When more than one party contributed to an injury, the damages a plaintiff can recover are usually reduced by the plaintiff's own share of fault — but exactly how that reduction works, and whether any recovery survives at all, depends entirely on which fault rule the governing jurisdiction follows. The arithmetic itself is simple once the fault percentages are set. The part that actually decides the outcome is knowing which rule applies and where its cutoffs sit, because the same set of facts can produce a full recovery, a partial recovery, or zero recovery depending on the jurisdiction.

Fault percentages are also rarely a fixed, known number going into a case — they are typically the product of disputed evidence a jury ultimately assigns. That makes comparative fault a genuinely uncertain input, not a settled fact, right up until a verdict or a negotiated agreement fixes it. Modeling a range of plausible fault percentages, rather than betting the case on a single assumed number, is the same discipline a litigation simulation applies to any uncertain input — treat the fault percentage as a dial to test, not a number to assume.

Three basic frameworks exist. Pure comparative negligence allows a plaintiff to recover damages reduced by their percentage of fault no matter how high that percentage is — even a plaintiff found 90% at fault can recover the remaining 10%. Modified comparative negligence allows recovery only if the plaintiff's fault stays below a stated threshold, commonly 50% or 51% depending on the jurisdiction; cross that threshold and recovery is barred entirely. Contributory negligence, followed in a small number of jurisdictions, bars any recovery at all if the plaintiff bears any fault whatsoever, however small.

Because these rules produce dramatically different outcomes from the identical facts, confirming the governing jurisdiction's rule is the first step in any damages calculation involving shared fault — before spending time refining the fault percentage itself.

Fault percentages are typically determined by the fact-finder — a jury or judge in a bench trial — after weighing evidence of each party's conduct against the applicable standard of care. In a negotiated settlement, the parties instead reach the percentage by agreement, often as a proxy for what each side believes a jury would likely find.

The party asserting the other side's comparative fault generally carries the burden of proving it, just as the plaintiff carries the burden of proving the defendant's liability in the first instance — comparative fault is typically pleaded and proven as an affirmative defense, not assumed.

Once total damages and a fault percentage are established, the calculation itself is a single multiplication: the plaintiff's recoverable damages equal total damages multiplied by the defendant's percentage of fault (equivalently, total damages reduced by the plaintiff's own percentage of fault).

In a modified comparative negligence jurisdiction, the single most consequential number in the case is often not the exact fault percentage but which side of the statutory threshold it falls on. A finding of 49% plaintiff fault and a finding of 51% plaintiff fault can be the difference between a substantial partial recovery and nothing at all, even though the underlying facts barely moved.

This makes disputes at or near the threshold disproportionately high-stakes, and it is a common area where a defendant will invest heavily in evidence pushing the plaintiff's fault percentage just over the line, and a plaintiff will invest just as heavily in keeping it just under.

Because the fault percentage is contested until a verdict fixes it, treating it as a single assumed number when evaluating settlement can badly misstate the case's expected value — particularly near a modified-comparative cutoff, where a small shift in the assumed percentage can swing the outcome from a partial recovery to zero. Building a distribution of plausible fault findings, and running the expected-recovery calculation across that range rather than at one point estimate, produces a far more defensible settlement position than a single confidently-stated number.