Methodology

How LexCases scores cases.

Every case LexCases returns has an authority score from 0 to 100. This is not a black box. It is a transparent four-component rubric, visible for every result.

Score composition: total 100 points

40

25

20

15

Doctrinal

Factual

Authority

Currency

Why this matters

The primary reason you cite a case is to establish a legal principle. A case with high doctrinal alignment states, as its ratio decidendi, exactly the rule your argument turns on. A case that only touches your area obliquely scores low here, even if every other factor is strong.

Example

You argue that a landlord owed a duty of care to a visitor injured on the premises. Donoghue v Stevenson [1932] AC 562 scores high: the neighbour principle is the direct ancestor of occupier liability doctrine. A case about economic loss from negligent misstatement scores low. The area is related but the rule is different.

Illustrative sub-score

40/40pts

Donoghue v Stevenson [1932] AC 562

Worked example

A score of 94/100: what does it actually mean?

Brief: cases on the neighbour principle and duty of care in product liability.

  • 40

    Doctrinal alignment

    The ratio of Donoghue is the neighbour principle, the exact duty-of-care rule the brief turns on. A direct match stated as the binding ratio, so it takes full marks.

  • 19

    Factual proximity

    Measured by comparing the case facts to your brief. The consumer-manufacturer relationship is the same pattern, though the snail-in-a-bottle facts sit at one remove, so it lands high but short of full.

  • 20

    Authority weight

    A House of Lords decision, binding in a UK search. Read straight from the court, so there is no judgement call: the apex court takes full marks.

  • 15

    Currency & standing

    Donoghue has never been overruled and is still applied today. Read from its treatment record. Being refined by Caparo narrows the test, it does not unsettle the case, so it keeps full standing.

Donoghue v Stevenson [1932] AC 562

  • Doctrinal alignment

    40 /40
  • Factual proximity

    19 /25
  • Authority weight

    20 /20
  • Currency & standing

    15 /15

Authority score

the four parts, summed out of 100

0

The formula

The exact maths, in the open.

No black box. This is the whole calculation, the same one that runs on every result. Two parts follow hard rules from the court and its treatment, one is a real similarity distance, and one is broken down so you can see the judgement.

Authority score = Doctrinal 40 + Factual 25 + Authority 20 + Currency 15

summed to a score out of 100

40

Doctrinal alignment

  • How directly the ratio answers your point direct 32 / partial 22 / tangential 10
  • Is that point the binding ratio ratio +6 / mixed +2 / obiter -2
  • Does it back your position clear stance +2 / neutral -6
25

Factual proximity

  • The real distance between your brief and the case facts vector similarity, scaled to a 0 to 25 score
20

Authority weight

  • Your courts outrank foreign ones of the same level your apex 20 / your appeal court 16 / your first instance 12
  • A foreign court is only ever persuasive foreign top court 11 / foreign lower court 7
15

Currency & standing

  • Is it still good law good law 15 / distinguished 11 / doubted 10 / overruled 1

The model never sets the score. It reports the facts, the court, the treatment, how directly the ratio answers your point, and LexCases does the arithmetic, so the same brief and the same case always produce the same number.

What AI tools give you

  • No explanation of why a case is relevant
  • Citations may be fabricated or misquoted
  • No check on whether the case is still good law
  • No authority ranking, all results look equal

What LexCases gives you

  • A transparent score broken into four readable components
  • Every case independently re-checked before it reaches you, and anything unconfirmed is dropped rather than replaced
  • Treatment flagged as good law, criticised or overruled, from the model's reading and not a citator
  • Ranked by relevance to your specific argument, not keyword density

Read next: How to study law cases and Choosing a legal research tool.

Run a search and see the scores for your argument.

Every result shows the score and the four-component breakdown.

Search cases
  • Home
  • Search case law
  • About
  • Pricing
  • Built with you: the changelog
  • Question Advisor
  • Foundations
  • Study method
  • Resources
  • Contact
  • Terms of Service
  • Privacy Policy
  • Refunds Policy
  • Legal research for Dublin law students
  • How to study law cases
  • Best legal research tools for law students
  • Best AI Tools for Law School: What Reddit Says
  • AI Tools for Studying Law in the UK: Reddit
  • Can AI Find Case Law for You? What Reddit Says
  • Do AI Case Summaries Hurt You? What Reddit Says
  • Are Law School Study Aids Worth It? Reddit Says
  • AI Disclaimer
  • Duty of care (negligence)
  • Negligent misstatement (pure economic loss)
  • Consideration
  • Misrepresentation
  • Mens rea, intention and recklessness
  • Necessity and murder
  • Judicial review, grounds of review
  • Frustration of contract
  • Promissory estoppel