A supplier offers to settle your claim at a discount. Accepting feels like losing and litigating feels like winning — both instincts are wrong. Here is the equation that actually decides it, input by input, from a lawyer with an accounting degree.
The email arrives on a Friday: "We value the relationship and would like to put this behind us. We can offer 70% of the amount outstanding, paid within 30 days of settlement."
Most claimants answer that email with instinct. Either pride — "I'm right, I'll take 100% or see them in court" — or fear — "money on the table, take it before it disappears." My training was in accounting before law, and I confess both instincts irritate me, because the question has an actual structure. A discounted settlement is an offer to buy your litigation risk. Whether the price is fair depends on four variables you can — and should — estimate before replying.
Stripped of rhetoric, the decision compares two things:
Settlement value = the discounted amount on the table, adjusted for when and whether it actually gets paid.
Litigation value = (probability of winning × amount realistically enforceable) − remaining costs of the fight − the time cost of getting there − the residual risk that even a win collects nothing.
Take the settlement when the first number comfortably exceeds the second; decline it when it doesn't. That is the entire framework. The rest of this article is about estimating each input honestly, because every bad settlement decision I have seen was a bad estimate of one of these inputs, not a bad equation.
Before the inputs, one structural point: a settlement is a transaction, so it has terms beyond the headline number. When you value an offer, look at the whole package — payment timing and installments (a 70% offer in three tranches over a year is not a 70% offer), who pays first (cash simultaneous with signature, or your release in exchange for a promise?), the scope of any release you give (this claim only, or "all claims arising from the relationship"?), and what happens on default (a consent-to-judgment clause or a penalty makes a settlement self-enforcing; a bare payment promise just starts a new dispute). Two offers with identical percentages can differ by half in real value. The equation below prices the headline number; your contract discipline prices the rest.
"We have a strong case" is not an input; it is a mood. What actually moves win probability is the tier of your evidence:
Two adjustments matter in cross-border claims against Chinese suppliers. First, evidence you intend to use in a Chinese court generally needs to be properly formalized — foreign documents typically require notarization and legalization or apostille, and foreign-language evidence requires translation; budgets and timelines must assume it. Second, your own conduct is evidence: a documented history of demands, partial payments, and recorded promises strengthens the story; years of WhatsApp informality weaken it. When I assess probability for a client, I am really grading the file against these tiers — and the grade predicts not just litigation but how a serious supplier-side lawyer will grade it too. That matters, because settlement pricing is set by their estimate of your case as much as yours.
This is the input most claimants get wrong, and it is the one my accounting background won't let me round off. A judgment for the full claim is a legal statement; what you can actually collect is a balance-sheet fact. The gap between the two is the collection risk, and in disputes with Chinese suppliers it is driven by:
So the input isn't "amount claimed" — it's amount claimed × realistic recovery rate. A full-value judgment against a shell is worth less than a 70% settlement from a company with visible cash flow. Say that plainly and half the emotional confusion around settlement evaporates: you are not comparing 70 to 100. You are comparing 70 to a probability-weighted fraction of 100.
The cost input that matters is remaining cost — sunk costs are sunk, however annoying they were. Depending on posture, the remaining column includes: court fees (which in China scale with claim size), preservation fees and any required security, translation and notarization of evidence, counsel time through first instance and potentially enforcement, and travel or local representation. Against these, a settlement today costs almost nothing to finalize — a settlement agreement, perhaps a mutual release, done.
The asymmetry matters most in mid-sized claims: where the absolute cost of full litigation is a material share of the amount in dispute, the settlement discount is competing against a cost base that eats a real percentage of any recovery. On very small claims, litigation is arithmetic nonsense and even a steep discount can be the rational answer; on very large ones, the fixed costs are proportionally minor and the case should usually be tried unless the merits are genuinely doubtful. The uncomfortable middle is where the math deserves actual computation.
Two further inputs, both routinely ignored.
Time. A judgment recovered in two years is worth less than cash in 30 days — by your cost of capital, by inflation, and by the risks the world injects into a two-year window: the company deteriorates, the market turns, the assets evaporate. The longer the realistic path to final recovery, the higher the discount you should rationally accept. "70 cents today versus 100 cents eventually" is never the real comparison.
The counterparty's pulse. This is the signal I weigh most heavily when advising on settlement, and it comes straight from credit practice: a supplier that is still taking orders, still shipping, still hiring, is a supplier with cash flow — and a cash-flow business that wants this dispute off its desk is the best settlement counterparty in existence. They are buying peace at a discount because operations are worth more to them than the discount. By contrast, a supplier that has stopped production, laid off staff, stopped answering new customers — the discount it offers is not generosity, it is its last liquidity, and "pay 70% in 30 days" from a dying company is a promise you will be re-litigating with an insolvency administrator. Same number on the paper, opposite answers to the equation. Read the vital signs before you price the offer: recent shipments, new product listings, an active workforce, answers that arrive within days rather than weeks.
| Step | Question | What goes in |
|---|---|---|
| 1. Grade the evidence | What tier is my file — dispositive, contested, or oral? | Assign a realistic win-probability range. If you can't defend the range to a skeptical colleague, it's a wish, not an estimate. |
| 2. Size the collectable amount | If I win, what is actually there? | Registered capital, asset map, structure risk. Amount claimed × realistic recovery rate. |
| 3. Net out remaining costs | What does the rest of the fight cost? | Court fees, preservation, formalization of evidence, counsel, enforcement. Subtract from step 2. |
| 4. Discount for time | When does each path actually pay? | Settlement in 30–60 days vs. realistic litigation-plus-enforcement horizon. Apply your cost of capital, plus deterioration risk over the window. |
| 5. Read the vital signs | Is this counterparty alive and trading? | Still taking orders → discount is credible and negotiable. Contracting operations → take the promise less seriously than the number. |
Then compare: the offer, adjusted for payment certainty and timing, against the litigation value from steps 1–4. Decide. And whatever you decide, write it down with the numbers — because the discipline of writing the numbers down is what stops the Friday-afternoon instincts from making the decision instead.
None of these inputs is a measurement. Win probability is judgment; recovery rate is judgment; the counterparty's remaining life is a forecast. Two competent lawyers can run the same file and land on meaningfully different numbers — I know, because I have argued both sides of that conversation. The framework's value is not that it turns disputes into arithmetic; it is that it forces every assumption into the open where it can be argued, tested, and revised, instead of hiding inside feelings about fairness or fear.
The pattern in practice: claimants overpay for vindication and underpay for risk. They refuse credible discounts because 100% feels like justice, and they accept hollow promises because a number on paper feels like resolution. The equation is the antidote to both. And note the timing wrinkle: a settlement conversation works best before litigation starts — after a claim is filed, the supplier's lawyer's advice gets defensive, and the discount window narrows. If the math says settle, the moment to move is usually now; a formal demand on the record can open the conversation without surrendering anything, which is the role the structured demand letter plays in the sequence. If the math says fight, then fight deliberately — starting from a collectability screen, through preservation and litigation, and out the far end at enforcement of the judgment, where the collection assumptions in this equation finally get graded.
A final word on the negotiation itself: the equation doesn't just tell you whether to accept — it tells you what a rational counteroffer looks like. If your litigation value works out to, say, the low seventies as a percentage of the claim, an opening offer of 70% is not an insult to be refused; it is an invitation one round away from your number, and the correct move is a counter anchored slightly below your walk-away arithmetic with the non-price terms tightened — faster payment, this-claim-only release, default consequences. If your litigation value works out to forty, the right answer is to close at 70% quickly and gracefully. Parties who negotiate every offer down to the last point regardless of the math are paying litigation prices for settlement outcomes, just with extra months attached.
Seventy cents on the dollar is neither victory nor defeat. It is a price. Price it against the math, not the mood — and the math starts with knowing what the counterparty's money actually looks like.
This article is general information, not legal advice, and does not create an attorney–client relationship. Outcomes vary by case; nothing here is a guarantee of results.
Send the file and the offer. We'll run the five steps on your actual numbers — evidence tier, collectable base, remaining costs, counterparty vital signs — and give you a defensible answer: settle, or fight, and why.
Run the numbers with us