Forensic Behavioral Analysis Engine

CHESSFORENSICS

Deep Game Analysis & Coaching Intelligence

→ Now with personalized AI coaching based on 75 behavioral signals

Find out which SuperGM you play like. Get letter grades, percentile rankings, and a personalized training plan based on 75+ signals — all from a single PGN. Whether you want to improve your game or need forensic evidence for a dispute, our engine delivers insights no free tool can match.

or
Download Sample Report
$9.99
Game Review
coaching + SuperGM comparison
75+
Analytical Signals
per game analyzed
10,000+
Games in Database
SuperGM + engine baseline

Three Steps. One PGN. Deep Insights.

Upload any PGN and get a full coaching breakdown in minutes. No software to install, no engine setup required.

01
Upload Your PGN
Drag and drop any PGN file. We accept games from Lichess, Chess.com, FIDE events, or any standard PGN export. Select which side to analyze. PGNs with clock data unlock additional timing dimensions.
02
35-Dimension Analysis
Our engine evaluates every position at depth 20+, then runs behavioral forensics across crisis correlation, error texture, tilt recovery, Markov chain analysis, and 31 more signals.
03
Download Your Report
Receive your evidence-grade forensic report in under 5 minutes. Full behavioral dissection, phase analysis, verdict with confidence score, and forensic charts.

Know Your Game. Own Your Improvement.

Deep coaching + forensic verdict starting at $9.99. Full evidence-grade reports for $29.

Personal
Game Review
$9.99
per game
  • Engine likelihood verdict (0-100)
  • SuperGM comparison (Magnus, Hikaru, Firouzja...)
  • Play style archetype
  • Performance scorecard with letter grades
  • Percentile rankings vs 4,000+ SuperGM games
  • Weakness map with signal radar
  • Move-by-move game narrative
  • 30-day personalized training plan
Organization
Event Retainer
$499
per event
  • Unlimited game analyses for 7 days
  • White-label forensic reports with your branding
  • Bulk PGN upload (entire tournament)
  • Dedicated support channel
  • Arbiter summary dashboard
Inconclusive? You pay nothing. If the analysis returns an INCONCLUSIVE verdict due to insufficient game data, we refund your purchase in full. No questions asked.

Your $9.99 Report Comes With Your Own Personal AI Coach

We don't just show you what went wrong.
We turn your unique 75-signal fingerprint into a complete personalized training system.

🗺

Your Personal Weakness Map

Exact breakdown of where you lose the most rating points — critical moments, time pressure, specific openings, endgames, and tilt spots.

Ranked by impact. No more guessing what to study.

🔥

Tilt Recovery Drills

Custom mental-game exercises based on your real recovery patterns after blunders.

Learn to stop the bleeding when the game goes south.

📅

30-Day Personalized Training Plan

A complete day-by-day roadmap built only from your games and your 75 signals.

Targeted puzzles, opening fixes, endgame focus — all unique to you.

Plus: SuperGM comparison · play style archetype · percentile rankings · forensic verdict

75+ Analytical Signals

Validated on 10,000+ games — 4,000+ verified SuperGM games and 6,000+ confirmed engine games. Each signal targets a distinct fingerprint that separates human cognition from engine assistance.

Pressure Response
How accuracy changes in critical, high-stakes positions. Humans react differently to pressure than engines.
Behavioral
Error Distribution
The shape and texture of errors across the game. Human error patterns are fundamentally different from algorithmic ones.
Statistical
Recovery Dynamics
What happens after a mistake. The speed and pattern of recovery reveals whether a human or machine is playing.
Behavioral
Complexity Awareness
Whether accuracy correlates with position difficulty. Engines don't care how hard the position is. Humans do.
Statistical
Move Selection Profile
The distribution of move quality across the full game. Patterns invisible to the naked eye but clear to forensic analysis.
Statistical
Baseline Comparison
Every metric compared against our authenticated database of verified SuperGM performances and confirmed engine games.
Statistical
Phase Analysis
Opening, middlegame, and endgame evaluated independently. Each phase tells a different story about the player's cognition.
Behavioral
Strategic Tension
How board complexity is managed over time. Humans simplify to reduce cognitive load. Engines sustain maximum tension.
Behavioral
Risk Tolerance
Decision patterns in winning positions. Humans seek safety. Engines maximize regardless of volatility.
Behavioral
Brilliance Detection
A secondary system that identifies genuine human flow states — protecting elite players from false accusations.
ML-Enhanced
Tail Distribution
Mathematical analysis of the error distribution's shape. Human errors follow distinct statistical laws that engines cannot replicate.
Statistical
ML Verification
An independent machine learning model — trained on 10,000+ games — verifies the behavioral findings. Under 1% false positive rate on SuperGM play.
ML-Enhanced

Who Uses ChessForensics

Serious players who want to improve, competitors appealing bans, coaches protecting students, and tournament directors who need evidence.

Serious Players
Understand your play at a level no free tool can match
Not worried about cheating? Neither are most of our customers. ChessForensics gives you a deep-dive analysis of your play — phase-by-phase accuracy, pressure performance grading, recovery patterns after mistakes, and a personalized training plan based on YOUR data. No black boxes. Every signal explained. Better than any post-game review you've ever done.
Tournament Directors
Protect your prize fund with defensible evidence
When a $500 prize is on the line, "it felt suspicious" isn't enough. ChessForensics gives you a comprehensive forensic audit with a confidence-scored verdict. Make decisions you can defend to players, parents, and your federation.
Ban Appeal Players
Prove your innocence with data, not promises
Wrongly flagged by an automated system? A ChessForensics forensic report provides independent, third-party behavioral analysis that speaks the language arbiters understand. Your best evidence in a ban appeal.
Chess Coaches
Verify your students' competitive integrity
Protect your reputation and your students'. Run forensic checks on tournament games before results become official. Identify students who need guidance before it becomes a crisis.
League Administrators
Scalable integrity for online leagues
The Event Retainer covers your entire tournament. Upload every game, get every report. White-label reports with your league branding. Arbiter dashboard gives you an at-a-glance integrity overview.
CHESSFORENSICS
Game Review · Coaching Intelligence
LIKELY HUMAN — Natural Play
A+
Accuracy
C
Blunders
D
Best Moves
D
Consistency
Grades calibrated for Class A level
You played most like Alireza Firouzja
Play Style The Adaptable Player
Biggest Weakness Blunder Rate (56%)
Training Priority Tactical Vision
🎯 30-Day Training Plan
📚 Move-by-Move Narrative
♕ Weakness Map & Signal Radar

Sample Reports

See exactly what you're getting. From coaching-focused Game Reviews to full forensic breakdowns with behavioral dissection and signal charts.

Engine Flagged Sample (Full Forensic) Human Confirmed Sample (Magnus Carlsen) Game Review Sample ($9.99)

Reports are generated with full behavioral analysis and forensic charts.

Read: Did Hans Niemann Cheat? — Full Forensic Case Study →

Why This Exists

You finish a game. You click "Analysis." Lichess shows you a graph and a number. Average centipawn loss: 34. Cool. What does that mean? What should you study? Where exactly are you losing games? The free tools show you what happened. None of them tell you why, or what to do about it.
You watch YouTube videos. You solve random puzzles. You play 200 more blitz games and your rating stays exactly where it was six months ago. You're not lazy — you're practicing blind. Without knowing your specific weaknesses, every hour of study is a coin flip. Most players plateau not from lack of effort, but from lack of direction.
ChessForensics was built to solve that. We run your games through 75 behavioral signals — the same forensic depth used to analyze SuperGM play — and turn the results into something no other tool gives you: a personalized weakness map, a SuperGM comparison that shows who you actually play like, and a 30-day training plan built from your data, not someone else's curriculum.
Upload more games and the picture sharpens. One game catches a snapshot. Three games reveal patterns. You'll see exactly where your rating points are bleeding — whether it's opening preparation, middlegame calculation, time pressure decisions, or post-blunder tilt — ranked by how much each weakness actually costs you.
And yes — every report also includes a full forensic integrity verdict. If you need evidence for a ban appeal, a tournament dispute, or just want to prove your play is legitimate, the same engine that coaches you can defend you. One upload. Coaching + proof.

Common Questions

How accurate is ChessForensics?
Validated on 10,000+ games — 4,084 verified SuperGM games and 5,936 confirmed engine games. Cross-validated false positive rate: under 1% on SuperGM play. Zero false positives on named SuperGM accounts (Magnus Carlsen, Alireza Firouzja, Nihal Sarin, David Navara, Etienne Bacrot). Engine detection rate: 99%+ on depth-20 engine play. Every report includes a confidence score and full signal breakdown.
Will this work for my Elo range?
ChessForensics is calibrated against SuperGM blitz play (2700+ Elo). For players below 2000, the behavioral signals are less precise because lower-rated games have higher natural variance. The report will clearly state when confidence is reduced due to game characteristics.
Can I use this to appeal a Chess.com or Lichess ban?
Yes. Export your game as a PGN (both platforms allow this even after a ban), upload it, and attach the ChessForensics report to your appeal. It provides independent, third-party evidence that your play matches human behavioral patterns. Whether the platform accepts it is their decision, but you'll have the data on your side.
What if the result is inconclusive?
You pay nothing. If ChessForensics returns an INCONCLUSIVE verdict due to insufficient game data (too few moves, missing critical positions), we refund your purchase in full. No questions asked.
Do I need clock data in my PGN?
Strongly recommended but not required. PGNs with clock annotations (like Lichess exports with %clk tags) unlock additional timing-based behavioral dimensions. Without clock data, the analysis relies purely on move quality signals — still effective, but fewer dimensions are available.
How long does the analysis take?
Typically 2–5 minutes per game depending on game length. Our engine evaluates every position at depth 20+ with multi-PV analysis, then runs the full 35-dimension behavioral pipeline. You'll see a real-time status update while it processes.
Is ChessForensics affiliated with Chess.com, Lichess, or FIDE?
No. ChessForensics is a fully independent forensic analysis tool. We are not affiliated with any chess platform or governing body. Our reports are independent third-party assessments — that independence is what makes them valuable as evidence.
What's included in the Event Retainer?
Unlimited game analyses for 7 days, white-label forensic reports with your tournament branding, bulk PGN upload (entire tournament at once), a dedicated support channel, and arbiter summary dashboard showing flagged games at a glance. Contact us for custom API integration with your pairing software.
10,000+
Games in Validation Dataset
4,000+ verified SuperGM games · 6,000+ confirmed engine games · 80+ titled player accounts · Under 1% false positive rate

Get In Touch

Questions about reports, event retainers, API access, or partnerships? We respond within 24 hours.

support@chessforensics.com

♛ Begin Your Analysis

Upload your PGN. Get evidence-grade evidence in under 5 minutes.

Drop your PGN(s) here or click to browse
Up to 3 games for a combined analysis — Lichess, Chess.com, FIDE, or any standard PGN
ⓘ PGNs with clock data unlock additional timing dimensions
Your report includes:
✓ full behavioral analysis ✓ ML-verified verdict ✓ Phase-by-phase ACPL ✓ Worst-move forensic dissection ✓ Confidence score (0–100) ✓ Arbiter-ready Report download
Uploading...
🔒 Stripe Secured 🗑️ Deleted in 72h 💰 Inconclusive = Refund