analyzed
50
MLP, PPA matches
Rallies Tracked
5,666
Every point logged
Shots Cataloged
48,550
Per-shot classification
Teams Profiled
13
2025-26 Season
Current Coverage
MLP45 matches
MLP New York (6)MLP San Clemente (12)MLP San Clemente (16)MLP Daytona Beach (2)MLP Cup (8)MLP Phoenix (1)
PPA5 matches
Veolia Texas Open (4)US Open 2024 (1)
APPComing soon
Coverage expands as new matches are analyzed — any pro broadcast can be processed
How It Works
Match Ingestion
Full match broadcasts are collected from publicly available sources
Video Segmentation
Long matches are split into manageable segments for analysis
AI Video Analysis
Each segment is processed by our AI model, which identifies every shot, player position, and rally outcome
Data Extraction
Rally boundaries, shot types, court positions, and outcomes are extracted and merged across segments
Storage & Aggregation
Individual shots and rallies are stored in our database, then aggregated into per-match statistics
Approach
Proprietary AI video analysis · Automated ingestion pipeline · Cloud-hosted database · Custom data extraction
Transparency & Methodology
Data SourcePublicly available pro match broadcasts
AnalysisProprietary AI video understanding pipeline
ProcessingAutomated ingestion with retry logic and quality checks
Last UpdatedMarch 2026
CoverageMLP 2025–26 Season — multiple events and divisions
Confidence Note
AI-assisted analysis provides estimates, not ground truth. Shot counts may vary ±5–10% from manual human annotation. Rally boundaries and shot classifications are determined by the AI model's visual interpretation of broadcast footage. We report what the model sees — and note when uncertainty is high. Shot type labels have been normalized across matches for consistency.
Shot Type Distribution
3rd Shot Drive (1,029)2.5%
Source: Pickleball Research match database · 50 matches · 5,666 rallies · 48,550 shots
Court Position Breakdown
65.2%
Kitchen (NVZ)
26,685 shots
10.7%
Transition Zone
4,398 shots
Source: Pickleball Research match database · 50 matches · 5,666 rallies · 48,550 shots · Per-shot position classification
Shot Outcomes
Source: Pickleball Research match database · 50 matches · 5,666 rallies · 48,550 shots · Per-shot outcome classification
How Rallies End
Error (unforced + forced) (2,891)60.9%
Winner (outright) (1,789)37.7%
Kitchen Violation / Other (70)1.4%
Source: Pickleball Research match database · 50 matches · 5,666 rallies · 48,550 shots
Dinks account for over half (52.8%) of all shots in our dataset, confirming that soft-game proficiency is the foundation of elite pro doubles. Only 4.3% of shots are outright winners — the game is won through patience and forcing opponent errors. Two-thirds of all shot activity occurs at the kitchen line.
Derived from Pickleball Research match database · 50 matches · 5,666 rallies · 48,550 shots
Rally Termination Analysis
Errors end the rally60.9% of rallies
2,891 rallies ended on unforced or forced errors
Winners end the rally37.7% of rallies
1,789 rallies ended on outright winners
Other (kitchen violation, etc.)1.4% of rallies
70 rallies ended on violations or edge cases
Source: Pickleball Research match database · 50 matches · 5,666 rallies · 48,550 shots