The Rating System
How Jits Ratings Work
Baseball has batting averages. Wrestling has seedings. Soccer tracks goals across entire seasons. Youth jiu-jitsu has none of it — competition data is scattered across dozens of organizations, buried in brackets that vanish days after an event.
Jits ratings change that. One number. Updated after every tournament. Built from every match. Calibrated against every opponent.
Every rating is recalculated from scratch using the complete match history — not patched incrementally. When the model improves, every fighter's rating improves with it.
Ratings give parents preparation. Coaches get game plans. Fighters get a clear picture of where they stand. When a family opens a bracket and sees a rating next to every name, the entire competitive experience changes.
Why Not Just Use Elo?
Most people are familiar with Elo — the rating system originally designed for chess in the 1960s. It's simple, well-known, and completely inadequate for youth jiu-jitsu. Here's why we built something better.
Elo treats every win the same. Beat a first-time competitor or beat a national champion — same math, same points. Jits doesn't. Who you beat matters enormously, and the system reflects that.
Elo has no concept of confidence. A fighter who's 2-0 and a fighter who's 40-2 with the same win percentage are treated identically. Jits tracks how certain we are about every rating. A small sample is treated with appropriate skepticism. A deep track record is trusted.
Elo can be farmed. Compete only against beginners, rack up wins, inflate your number. Jits detects this pattern and adjusts. You can't build a real rating by avoiding real competition.
Elo ignores context. Was this an IBJJF Worlds final or a local round-robin? Did you submit your opponent or win by a single advantage? Did you compete up in weight or age? Elo doesn't care. Jits does.
Elo punishes kids who challenge themselves. A young fighter who competes up an age division and loses gets crushed in a basic Elo system. Jits protects competitors who test themselves against tougher opponents. We believe challenging yourself should never be penalized.
Elo creates rating inflation over time. Without decay and recalibration, numbers drift upward and become meaningless. Jits uses confidence intervals that widen with inactivity, keeping ratings honest and current.
How Jits Ratings Actually Work
Jits ratings are powered by a proprietary engine we built from the ground up for combat sports. It solves the problems Elo can't — modeling uncertainty, adapting to volatility, and handling the irregular competition schedules that define youth BJJ.
Three components are tracked for every fighter: Rating (the skill estimate), Confidence (how certain the system is about that estimate), and Volatility (how much performance varies from event to event). All three work together. Rating is the headline number, but confidence is what determines how much that number can move.
New fighters start with wide confidence intervals — the system is openly saying I don't know this person yet. Every match narrows that window. A fighter with fifty matches has a rating the system trusts deeply. A fighter with three matches has a rating the system treats as a rough draft. This is by design. The alternative — pretending early ratings are as reliable as proven ones — would be dishonest.
Every match is processed through multiple layers of context before a rating change is calculated. The raw result — win or loss — is just the starting point.
What Goes Into Every Rating Change
Opponent Strength. Who you beat matters more than how many you beat. Defeating a proven competitor moves your rating more than beating someone in their first tournament.
Match Quality. Submissions carry more weight than points. A rear-naked choke sends a stronger signal than a 2-0 advantage win. How you win matters.
Tournament Prestige. Competing at the highest levels — IBJJF, ADCC — is rewarded. These are the hardest brackets in the sport, and results there carry amplified significance.
Recency. Recent matches count more than old ones. Your rating reflects who you are now, not who you were two years ago.
Division Difficulty. Competing in a tougher division — against higher-rated opponents or at a higher belt level — amplifies the significance of each result.
Sample Size Regression. Small records are treated with appropriate skepticism. A 2-0 fighter isn't rated like a proven champion — the system waits for evidence before committing to a rating.
What Moves Your Rating
Opponent strength is the biggest factor. Victories against higher-rated, proven competitors are worth significantly more than wins against newcomers. On the flip side, losing to an unproven opponent hurts more than losing to a proven one.
How you win and lose matters. Submissions carry more weight than points — on both sides. Finishing your opponent signals dominance. Being finished signals the opposite.
Consistency builds durability. The more data the system has, the more accurate and stable your rating becomes. One big weekend doesn't make a rating — sustained performance across many matches does. A fighter with a deep match history will see smaller, steadier changes. A fighter still building that history will see bigger swings. Both are correct behavior.
Prestigious events carry amplified weight. IBJJF and ADCC results are rewarded. Show up where the best compete.
Farming is detected and penalized. If you're consistently beating beginners and avoiding ranked opponents, the system recognizes the pattern and progressively reduces the value of those wins. You can't inflate your way to the top.
Why Ratings Swing — and When They Settle
If your child's rating just moved a thousand points in one weekend, that is the system working correctly — not a bug. Big swings are the most common source of confusion, so here's exactly what's happening.
Early ratings are supposed to be volatile. When the engine has only seen a handful of matches, its confidence is low. Low confidence means every new result carries enormous weight. A 4-0 weekend for a fighter with five career matches might move their rating by 2,000+ points. That same 4-0 for a veteran with forty matches might move it by 300. This isn't unfair — it's the system saying I'm still learning who this person is. The more matches you have, the more stable your rating becomes. Most fighters see their rating settle into a tighter range after 10–15 matches.
Elijah Gil goes 4-0 on January 10 → +920 Jits (230 per match). Wide confidence interval, meaningful move.
Same fighter, March 7: goes 8-0 — double the wins — but gains only +680 Jits (85 per match). His confidence interval tightened between January and March. Same dominance, smaller swing per match.
Who you face matters more than how many you beat. Going 5-0 against first-time competitors at a local event will not move your rating the same as going 3-2 against proven, rated opponents at a major. Parents sometimes see a dominant weekend produce a modest rating change and wonder why. The answer is almost always opponent strength — the system doesn't reward wins against unknown competition the same way it rewards wins against fighters with deep track records.
January 24: Ruben Perez goes 2-0 → +5,130 Jits. Madyson Bilbrey goes 13-0 → +3,590 Jits. Madyson won six times more matches, undefeated, and gained less. Ruben's two opponents were stronger. The engine doesn't count wins — it weighs them.
Losses to lower-rated opponents sting. The engine expected you to win. When you don't, the correction is larger. A loss to a higher-rated opponent barely moves the needle — the system already expected that to be a tough match. But losing to someone the system expected you to beat signals that the current rating might be too high, and it adjusts accordingly.
Michael Jones goes 2-1 → -2,980 Jits (dropped to Advanced, 2,363). He won more than he lost and still dropped nearly three thousand points — the loss was to a lower-rated opponent the engine expected him to beat. Meanwhile, Charlie Trejo goes 1-4 → -1,150 Jits. Charlie lost more matches but dropped less — the system expected those losses against tougher competition.
Ratings recalculate from the entire history. Every week, the engine replays every match from scratch. This means your rating can shift even on a week where you didn't compete — because an opponent's rating changed, which retroactively changes what beating or losing to them was worth. This is a feature: as the system learns more about the competitive landscape, everyone's rating gets more accurate, not just yours.
Inactivity widens the window. If a fighter hasn't competed in several months, their confidence interval gradually opens back up. The system isn't punishing inactivity — it's acknowledging that kids develop (or regress) between tournaments, and it shouldn't pretend a six-month-old rating is as reliable as a fresh one. The first tournament back will produce a bigger swing than usual as the system recalibrates.
What a breakout looks like: Coraline Costello Airheart of Pedigo Submission Fighting goes 11-0 in a single weekend → +4,770 Jits to reach S-Tier. Her confidence was wide, her weekend was dominant, and the system responded with a massive update. A veteran with the same 11-0 record and a tight confidence window would see a fraction of that move.
The bottom line: big swings are normal and expected for new or returning competitors. They are the system being honest about what it knows and what it doesn't. If your child has been competing regularly for a year, their rating will be stable and meaningful. If they're in their first season, treat the rating as a work in progress — it's getting smarter with every match.
What Ratings Don't Tell You
Ratings are probabilities, not prophecies. They capture tournament performance — not the full picture of who someone is on the mat.
David Israel-sanchez, rated 1,185 Jits (Beginner), beats Lucas Helio Rubio, rated 5,130 Jits, by points at the ADCC US Open in San Diego. A 3,945-Jits gap. The engine gave him a 1.1% chance of winning. He won anyway. Upsets like this happen across our platform every single weekend.
Ratings start at the same baseline regardless of belt — a 1200 is a 1200, earned through competition. Context still matters: a grey belt and a blue belt with identical ratings earned them against different opponent pools and competition levels.
Training Experience Points (TXP)
TXP credits the training time that doesn't show up in tournament brackets. A fighter who's been training for eight months but only competed twice has developed real skills the system hasn't seen yet. TXP bridges that gap with a small display bonus (up to 200 points) that phases out as match volume grows.
Fighters with fewer than fifteen matches at their current belt are eligible. After that, the rating engine has enough signal and your results speak for themselves.
Tiers
Every rated fighter is placed into a tier based on where they stand relative to all rated competitors. Tiers are percentile-based — they reflect your position in the overall distribution, not an arbitrary cutoff.
The tier system runs from Beginner through S-Tier, with each tier representing a progressively smaller slice of the competitive population. Most fighters fall in the middle tiers. The highest tiers are reserved for proven competitors who have demonstrated sustained excellence across a meaningful sample of matches.
Tiers update when ratings update. They're a quick way to understand competitive standing at a glance — especially useful when scanning a bracket or comparing opponents.
Our Commitment
The model is continuously improved. We study edge cases, adjust for fairness, and recalculate the entire history when the model updates. Accuracy over speed — ratings update after events are fully processed and validated, not in real-time. The data has to be right before the numbers move.
The rating system is the backbone of everything we build. We take it seriously.