LABORISAI
Jurimetrics

Jurimetrics that holds up.

Measuring, over real and public labor-court decisions, how often each claim is granted — by topic, court, jurisdiction and judge. With honest method and the sample in plain sight.

Methodology

The number only counts if the math is honest.

On the merits, not the dismissal

The rate counts only what was decided on the merits — granted, partial or denied. Dismissals without merit are excluded; we don't inflate the number.

Real decisions, not headnotes

We read each claim's outcome in the decision, reconciled claim by claim. Not generic case law, not opinion.

Only where there's a base

Court and judge get a profile only with a sufficient sample. Where there's no base, we don't assert — we flag low confidence.

Granularity

Four grains, from topic to judge.

01

Topic

Pattern by claim type — overtime, vacation, moral damages.

02

Court

Behavior of the unit where the case runs.

03

Jurisdiction

The set of courts in a region.

04

Judge

Individual profile and deviation from the pool.

The judge matters

The same claim, different judges, different outcomes.

Around a pool of ~36% grant rate, judges with a sufficient sample range from −36 to +64 percentage points. Ignoring that is defending in the dark.

  • Grant rate vs. the pool, with the deviation
  • Defensibility by judge and by court
  • approximately 400 judges meet the sample threshold
Judge profile · grant rate vs pool
Pool (average) 36%
Judge A 96% +60pp
Judge B 58% +22pp
Judge C 6% −30pp

approximately 400 judges meet the sample threshold

Pattern by topic

Topics range from 40% to 80%.

Legal fees 80%
Reflex effects 79%
Vacation 75%
FGTS 72%
Prior notice 71%
13th salary 68%
Overtime 56%
Moral damages 40%

Base: 14k+ real cases, nationwide, totaling ~62k merit decisions. Rate = granted + partial over merit-decided.

From base rate to your case

The base is public. Your number is yours.

The grant pattern is the starting point — what the courts do in general. Your case's probability emerges when that base is cross-checked with your evidence on that claimant.

And the claimant is almost always a former employee: you hold their full record — payroll, timesheets, eSocial, PPE log, termination. Each claim is checked against your evidence. Overtime granted 56% in the pool? With your sound time records, it drops to 28% — in your case. This cross-check runs inside your isolated instance and refines only your number.

Cross-check · Claim × Evidence
Overtime
Time records
Contradicts
Hazard pay
Report + PPE log
Confirms
Salary balance
Payroll + receipt
Paid
Probability adjusted by evidence 56% → 28%
Data provenance

Public, aggregated, anonymized.

The collective jurimetrics comes from public sources of the Regional Labor Courts, aggregated and anonymized — it's the base rate. Your internal data never enters that collective layer: it works on your side, in your isolated instance, refining your case's probability. The boundary is strict and one-way — your data helps you, without ever feeding the collective.

Glossary: jurimetrics · grant pattern.

See the jurimetrics applied to your caseload.

A demo with the grant pattern of your courts and judges.

Book a demo