casesim-settlement-vs-trial-ev-calculator

Category: General Risk: Unknown ★ 3.9 · Rating 3.9/5 (8) sboghossian/mini-claude-for-legal MIT

Rating is derived from the repo's GitHub stars and shown for reference.


name: casesim-settlement-vs-trial-ev-calculator
description: Use when an attorney or client needs to compare the expected value of accepting a settlement offer against the expected value of proceeding to trial — incorporating P(win), damages range, costs to trial, loser-pays rules, discount rate, and strategic or reputational factors. Produces an EV comparison (trial vs. settlement), a recommendation (take/counter/decline), and a sensitivity analysis showing which assumptions matter most. Always disclaimed as a probabilistic estimate. Always paired with the outcome probability estimator.
license: MIT
metadata:
id: casesim.settlement-vs-trial-EV-calculator
category: casesim
jurisdictions: [multi]
priority: P1
intent: [settlement, ev, client-counseling, litigation-strategy]
related: [casesim-outcome-probability-estimator, casesim-opposing-counsel-simulator, casesim-judge-bench-perspective, casesim-client-q-and-a-prep]
source: Louis — HAQQ Legal AI (github.com/sboghossian/mini-claude-for-legal)
version: "1.0"

Settlement vs. Trial EV Calculator — Should You Take the Offer?

When to use this

Invoke when:

  • Opposing counsel has made a settlement offer and the client asks "should we take it?"
  • A litigation team is preparing for a settlement conference and needs an internal floor / ceiling analysis
  • A client's board asks for a "go / no-go" recommendation on litigation versus early resolution
  • An attorney needs to explain the settlement vs. trial decision in quantitative terms to a client who wants numbers
  • A funding decision for continued litigation needs to be made

This skill works alongside [[casesim-outcome-probability-estimator]], which generates the P(win) and damages range inputs. Use both together for a complete client counseling package.

Inputs

Input Required Notes
Settlement offer amount Yes The net figure the client would receive (or pay)
P(win at trial) Yes From [[casesim-outcome-probability-estimator]] or attorney's own assessment
Damages range at trial (low / mid / high) Yes If liability established, what would the court award?
Costs to trial (own side) Yes Estimate to judgment; separate from costs already spent
Loser-pays rule Yes Does the forum apply loser-pays (English rule) or each side bears its own costs (American rule)?
Discount rate Optional Default: 5–8% depending on jurisdiction and client risk profile
Time to judgment Optional Months from today to expected judgment date
Strategic / reputational factors Optional Confidentiality value, business relationship, precedent-setting risk, management distraction
Appeal probability Optional Will opposing counsel appeal a first-instance win?

Calculation methodology

EV(Trial) — Expected Value of Going to Trial

EV(Trial) = [P(win) × E(award | win) - Costs(own)]
            + [P(lose) × -Costs(own)]
            ± [Loser-pays adjustment]
            × [Time discount factor]

Where:

  • E(award | win) = probability-weighted average across the low / mid / high damages scenarios
  • Costs(own) = estimate of legal fees and disbursements to trial
  • Loser-pays adjustment:
    • If win: add recovery of opponent's costs (partial — typically 60–70% recovery in English-rule forums)
    • If lose: subtract payment of opponent's costs
  • Time discount: apply the discount rate over the expected time to judgment

Example:

Variable Value
P(win on liability) 65%
Expected award (mid scenario) USD 800,000
Costs to trial (own) USD 120,000
Loser-pays rule English rule (DIFC)
Opposing costs if lose USD 100,000
Discount rate 6%
Time to judgment 18 months

EV(Trial) = [0.65 × (800,000 + 0.65 × 100,000) - 120,000]
+ [0.35 × -(120,000 + 100,000)]
discounted at 6% over 18 months
≈ [0.65 × 865,000 - 120,000] + [0.35 × -220,000] × 0.915
= [562,250 - 120,000 - 77,000] × 0.915
= 365,250 × 0.915
≈ USD 334,200

EV(Settlement) — Expected Value of Accepting Settlement

EV(Settlement) = Settlement amount - Costs already spent (sunk) + [Optional: value of strategic benefits]

Note: sunk costs do not affect the forward-looking decision. The only relevant comparison is future EV(Trial) against the net benefit of settlement from today. Costs already spent are the same whether the client settles or litigates; they do not change the calculus.

However, where the client is framing the question as "what do I net from this whole matter?" — include sunk costs in the total cost presentation (not in the EV comparison but in the client narrative).

Net Comparison

Settlement advantage / disadvantage = EV(Settlement) - EV(Trial)
  • Positive: settlement is worth more (in expected value terms) than going to trial
  • Negative: trial has higher expected value; settlement offer is below fair value
  • Near zero: the decision turns on strategic / non-quantitative factors

Sensitivity analysis

The most important output after the EV comparison is the sensitivity analysis: which assumption matters most?

Louis generates a table showing how the recommendation changes as key variables shift:

Variable Base case Sensitivity test Effect on recommendation
P(win) 65% 55% / 75% At 55%, settlement more attractive; at 75%, trial clearly preferred
Costs to trial USD 120k USD 80k / USD 180k Higher costs make settlement more attractive
Time to judgment 18 months 12 months / 30 months Longer time = greater discounting = settlement more attractive
Award (mid scenario) USD 800k USD 600k / USD 1M Affects the upside case significantly

This tells the attorney: "The decision is most sensitive to P(win). If we have reason to believe our win probability is above 70%, we should push back on this offer. If it's below 60%, the settlement looks fair."

Recommendation framework

Based on the EV calculation and sensitivity analysis, Louis outputs one of:

TAKE — settlement offer exceeds or approximates EV(trial)

  • Recommended when: settlement is within 10–15% of EV(trial) and strategic factors favor resolution
  • Key message to client: "The offer is fair value. The certainty of settlement is worth the small discount versus the expected trial outcome."

COUNTER — settlement offer is below EV(trial) but trial is not clearly superior

  • Recommended when: settlement offer is 20–40% below EV(trial) but costs, time, or strategic risks make trial unattractive
  • Key message to client: "Push back. Make a counter at [X], which reflects a reasonable settlement range. Trial is an option but not our preference given [costs / timing / relationship]."

DECLINE — EV(trial) materially exceeds the settlement offer

  • Recommended when: settlement offer is more than 40% below EV(trial) and the case is strong
  • Key message to client: "This offer does not reflect the value of your claim. Proceed to trial unless opposing counsel substantially improves their position."

Strategic and non-quantitative factors

The EV calculation captures the financial dimension. These factors may change the recommendation:

Factor Effect
Confidentiality If reputational or commercial harm from a public trial is material, settlement has a premium beyond its financial value
Business relationship Settlement may preserve a commercial relationship that has ongoing value
Precedent / principle A public judgment may be worth litigating for even if the EV is comparable — signals to other potential claimants or deters future breach
Management distraction Litigation absorbs executive time; cost is real but hard to quantify
Counterparty financial risk Is the counterparty likely to be able to pay a judgment? An insolvent win has limited value

Mandatory disclaimer

Every output of this skill must include: "This expected value analysis is a probabilistic planning tool, not a prediction of actual outcomes. Legal proceedings involve inherent uncertainty. All assumptions should be reviewed with qualified legal counsel before making any decision."

  • [[casesim-outcome-probability-estimator]]
  • [[casesim-opposing-counsel-simulator]]
  • [[casesim-judge-bench-perspective]]
  • [[casesim-client-q-and-a-prep]]