Rescission treats a contract as if it never happened. Rather than awarding damages for a broken promise or ordering the promise carried out, rescission cancels the agreement and requires both sides to return whatever they received under it -- money refunded, goods returned, property reconveyed -- so the parties end up, as nearly as practical, where they started. It is available both as a remedy a court orders and, in many circumstances, as a self-help right a party can exercise unilaterally by giving notice, subject to the other side's ability to contest it in court.

Because rescission unwinds the whole bargain rather than compensating for one broken piece of it, courts require a defect that goes to the heart of the agreement -- not every breach justifies undoing the entire contract. The available grounds cluster around problems in how the agreement was formed or in whether the parties actually agreed to the same thing, which is what separates rescission from an ordinary damages claim for breach.

Rescission is most commonly available for mutual mistake -- both parties shared a false belief about a basic assumption on which the contract was made, and the mistake has a material effect on the exchange. It is also available for fraud or fraudulent misrepresentation, material (non-fraudulent) misrepresentation that induced the agreement, duress, undue influence, and lack of capacity.

A material breach by the other party can also support rescission as an election of remedies: rather than suing for damages while treating the contract as still binding, the non-breaching party can instead treat the material breach as terminating its own obligations and seek restitution of what it has already conferred. This overlaps with, but is doctrinally distinct from, simply suing for breach-of-contract damages.

Rescission cancels the contract; reformation rewrites it to match what the parties actually agreed to, then enforces the corrected version. A party seeking to unwind a bad deal wants rescission; a party who agrees the deal is sound but says the written document got a term wrong wants reformation instead. The two are frequently pleaded in the alternative because they respond to different theories of what went wrong.

Rescission also differs from a damages remedy in what it restores. Damages try to give the injured party the benefit of the bargain (or, for reliance damages, to make the party whole for expenditures made in reliance on the contract). Rescission plus restitution instead tries to erase the bargain altogether, returning both sides to their pre-contract position -- which can be more or less favorable to a given party than a damages award would have been, depending on how the deal has performed.

A party seeking rescission must generally act promptly after discovering the ground for it, must be able to return (or offer to return) what it received under the contract, and must show the defect was material to the bargain rather than a minor or immaterial flaw. Courts also weigh whether an innocent third party has since acquired rights that unwinding the contract would unfairly disturb.

Defendants resist rescission by showing the plaintiff ratified the contract after discovering the defect (continuing to accept its benefits instead of promptly disaffirming), that the plaintiff cannot restore what it received, that the alleged mistake or misrepresentation was not material, or that too much time has passed and intervening reliance by third parties now makes unwinding the deal inequitable.

In Juricratic, rescission is modeled as an alternative claim path running parallel to a straight breach-of-contract damages path from the same underlying facts, since the two theories can lead to materially different recoveries and are frequently pleaded together in the alternative. Which path dominates the case's expected value depends heavily on how the deal has since performed -- rescission is more attractive when the contract has turned out badly for the party seeking to unwind it than a damages calculation would reflect.

Because materiality of the mistake or misrepresentation and the promptness of the election to rescind are both fact-intensive and separately contestable, a user can adjust dials for each -- how material was the defect, how quickly did the party act after discovering it -- and observe how the modeled viability of the rescission path shifts relative to the damages alternative. These are simulation inputs, not predictions of how any court will rule.