Our work
Decision Records
Decision Records make NLI’s work inspectable. Each record fixes the decisive question, controlling dependency, falsification and defeat conditions, classification, NLI’s exact role, the result the evidence can carry, and the point at which the claim must stop.
What a Decision Record does
It lets a reader evaluate one consequential claim without relying on NLI’s confidence, reputation, or account of its own effect.
A case narrative can describe events. A Decision Record must additionally distinguish fact from allegation, NLI’s role from counsel’s and the institution’s, analytical result from institutional outcome, and participation from causation.
How to read a Decision Record
Three record classes, three different claims
Method Demonstrations show how the method works. Comparative Determinations test dated predictions against what an institution actually decided. Future Verified Engagement Records will document authorized, evidence-supported NLI work in actual matters.
The same format can carry all three, but the status and permissible attribution are not interchangeable.
Method Demonstrations
Representative constructions used to demonstrate the proof procedure, classification discipline, limits, and publication format. They do not establish a client relationship, filing, NLI participation in a named matter, institutional adoption, observed outcome, or causal effect.
Comparative Determinations
A determination NLI published on a date, later set against the decision the institution actually reached. The prediction is timestamped by a third party before the outcome is known, so the comparison can be checked by anyone. It establishes no client relationship and no participation in the case.
Verified Engagement Records
A future record class for actual NLI work published only when the matter, publication authority, NLI role, institutional response, observed result, attribution, and material limitations can be responsibly verified.
What the published records do and do not claim
NLI was not a party, counsel, or participant in any case below. These are records of published analysis, not reports of NLI involvement in a named matter.
The Comparative Determinations below score NLI’s own earlier published determinations against subsequent Supreme Court decisions. They establish that a determination was made on a stated date and how it compared to the outcome. They do not establish that NLI influenced any decision, and no favorable comparison should be read as a prediction of results in your matter. Method Demonstrations remain demonstrations unless and until a separate, verified record is published under the engagement standard described below.
What Governs the Number of Comparisons
How many comparisons exist is set by how often the Court issues a determinative resolution, not by which ones NLI elects to publish.
NLI has published determinations on many more matters than appear below. A reader who knows that will reasonably ask what happened to the rest, and the answer is structural rather than editorial.
Petitions reach the Court each year
The overwhelming majority are denied without a merits determination. A denial leaves nothing against which an earlier determination can be scored.
Receive a determinative resolution
That is the entire population from which any comparison of this kind can be drawn, by NLI or by anyone else attempting the same exercise.
Expected to become scoreable in a year
NLI publishes on more matters than the Court will resolve, because the criterion is whether a matter warrants NLI’s participation rather than whether the Court is likely to take it. Some are never addressed at all.
What makes the record checkable is the rule rather than the count: when a matter NLI has published a determination on receives a determinative resolution, the comparison is published, whatever it shows. Two of the records below score NLI’s own determinations as wrong for exactly that reason. A record that reported only its successes would be equally consistent with any level of accuracy, and would therefore establish none. These figures describe the Court’s docket and NLI’s publication practice; they are not a projection of any result in your matter, and NLI was not a participant in any case scored below.
Comparative determination
A determination is only as good as the petition object it addresses
QuestionWhich authority governed the petition: the constitutional standard for detainee medical care, or the federal habeas limits on reviewing a state court's Brady materiality ruling?
Turns onWhether the operative legal object was the custodial-care standard or AEDPA deference to the state court's ruling.
Defeated byCorrect identification of the petition's actual legal object defeats any determination addressed to a different one.
Constitutional decidability
Authority cannot substitute for the question it must answer
QuestionDid the disposition defeat the material construction—or merely avoid its consequences?
Turns onWhether the necessary constitutional question was material, preserved, within authority, and capable of remedy.
Defeated byA controlling limitation, valid waiver, absence of remedy, or superior construction that expressly resolves the question.
Institutional accountability
Authority, injury, responsibility, and quo warranto
QuestionWhich actor possesses authority, which act exceeded it, and which remedy remains available?
Turns onA traceable chain from delegated authority through act, breach, injury, responsibility, and lawful correction.
Defeated byValid authority for the act, absence of cognizable injury or responsibility, or lack of an available corrective instrument.
Comparative determination
Courts may not supply a cause of action Congress did not create
QuestionMay a federal court create a cause of action under the Alien Tort Statute, or recognize aiding-and-abetting liability under the Torture Victim Protection Act, without congressional authorization?
Turns onWhether the asserted liabilities were created by Congress or supplied by judicial implication.
Defeated byA congressional authorization of either the ATS cause of action or TVPA secondary liability would have defeated the construction.
Comparative determination
Emergency power does not silently include the power to tax
QuestionDoes the International Emergency Economic Powers Act clearly authorize the President to impose tariffs?
Turns onWhether tariffs are taxing instruments, and therefore require a clear congressional delegation rather than an inference from emergency authority.
Defeated byA clear statutory delegation of tariff authority within IEEPA would have defeated the construction.
Comparative determination
Facts known to the officer may not be excised before totality review
QuestionMay a court categorically exclude facts known to the officer before assessing reasonable suspicion under the totality of the circumstances?
Turns onWhether reasonable suspicion is assessed on the whole picture, or only on the facts surviving a prior round of excision.
Defeated byAuthority permitting the exclusion of known circumstances before totality analysis would have defeated the construction.
Comparative determination
General similarity between a state duty and a federal requirement is not equivalence
QuestionDoes FIFRA preempt a state-law failure-to-warn claim that would require a cancer warning the EPA did not require?
Turns onWhether the state duty enforced a requirement parallel to the federal misbranding standard, or imposed an additional one the manufacturer could not lawfully satisfy.
Defeated byProof that the state duty enforced the same federal requirement would have sustained the construction.
Comparative determination
Prejudice must be measured against the evidence the jury actually saw
QuestionMust prejudice from constitutional trial error be evaluated solely on the evidence actually presented to the jury?
Turns onWhether prejudice review may take account of evidence that was never before the jury.
Defeated byAuthority permitting consideration of evidence outside the trial record would have defeated the construction.
Comparative determination
Qualified immunity requires close factual correspondence, not abstract generality
QuestionDid existing precedent clearly establish that this specific wristlock-removal sequence, following warning and passive resistance, violated the Fourth Amendment?
Turns onWhether the precedent relied on below described conduct similar enough to the officer's actual sequence to place the question beyond debate.
Defeated byA precedent addressing materially similar removal conduct would have defeated the qualified-immunity conclusion.
Counter-lawfare
When procedure becomes the effective remedy
QuestionWould the action retain its strategic value if the merits became immediately decidable?
Turns onWhether delay, expense, discovery, publicity, or procedural asymmetry supplies the effect the merits cannot independently warrant.
Defeated byA merits theory sufficient to justify the demanded result without relying on process-created cost or delay.
Publication standard for a Verified Engagement Record
A reader must be able to tell what happened, what NLI did, what the work establishes, and where the claim ends.
Named or identifiable matters require a verified public record, publication authority, protection of confidential material, and a result that may responsibly be attributed.
Context and available record
The parties or lawful anonymization, forum, posture, chronology, supplied evidence, adverse facts, missing material, and restrictions governing use.
NLI’s exact role
The construction or instrument NLI supplied—separate from advice, representation, filing, advocacy, decision, remedy, and acts of other participants.
Decisive construction
The question, controlling dependency, proof tests, classification, defeat conditions, and lawful or evidentiary limits.
Delivery and institutional response
Who received or used the instrument and whether the institution accepted, refuted, lawfully limited, replaced, or omitted the material construction.
Observed result
The event that can be verified and the narrower causal or analytical claim the evidence supports—without converting sequence or participation into causation.
Limitations and correction
What the result does not establish, what remains unresolved, what would change the classification, the correction path, and whether the record is current or superseded.
Attribution standard
Participation does not establish causation. Sequence does not establish effect. A published record claims only what its evidence, role, authority, and verified result can carry.
Bring a matter into review
A fit review tests the material question, available record, proposed instrument, authorized user, professional dependencies, and whether NLI can responsibly accept the work.
Start an inquiry