The Mechanism
Underwriting takes the distribution prediction produced and asks a different question: at what price, and under what structure, does deploying capital against this position clear the required return once the position's risk and duration are priced in. The mechanism runs the prediction's outcome and duration bands through a pricing function that incorporates the capital's own cost, the specific structure being offered, non-recourse advance, portfolio facility, structured product, and the concentration effect this position would have if added to the existing book. The output is a priced reference for the independent capital partner reviewing the position: a suggested amount range, a term, a repayment mechanic, and the conditions under which the reference is withdrawn or repriced. The partner decides on its own whether and on what terms to proceed.
None of this belongs inside the prediction function itself. A model that adjusts its outcome estimate to justify a deal the desk wants to close is no longer measuring the matter; it is rationalizing a decision that has already been made, which is why underwriting is treated as a distinct, downstream consumer of prediction rather than a collaborator in producing it.
Structure selection is itself part of the mechanism, not an afterthought applied once a price is set. The same predicted distribution can support a straightforward non-recourse advance, a structured product with staged tranches tied to procedural milestones, or a portfolio facility spanning many matters at once, and which structure fits depends on the position's duration profile and the capital source's own liquidity needs as much as on the underlying probability itself.
Inputs and Outputs
Inputs are the prediction stage's outcome and duration distributions, the capital's cost and available structures, and the current portfolio's existing concentration by venue, judge, defendant, provider, and referral source. Outputs are a priced reference for the independent partner and, critically, a documented record of exactly which inputs produced it, so that a position that performs unexpectedly can be traced back to whether the prediction was wrong or the pricing was wrong, two very different problems requiring different fixes.
The documented record is what allows a later review to separate two very different failure modes that otherwise look identical from the outside: a model that mispredicted the matter's true distribution, and a desk that priced correctly against a good prediction but chose a structure that did not fit the position's actual duration risk.
What Can Go Wrong
The dominant failure mode is economics override: a sufficiently attractive projected return causing an unresolved integrity deficiency, a break in the evidence chain, an unverified lien, a duplicate advance risk, to be waved through rather than cured first. The standard this platform follows treats that sequence as structurally backwards and requires integrity resolution before economic terms are finalized. A second failure is pricing a position in isolation without checking what it does to existing portfolio concentration, since an asset acceptable on its own can be unacceptable once it is the fifth position this book holds against the same defendant or the same referral source.
A third failure is structural mismatch: pricing a position correctly on expected value while selecting a structure whose repayment mechanics assume a duration profile the prediction did not actually support, which produces a position that is economically sound on paper but liquidity-mismatched against the capital source's own needs in practice.
A fourth, easily missed failure is pricing against a stale prediction. A position underwritten weeks or months after the prediction was generated, without a refresh, is priced against a snapshot that may no longer reflect the matter's current procedural posture, particularly in fast-moving verticals where docket events accumulate quickly between the initial prediction and the underwriting decision that follows it.
How Criterica Implements This Stage
Criterica Capital prices each position against Criterica Intelligence's models rather than file review alone, with concentration checks run against the live book, and prepares it for independent capital partners. Each partner makes its own decision. Criterica Group's Legal Asset Integrity Standard sits above this stage as a gate: an asset with an unresolved hard-stop integrity failure does not reach pricing at all, regardless of its projected return, and the standard's Integrity-Economics Matrix makes that sequencing explicit rather than leaving it to individual underwriter judgment.
The handoff to deployment is deliberately narrow: once terms are finalized and integrity is confirmed, the next stage's job is to execute exactly those terms against a fully verified baseline, not to revisit the pricing decision itself.
Discuss the Underwrite Stage
Institutional buyers and capital partners evaluating the Regulated Outcome Lifecycle.