HARTZ.LEGAL

Quantitative Case Valuation

Litigation Risk Analysis

JavaScript is required. This tool is an interactive application that runs entirely in your browser — no data leaves your device. To use it, please enable JavaScript and reload the page. The methodology overview below is available without JavaScript so you can decide whether the tool fits the matter before enabling.

How to use and interpret this tool

This tool values a contested matter by decomposing it into its case-dispositive inflection points — the discrete events at which the case can end — and combining the conditional probability of advancing past each one with the financial outcome at stake. Where the procedural stage doesn't capture enough nuance, individual factual or legal issues can be attached to stages to model their conditional outcomes inside the same chain. The sections below walk through each input in the order they appear in the form column, then explain how to interpret the valuation widgets the inputs drive.

Saving your work

Save keeps the current matter in this browser's local storage (instant, no dialog); it reappears in the saved-cases list, but only in this browser on this device. Export writes the matter to a JSON file at a location you choose, which you can reload later via Import or carry to another machine. Both the in-browser save and the exported JSON are unencrypted plaintext stored locally — nothing is uploaded, but anyone with access to your computer, browser profile, or the saved files can read the matter, so safeguard access accordingly.

Inflection points

An inflection point is a discrete event at which a contested case can end, advance, or be redirected — typically a motion to dismiss, summary judgment, class certification, Markman, Daubert, trial / final judgment, or appellate outcome. The tool ships with a standard catalog you can toggle on or off, and you can add custom stages when the procedural path of the case calls for it. The ordered chain of inflection points is the structural skeleton of the model: every dollar of expected value, expected cost burn, and option premium is derived by walking the chain stage-by-stage and combining the conditional probability of advancing past each stage with the financial consequence of advancing or exiting at it.

Configure only the stages the case actually faces. A case in arbitration won't see Markman or class cert; a single-defendant breach action probably won't see Daubert. Removing inapplicable stages keeps both the math and the visualizations focused, and avoids forcing nominal 100%-pass placeholder probabilities that distort the chain.

Issues and the decision tree

An issue is a discrete factual or legal question with a binary outcome (plaintiff prevails or defendant prevails). Each issue is assigned to a chain stage (the stage where that question gets decided). When a stage has one or more assigned issues, the stage's overall advance probability is derived from the issue tree rather than the stage slider — the user has explicitly modeled the stage's outcome.

Each side of an issue (plaintiff prevails / defendant prevails) can resolve into one of three terminal outcomes — Award damages, Defense win, or Continue to another issue. Linking to another issue forms a downstream decision tree; the linked issue can sit at the same stage or any later stage. The same issue can be referenced from multiple branches via reference linking — useful when the same underlying question can be reached by more than one procedural path. The tree is visualized and edited in the Decision Tree Canvas.

Issue Assessment

For each modeled issue, you can record a short free-text assessment of why the plaintiff might prevail and why the defendant might prevail. These notes are stored alongside the case, surfaced in the PDF export, and are intended to force the kind of two-sided diligence that produces well-calibrated probability inputs.

The discipline is symmetrical: if you can't articulate the defendant's best argument on an issue, your plaintiff-prevails probability is probably overstated; if you can't articulate the plaintiff's best argument, it is probably understated. The assessment fields are a place to write out those arguments before committing to a number on the slider in the section below.

Probabilities and the tails of the distribution

Each slider is a conditional probability: given that the case has reached this stage, how likely is it to advance? The complement is the chance of the case ending here. Path probabilities are products of conditional probabilities along the chain.

Be willing to use the tails of the distribution. Lawyers tend to clump probability assessments in the 30–70% range out of caution. The compounding effect of moderate probabilities over multiple stages can dramatically suppress apparent value (60% × 60% × 60% = 22%). Over- or under-stating tail probabilities is one of the largest single sources of valuation error.

For very low probabilities, the slider becomes hard to use precisely. Click any percentage value to type it directly. Real-world frequency analogues at the tail:

1% — like rolling double sixes twice in a row
0.1% — like having identical twins
0.01% — like guessing a four-digit PIN on the first try
0.001% — rarer than being dealt a straight flush in poker (1 in ~72,000)

Granularity matters — Tetlock's superforecasters. The Good Judgment Project found that the best probabilistic forecasters distinguish themselves not by domain expertise but by granularity: they assign 73% rather than "around 70%," and rounding their forecasts to coarser bins measurably degrades accuracy. Equally important is resisting what Tetlock calls "wrong-side-of-maybe" thinking — collapsing 70% to "yes" or 30% to "no" rather than treating each as the actual quantitative claim it is. The 0.5%-step slider and click-to-type input are designed for this discipline: pick the number you actually believe, not the nearest round one.

Costs and cost burn

The total litigation budget is allocated across stages — each stage carries a dollar amount representing what the case costs while it's in that phase. Expected cost burn weighs each stage's allocation by the probability the case actually reaches it (derived from the scenario tree), then discounts to present value by stage timing. Any unallocated portion of the total budget is treated as overhead incurred at probability 1.0, discounted to mid-litigation. The opposing party's budget is captured separately and is used in the settlement-zone and Divergent Expectations diagnostics, where each side's cost-of-fighting matters to the rational settle/litigate decision.

How to enter damages (P10 / P50 / P90)

Damages are entered on each terminal "Award damages" leaf in the issue tree — every damages outcome carries its own three-point distribution that approximates the range of plausible verdicts on that specific path. Set them by clicking the damages leaf in the canvas, or edit them in the Damages section below the tree.

For each damages leaf, imagine 100 versions of this matter resolving on this path — different juries, arbitrators, or judges, with the same underlying facts. Sort the resulting awards from lowest to highest:

Low (P10) — the 10th lowest. Only 10 of the 100 outcomes fall below.
High (P90) — the 10th highest. Only 10 exceed.
Median (P50) — the middle. Assess last, after bracketing the low and high (anchoring on the median first tends to compress the range artificially).

Three-point weighting — methodology basis. The tool weights these 0.25 / 0.50 / 0.25. This is a member of the well-established Swanson / Pearson-Tukey family of discrete approximations used in decision analysis and quantitative risk assessment (petroleum reserves estimation, project valuation, etc.). The method approximates the expected value of a continuous outcome distribution by placing probability mass at representative points of its sub-ranges: each endpoint stands in for the conditional mean of its tail quartile and the median for the central half, so the weighted sum closely tracks the true mean for symmetric-to-moderately-skewed distributions. The 0.25 / 0.50 / 0.25 weighting is also chosen for interpretive transparency — the three inputs correspond to the intuitive elicitation questions "there is a 10% chance the award is below ___," "there is a 90% chance it is below ___," and "the 50/50 point is ___." For strongly right-skewed (catastrophic) cases, where a three-point summary understates the tail's contribution to expected value, use the tail-branching approach described next rather than re-weighting.

Modeling catastrophic-verdict risk

The three-point damage estimate (low / median / high) approximates a roughly symmetric or mildly-skewed range well. For case types with a meaningful chance of a runaway or punitive-driven "nuclear" verdict — catastrophic personal injury, medical malpractice, product liability, or cases with significant punitive exposure — the outcome distribution has a long right tail that a single "high" figure cannot capture, because the tail's contribution to expected value is disproportionate and a three-point summary collapses it into one number.

For these cases, model the catastrophic outcome as its own terminating branch rather than stretching the "high" value of the main damages node. For example, after a liability win, branch into (a) an "ordinary award" node with a normally-calibrated low / median / high range, and (b) a separate low-probability "catastrophic / punitive verdict" node with its own high range (e.g., a 5% path to a $40–80M range). This keeps each damage node's three-point estimate well-calibrated, and — more importantly — makes the tail risk explicit and attackable: you can see its probability, its contribution to your settlement exposure and to the plaintiff's reservation value, and assess whether a Daubert motion, a damages-theory exclusion, or a punitive-damages cap would prune that branch. A catastrophic risk buried in an upper percentile is invisible; a catastrophic risk modeled as its own branch is a discrete, priceable, defensible target.

Counterclaim

When the defendant has a counterclaim, mark it on the cc root issue. Two modes:

Dependent — the counterclaim is reachable only through outcome edges pointing at it from main-tree issues. Right when the counterclaim is a legal fallback or alternative theory contingent on the outcome of an issue in the main case.
Independent — the counterclaim is litigated in parallel with the main case on its own merits. The two are valued separately and combined at the EV level (Net = main + counter).

On any counterclaim issue, the case-defendant is the counterplaintiff and the case-plaintiff is the counterdefendant. Damages awarded on a counterclaim flow defendant ← plaintiff; the engine auto-negates the EV contribution so a counterclaim award reduces the case-plaintiff's recovery and increases the case-defendant's.

Fee-shifting

Many statutes and contracts shift attorneys' fees to the prevailing party — Lanham Act § 35, Copyright Act § 505, civil rights actions under 42 U.S.C. § 1988, FRCP 68 offers of judgment, contractual prevailing-party clauses. The tool models fee-shifting as an additive adjustment to the gross plaintiff-perspective EV on the appropriate paths: when the plaintiff prevails on a fee-shift-eligible path, the defendant pays the plaintiff's accumulated costs through judgment; when the defendant prevails, the plaintiff pays the defendant's. Direction (one-way plaintiff-only, one-way defendant-only, two-way symmetric) is controlled in the section.

Fee-shifting is folded into the gross EV that drives the settlement-zone bounds rather than into the cost-burn calculation — the cost-burn line continues to reflect what each side pays its own counsel along the way, while fee-shifting flows as a payment between the parties at judgment. Watch for the asymmetric case: a one-way plaintiff-favoring fee-shift (typical of civil-rights and consumer-protection statutes) materially raises the plaintiff's settlement floor and the defendant's ceiling in close cases, because each marginal dollar of defense spend doubles as a transfer the defendant will owe back if the plaintiff prevails.

Prejudgment interest

Prejudgment interest accrues from the date of accrual to the date of judgment. The tool grosses up the damages distribution by (1 + rate)^{(years already accrued + years to resolution)} on prevail paths. The result is then discounted to present value at your discount rate.

Watch the rate comparison. When the PJI rate exceeds your discount rate, delay benefits the prevailing party in present-value terms — a major settlement-leverage consideration in jurisdictions with high statutory PJI (New York's 9% under CPLR 5004, for example, materially exceeds typical corporate discount rates of 5–8%). When PJI is below the discount rate, delay still erodes value even on prevail.

Static E-V Settlement Zone

The settlement zone is the range of dollar amounts at which both parties prefer settlement to continued litigation. It is bounded below by the plaintiff's expected recovery net of plaintiff's costs, and above by the defendant's expected liability plus defendant's costs. The width of the zone equals the sum of both parties' future litigation costs. Fee-shifting (with direction) is folded into the gross EV that drives the bounds.

Prospect theory — what it does

People don't value lawsuits the way a spreadsheet does. Decades of research — borne out in litigation specifically — show the same expected value feels different depending on whether you're chasing a gain or fending off a loss, and whether the key outcome is likely or a long shot. This view adjusts the case value for those documented patterns in how real litigants, including sophisticated represented ones, actually decide. It shows how understandable psychology can lead a plaintiff to take less (or demand more) than the case is mathematically "worth," and a defendant to pay more (or gamble more) than cold math predicts.

The four recognizable instincts. Real litigants pattern into four predictable postures depending on whether the case is in their gain or loss domain and whether the key outcome is likely or rare:

Likely gain → play it safe. Probable winner grabs a sure settlement at a discount (bird in hand).
Long-shot gain → chase it. Small chance at a big award gets over-valued; the long-shot plaintiff demands more than the math justifies.
Likely loss → gamble to avoid it. Probable payer rolls the dice at trial rather than accept a certain loss; the cornered defendant who fights.
Long-shot loss → buy insurance. Small chance of catastrophe gets over-weighted; the defendant who overpays to make tail risk disappear.

Both sides are subject to all four — this is as much a caution about your own tendency to mis-value as a read on the opponent's. In the sidebar, the subjective settlement zone shows how each frame moves the bounds; in some quadrants the floor and ceiling shift the same direction (Q2 lottery + Q4 insurance both push the zone up), in others they compress against each other (Q1 high-prob gain + Q3 high-prob loss tighten the band toward EV minus a discount). Watch where this case sits — the bias pulls you, not just them.

Reference point matters — pick the right premise. The same case under different psychological framings yields different behavioral reservations. A plaintiff who frames the case as "winning something from nothing" sees mostly upside, sits in the gain domain, and tends toward risk-aversion (settles for less). A plaintiff who frames it as "being made whole" after an injury sees everything as a restoration of loss, sits entirely in the loss domain, and tends toward risk-seeking (holds out for trial). The sidebar matrix surfaces these canonical frames so you can identify which fits the actual litigant — or, strategically, which you can induce by how you present the offer (gain-frame the same dollars to invite settlement, loss-frame them to provoke resistance).

Anchoring frame (not modeled, but worth tracking). A fourth common frame, not currently surfaced in the matrix, is relative to a stated demand or offer. Once a party has publicly anchored on a number — plaintiff's $5M demand, defendant's $1M offer — they tend to treat every subsequent settlement number as a gain or loss relative to that anchor rather than relative to EV or to "made whole." A plaintiff who has demanded $5M frames any settlement below $5M as a loss; a defendant who has offered $1M frames any movement up as a loss. This is concession-stickiness in PT terms. The tool doesn't model it because it would require explicit demand/offer inputs that drift quickly during active negotiation; mentally translate by reading the EV-anchored matrix rows and adjusting your sense of the party's position toward their stated number.

Guthrie on negative-EV / nuisance cases. For an improbable frivolous claim (low probability × high stakes), plaintiff sits in the long-shot-gain corner and over-values the shot; defendant sits in the long-shot-loss corner and overpays to extinguish the tail. Both effects compound in the same direction, producing a positive nuisance settlement zone even when static EV is negative — neither party is irrational, they simply have non-EU preferences. For a low-stakes frivolous claim (moderate-to-high probability × small stakes below transaction costs — small-claims/debt-collection territory), prospect theory does not generate a positive nuisance zone; case economics are dominated by transaction costs and the rational response is dismissal or default.

Using the prospect theory panel

Prospect theory is always applied; the Prospect Theory section exposes a behavioral intensity slider (0–10) that controls how strongly the K-T parameters are applied. At intensity 0 the model collapses to EV (no behavioral effect); at intensity 10 (the default) it reproduces the Kahneman-Tversky 1992 canonical population-average values:

λ = 2.25 — loss aversion coefficient (losses feel 2.25× more painful than equivalent gains)
α = β = 0.88 — value function curvature (diminishing sensitivity: the 10th $1,000 matters less than the 1st)
γ⁺ = 0.61, γ⁻ = 0.69 — probability weighting (small probabilities overweighted; γ controls the shape of the weighting curve, with separate values for gains and losses)

The panel reports the PT certainty equivalent (PT-CE) — the certain dollar amount that has the same subjective utility as the modeled gamble. Two decomposition lines below the PT-CE isolate the contribution of each effect: loss aversion contribution (PT-CE minus what PT-CE would be with λ=1) and probability-weighting contribution (PT-CE minus what PT-CE would be with γ=1). These are interaction effects rather than additive components — useful as a directional read on which behavioral mechanism is driving the result, not as a strict algebraic split.

Interpreting the K-T 10.0 result

At the canonical K-T 10.0 setting, the PT-CE reflects how the population-average decision-maker would subjectively value the case, accounting for all three departures from EV. How to read the gap between EV and PT-CE:

PT-CE more negative than EV — the user subjectively over-anticipates losses relative to expected value. Typical for defendants exposed to rare-but-large losses (tail risk in the small-probability loss corner of the four-fold pattern). Implication: defendant's willingness to settle is higher than EV suggests; effective settlement ceiling expands.
PT-CE more positive than EV — the user subjectively overvalues their position. Typical for plaintiffs holding low-probability large-recovery claims (lottery-like long shots). Implication: plaintiff's reservation price is higher than EV suggests; effective settlement floor rises.
PT-CE near EV — the case is in a regime where behavioral effects roughly cancel (e.g., moderate probabilities and balanced payoffs). PT and EV agree on valuation.

Calibration. Intensity 10 reflects population-average bias. Sophisticated repeat-player parties (insurers, large corporate defendants, plaintiffs' firms with portfolio diversification) often exhibit less bias than the K-T population mean — settings of 5–7 reflect "half-to-three-quarters of canonical bias" and are appropriate for institutional litigants. Intensity 0 reproduces EV exactly and is a useful sanity check.

Divergent Expectations Range

George Priest and Benjamin Klein (The Selection of Disputes for Litigation, 13 J. Legal Stud. 1 (1984)) modeled the choice between litigating and settling as a strategic decision driven by the parties' divergent beliefs. If both parties share the same forecast of who will win, settlement is always rational — there's a positive surplus equal to the sum of avoided litigation costs. Trial happens when the parties' subjective probabilities of plaintiff prevailing differ by enough that the plaintiff's expected trial recovery exceeds what the defendant is willing to pay to settle.

Let P_p be the plaintiff's subjective probability of winning, P_d the defendant's, J the judgment amount conditional on plaintiff prevailing, and C_p, C_d each side's remaining cost of going to trial. Trial occurs rationally iff:

P_p · J − C_p > P_d · J + C_d

⇔ (P_p − P_d) > (C_p + C_d) / J

The right side is the litigation threshold — the minimum belief divergence needed to overcome combined trial costs. Cases that settle are those where the parties' beliefs converge enough to make the cost of fighting unworthwhile relative to the disagreement. Cases that litigate are those where the divergence is large enough — usually because there's a genuine factual or legal uncertainty that one side reads more favorably than the other.

The tool's modeled P(plaintiff wins) is your subjective probability; the slider in the diagnostics section lets you set what you believe the opposing party's subjective probability is. Costs in this diagnostic are PV-discounted using each stage's timing but are not probability-weighted — every budgeted dollar is treated as avoidable by settling today.

Counterclaim handling. Priest-Klein is natively a single-claim framework, but generalizes cleanly when a counterclaim is present. Each party faces two sources of value at trial — gain from one claim, loss from the other — and the trial decision becomes:

(P_p − P_d) · J_main + (P_cc,d − P_cc,p) · J_cc > C_p + C_d

Each claim contributes a divergence-weighted value gap to total litigation pressure — the dollar value of the parties' disagreement about that claim's trial outcome. The convention is "the beneficiary's belief minus the other side's belief." For the main claim, plaintiff benefits, so the divergence is P_p − P_d. For the counterclaim, the counterplaintiff (= case-defendant) benefits, so the divergence flips: P_cc,d − P_cc,p. Either value gap can be negative — if one party believes the other side's claim more strongly than the other side does, that claim contributes settlement pressure that offsets the other claim's value gap.

Behavioral caveat. Priest-Klein predicts what's rational under expected-utility maximization with each party's own beliefs. Real cases settle above the threshold for non-modeled reasons (risk aversion, principal-agent pressure, urgency, future cost uncertainty) and try below it for non-modeled reasons (ego, fee dynamics, sunk-cost behavior). The verdict is a directional indicator, not a forecast.

Real options and negative-EV cases

The static EV computation values the case as a one-shot commitment to roll through every stage. In reality, parties can abandon at any stage. The right to walk away has value — particularly when the case might turn adverse upstream, before the most expensive stages are incurred.

The Real options / option premium section computes the case's value to a rational plaintiff who exercises an abandonment option optimally at each stage (Bebchuk-style real-options framework). At each stage, the plaintiff compares two values: continue (pay this stage's cost, observe outcome, then face the next stage's abandonment decision) versus abandon (drop the case, $0 going forward, sunk costs are sunk). Backward induction from the final stage yields the option-adjusted EV (OEV):

V_continue,i = −cost_i + p_advance · V_i+1 + p_exit · exit_value_i
V_i = max(0, V_continue,i)
OEV = V₀

The option premium is OEV − static EV. It is always ≥ 0 — adding the right to abandon can only increase a rational party's expected value. When premium is zero, the optimal policy is "continue through every stage" and the option is not value-additive (typical for moderate-to-strong cases). When premium is positive, there is at least one stage where the rational plaintiff abandons.

Why this matters for negative-EV cases. A case with negative static EV might have positive OEV if most of the probability mass is on early adverse outcomes where the plaintiff abandons before incurring the expensive downstream stages. This is the rational expected-utility justification for litigating a negative-EV case — distinct from the prospect-theory behavioral story. The tool surfaces a callout when OEV is positive but static EV is negative.

For defendants: OEV is always computed from the plaintiff's perspective (defendants don't have a meaningful symmetric abandonment option — defaulting is strictly worse than litigating for a defendant with a non-trivial chance of prevailing). For a defendant user, OEV is informative as the plaintiff's optimal reservation price — the floor below which the plaintiff will not settle. A defendant facing a plaintiff with positive OEV cannot rely on the plaintiff dropping the case even when static EV is negative; the abandonment option gives the plaintiff a positive reservation. Per-stage continue/abandon decisions in the policy table indicate which stages the plaintiff would rationally walk away at — useful for timing settlement offers.

Option-aware settlement zone

The EV-based settlement zone computes the plaintiff floor as expected recovery minus expected cost burn, and the defendant ceiling as expected liability plus expected defense cost. Both assume the case is litigated to verdict on every surviving path. The Option-aware settlement zone section recomputes both bounds under plaintiff's optimal abandonment policy and shows the two bands side-by-side.

Floor delta interpretation. The option-aware floor is the value of the case to a rational plaintiff who can drop his affirmative claim at any stage where forward EV goes negative. It is always ≥ the naive floor (option monotonicity). The floor delta is the dollar value the plaintiff places on retaining strategic abandonment — equivalently, the amount by which a defendant relying on "they'll drop this case" mis-estimates the plaintiff's reservation price. This is the quantity Grundfest & Huang (2006) formalize as the option premium and show is increasing in case-outcome variance.

Ceiling delta interpretation. The option-aware ceiling is the defendant's expected total outlay computed under plaintiff's actual abandonment policy, scenario-by-scenario. For each path the plaintiff would abandon: defendant pays $0 in affirmative-claim damages and incurs affirmative-defense cost only up to the abandonment stage. The ceiling delta is the dollar value plaintiff's affirmative-option is worth to the defendant: avoided judgment exposure plus avoided affirmative-defense spend on abandoned paths. The defendant doesn't have a symmetric defensive abandonment option (defaulting is strictly worse than litigating), but they're a passive beneficiary on the affirmative side. The ceiling is always ≤ the naive ceiling.

Counterclaim option (counter-plaintiff side). When a counterclaim is present, the case-defendant is also the counter-plaintiff and holds a symmetric abandonment option on the cc itself — they can rationally drop the cc when its forward EV after cc-prosecution cost goes negative. The tool computes the counter-plaintiff's cc-policy using per-stage cc-prosecution cost = (cc-prosecution share × defendant stage cost), defaulting to 30% on each cc-bearing stage and adjustable in the Costs section. Independent equilibrium is used (each side's policy treats the other's claim as exogenous), faithful to FRCP 41(c) (the cc survives plaintiff's voluntary dismissal of the affirmative claim).

Settlement point and the bargaining-power slider. The settlement point is α · ceiling + (1 − α) · floor, where α is the plaintiff's share of the bargaining surplus. α = 0.5 reproduces the Nash midpoint (equal bargaining power). α < 0.5 pulls the settlement toward the plaintiff's floor (defendant has more leverage); α > 0.5 pulls it toward the defendant's ceiling (plaintiff has more leverage). Grundfest-Huang Proposition 13 formalizes the disproportionate effect bargaining power has on the realized settlement when the bracket is constructed under option mechanics; the default α = 0.5 is the analytic anchor, not a prediction.

Credibility flag. When the option floor is at or near $0, the abandonment cascade has reached inception — a rational expected-utility plaintiff would not enter the case at all, so they cannot credibly threaten litigation. A rational defendant facing a non-credible threat has expected liability of $0, the option-aware ceiling also collapses to $0, and there is no ZOPA. The naive band's apparent overlap is structurally illusory. In practice, defendants in this regime do sometimes pay nuisance value for non-EU reasons (plaintiff behavioral biases, defendant uncertainty about plaintiff rationality, principal-agent dynamics) — captured by the prospect-theory diagnostic, not this one.

Bibliography

Decision analysis and three-point damages weighting

Pearson, E. S., & Tukey, J. W. (1965). "Approximate Means and Standard Deviations Based on Distances Between Percentage Points of Frequency Curves." Biometrika 52, 533–546.
Swanson, S. (1972). Decision-analysis applications using 0.30 / 0.40 / 0.30 weighting in petroleum reserves estimation.
Keefer, D. L., & Bodily, S. E. (1983). "Three-Point Approximations for Continuous Random Variables." Management Science 29(5), 595–609.
Calihan, R. K., & Victor, M. B. (2004). Three-point P10 / P50 / P90 elicitation methodology for litigation valuation.

Probabilistic forecasting and calibration

Tetlock, P. E. (2005). Expert Political Judgment: How Good Is It? How Can We Know? Princeton University Press.
Mellers, B., et al. (2015). "Identifying and Cultivating Superforecasters as a Method of Improving Probabilistic Predictions." Perspectives on Psychological Science 10(3), 267–281.
Tetlock, P. E., & Gardner, D. (2015). Superforecasting: The Art and Science of Prediction. Crown.

Settlement bargaining and the divergent-expectations tradition

Landes, W. M. (1971). "An Economic Analysis of the Courts." Journal of Law & Economics 14, 61–107.
Gould, J. P. (1973). "The Economics of Legal Conflicts." Journal of Legal Studies 2, 279–300.
Posner, R. A. (1973). "An Economic Approach to Legal Procedure and Judicial Administration." Journal of Legal Studies 2, 399–458.
Priest, G. L., & Klein, B. (1984). "The Selection of Disputes for Litigation." Journal of Legal Studies 13, 1–55.
Bebchuk, L. A. (1984). "Litigation and Settlement under Imperfect Information." RAND Journal of Economics 15, 404–415.
Bebchuk, L. A. (1996). "A New Theory Concerning the Credibility and Success of Threats to Sue." Journal of Legal Studies 25, 1–25.

Prospect theory and behavioral litigation economics

Kahneman, D., & Tversky, A. (1979). "Prospect Theory: An Analysis of Decision under Risk." Econometrica 47(2), 263–292.
Tversky, A., & Kahneman, D. (1992). "Advances in Prospect Theory: Cumulative Representation of Uncertainty." Journal of Risk & Uncertainty 5(4), 297–323.
Rachlinski, J. J. (1996). "Gains, Losses, and the Psychology of Litigation." Southern California Law Review 70, 113–185.
Guthrie, C. (2000). "Framing Frivolous Litigation: A Psychological Theory." University of Chicago Law Review 67, 163–216.
Korobkin, R. B., & Guthrie, C. (1994). "Psychological Barriers to Litigation Settlement: An Experimental Approach." Michigan Law Review 93, 107–192.

Real options in litigation

Grundfest, J. A., & Huang, P. H. (2006). "The Unexpected Value of Litigation: A Real Options Perspective." Stanford Law Review 58, 1267–1336.
Grundfest, J. A., Huang, P. H., & Wu, C. — lognormal extension of the Grundfest-Huang option-premium framework.
Antill, S., & Grenadier, S. R. (2023). "Financing the Litigation Arms Race." Journal of Financial Economics 149(2), 218–242.
Dixit, A. K., & Pindyck, R. S. (1994). Investment under Uncertainty. Princeton University Press.

Reporting frameworks

Financial Accounting Standards Board, ASC 450, Contingencies. — loss-contingency recognition thresholds referenced in the ASC 450 disclosure callout.

To begin, enable JavaScript and reload. Cases save to your browser's local storage; nothing is uploaded.