Ninety percent of
disputes settle. Ninety
percent of the data disappears.

What is this case worth? The answer depends on who is asking and what they mean by "worth." To the plaintiff, worth means the expected recovery net of fees and costs. To the defendant, worth means the expected cost of resolution — including damages, legal fees, and reputational impact. To a litigation funder, worth means the risk-adjusted return on invested capital. To an insurance carrier, worth means the probable loss exposure against the policy. Each stakeholder needs a different answer, and each answer requires a different model. Capital's Case Valuation engine builds multi-perspective financial models for individual matters. The engine begins with comparable matter analysis: identifying cases with similar characteristics (jurisdiction, cause of action, judge, opposing counsel, damages theory, industry) and examining their outcomes — both the public verdicts and, where available through proprietary data partnerships, the confidential settlement terms. From these comparables, the engine constructs a probability distribution of potential outcomes: not a single number ("this case is worth $14 million") but a range with confidence intervals ("the P25-P75 settlement range is $12M-$18M; the P10-P90 verdict range is $4M-$22M"). The engine incorporates judicial behavior data — how the assigned judge has ruled on motions to dismiss, summary judgment, Daubert challenges, and damages instructions in comparable cases — to adjust the probability distribution for venue-specific factors. It incorporates opposing counsel profiling — settlement patterns, trial behavior, resource capacity — to estimate the likely negotiation dynamic. And it produces a risk-adjusted expected value that incorporates the probability of each outcome weighted by the financial result: the single number that represents the statistically expected financial outcome of the case, taking into account every possible resolution pathway.

A corporation, law firm, or litigation funder does not manage one case. It manages a portfolio of hundreds. The financial performance of that portfolio depends not on any individual case outcome but on the aggregate economics across all matters — the same principle that governs investment portfolio management. Yet most legal departments manage litigation as a collection of individual matters, not as a unified financial portfolio. Each case has its own budget, its own timeline, and its own outside counsel — but the portfolio-level questions are rarely asked. What is the total expected value of all active matters? What is the total cost basis invested to date? What is the projected net return? What is the portfolio's internal rate of return? Where is the risk concentrated — by practice area, jurisdiction, case type, or outside counsel? Capital's Portfolio Economics engine answers these questions by aggregating the case-level valuations produced by Engine 01 into a portfolio view. The dashboard shows the weighted expected value of the entire portfolio — the sum of each case's risk-adjusted expected value — alongside the total cost basis (legal fees and expenses incurred to date), the projected cost-to-completion (remaining fees and expenses to reach resolution), and the projected net return (expected recoveries minus total costs). The internal rate of return (IRR) is calculated based on the timing of cash flows — when costs were incurred and when recoveries are projected — producing the time-value-adjusted return that investors and CFOs use to compare litigation investment against other capital allocation options. Risk concentration analysis identifies portfolio vulnerabilities: excessive exposure to a single jurisdiction (if the Ninth Circuit reverses a favorable precedent, how many cases are affected?), dependence on a single damages theory, concentration in a single opposing party, or over-reliance on a single outside counsel firm.

When a general counsel asks outside counsel "what will this case cost?", the answer is almost always wrong. Not intentionally — but because litigation costs are inherently uncertain and the incentive structure rewards underestimation (to win the engagement) and later surprise (as scope expands). Budgets set at engagement are often exceeded by 30-50% by the time the matter resolves. This budget uncertainty makes litigation spend the most unpredictable line item on any corporate balance sheet. Capital's Cost Forecasting engine replaces single-number estimates with phase-by-phase cost models calibrated against actual expenditure data from comparable matters. The engine decomposes litigation into its standard phases — pre-filing investigation, pleading and motion practice, fact discovery, expert discovery, dispositive motions, pre-trial, trial, and post-trial — and projects the cost of each phase based on the matter's specific characteristics. A patent case in the Eastern District of Texas with 10 million documents in the discovery universe has a statistically different cost profile than a breach of contract case in the Southern District of New York with 50,000 documents. The model knows this because it has been trained on thousands of completed matters with actual expenditure data. Within each phase, costs are broken down by category: partner hours, associate hours, paralegal hours, e-discovery vendor costs, expert witness fees, deposition costs, court filing fees, and travel. The model projects a burn rate curve — showing how costs accumulate over time — that enables the general counsel to manage cash flow and budget allocation with the same precision applied to capital expenditure projects. When actual costs begin to diverge from the projection (outside counsel billing at a faster rate than projected for the current phase), the engine flags the variance and updates the total cost-to-completion estimate in real time.

Litigation finance is an investment discipline that requires the same rigor as private equity or venture capital — but applies it to legal claims rather than companies. A litigation funder evaluating a $5 million investment in a commercial dispute needs to assess the probability of a favorable outcome, the expected recovery amount, the projected duration (which determines the time-value of the investment), and the risk of partial or total loss. These assessments are made by experienced legal professionals using judgment and expertise — but the scale of the modern litigation finance market ($19.3 billion in 2025 and growing at 10.7% annually) demands computational support. Capital's Underwriting engine provides investment-grade analysis for funding decisions. Case merit assessment evaluates the legal strength of the claim by analyzing the cause of action, the applicable precedent, the assigned judge's track record on similar claims, and the quality of the evidence. Return multiple modeling projects the expected return on the funding investment under different outcome scenarios: the funder invests $5 million, the case settles for $15 million, the funder's share is $10 million under the funding agreement terms — producing a 2x gross return multiple. But what if the case settles for $8 million? What if it goes to trial and the verdict is $25 million? What if the defendant files for bankruptcy and the recovery is zero? Each scenario produces a different return, and the weighted average across all scenarios — adjusted by probability — determines whether the investment meets the funder's return threshold. Duration projection is critical because litigation finance returns are time-dependent: a 2x return in 18 months is a very different investment from a 2x return in 5 years. The engine projects case duration based on comparable matter timelines adjusted for the assigned judge's historical pace and the parties' settlement posture.

Damages are the economic foundation of every commercial dispute. The strength of a damages model — its methodological rigor, its factual support, its resilience to Daubert challenge — often determines whether a case settles for a reasonable amount, proceeds to a favorable verdict, or collapses when the damages expert is excluded. Capital's Damages Quantification engine builds computational damages models across the standard methodologies that courts have endorsed. Lost profits models calculate the revenue the plaintiff would have earned but for the defendant's wrongful conduct, using the plaintiff's historical financial performance, industry growth rates, market share data, and the specific causal mechanism through which the defendant's conduct affected the plaintiff's revenue. Reasonable royalty models — particularly relevant in patent and trade secret cases — estimate the hypothetical royalty that a willing licensor and willing licensee would have agreed upon in a negotiation at the time of first infringement, using the Georgia-Pacific factors and comparable license agreements. Unjust enrichment models quantify the profits the defendant gained through the wrongful conduct, tracing revenue to the specific infringing products or services. Each model includes sensitivity analysis: how does the damages range change if the market growth rate assumption increases from 3% to 5%? What if the royalty rate benchmark is 4% instead of 6%? What if the infringement period is 3 years instead of 5? These sensitivity tables show litigation teams exactly which assumptions drive the damages range — and which assumptions the opposing side is most likely to attack.

Outside counsel fees are the largest component of litigation expenditure — and the most opaque. When a law firm bills 2,400 partner hours on a patent case, is that reasonable? The answer depends on comparison: how many partner hours did comparable cases in the same jurisdiction require? What is the partner-to-associate leverage ratio (are senior attorneys doing work that junior attorneys could perform)? Is the staffing level proportional to the case's complexity, or is it inflated by organizational inefficiency at the firm? Capital's Fee Analysis engine provides these comparisons at a granularity that transforms fee management from an adversarial negotiation into a data-driven conversation. Rate benchmarking compares every biller's rate against market data segmented by firm size, practice area, geography, and experience level. A litigation partner at $1,122/hour at a 1,000+ attorney firm is at the 75th percentile — 7% above the market median. Is the premium justified by the case's complexity and the firm's specific expertise? That is a legitimate question the general counsel can now answer with data rather than intuition. Staffing efficiency analysis examines the partner-to-associate-to-paralegal ratio and compares it against benchmarks for the case type. A case staffed with 60% partner hours and 10% paralegal hours when comparable cases are staffed at 40% partner and 20% paralegal suggests that the firm is not leveraging junior resources effectively — driving up costs without adding proportional value. Task-level analysis compares the time spent on individual tasks (document review, deposition preparation, brief writing, motion practice) against benchmarks, identifying tasks where the firm's expenditure is significantly above the comparable-matter median.

The fundamental data problem in litigation economics is that verdicts are public but settlements are confidential — and 90% of cases settle. A case valuation model built on verdict data alone sees only the tail of the distribution: the cases that were contentious enough or valuable enough to go to trial. Settlement data — which represents how the vast majority of cases actually resolve — is invisible. The result is systematic bias: verdict-based models overestimate case value because verdicts are higher than settlements on average (the cases that go to trial tend to be higher-value, and the jury outcomes include both high-plaintiff and defense verdicts). Capital's Verdict & Settlement Database addresses this gap by combining publicly available verdict data with proprietary settlement data obtained through data partnerships with law firms, corporations, and litigation funders who contribute their anonymized resolution data in exchange for access to the aggregate intelligence. The database enables comparable-matter searches across 14 parameters: jurisdiction, cause of action, industry, company size, damages theory, amount in controversy, judge, opposing counsel firm, case duration, number of defendants, presence of government parties, class vs. individual action, appeal risk, and funding status. Each search returns not just the comparable outcomes but the statistical distribution of those outcomes — enabling the Case Valuation engine to calibrate its probability distributions against actual resolution data rather than theoretical models. The database grows with every matter Capital processes, creating a compounding data advantage: the more cases are valued through the platform, the more resolution data feeds back into the comparable-matter universe, and the more accurate future valuations become.

A litigation portfolio is not a static collection of independent cases. It is a dynamic system where outcomes are correlated: an adverse ruling on a legal issue in one case affects every other case raising the same issue. A change in judicial sentiment toward a damages theory affects every case relying on that theory. A recession affects settlement behavior across the entire portfolio as defendants become more willing to litigate rather than deploy cash for settlements. Managing a litigation portfolio requires the same risk modeling sophistication that investment portfolio management requires — and the same tools: Monte Carlo simulation, scenario analysis, stress testing, and value-at-risk calculations. Capital's Portfolio Risk engine runs Monte Carlo simulations across the entire portfolio, randomly sampling outcomes for each case from its individual probability distribution (produced by Engine 01) and aggregating the results across thousands of iterations. The output is a distribution of total portfolio outcomes — showing the range from best case (every case resolves at or above its expected value) to worst case (multiple adverse outcomes simultaneously), and the probability of each portfolio-level result. Scenario analysis allows the general counsel or fund manager to test specific hypotheses: What happens to the portfolio if the Supreme Court grants cert on the issue underlying 30% of our cases? What happens if our largest defendant files for bankruptcy? What happens if outside counsel rates increase 10% across the board? Each scenario produces a revised portfolio valuation, enabling proactive decision-making rather than reactive crisis management. Value-at-risk (VaR) calculations quantify the maximum portfolio loss at a given confidence level — the metric that CFOs and boards use to understand and approve risk exposure.