The Group Decision Support System problem

The Missing System Theory states that the ordinary two-party payoff space does not carry the system-welfare coordinate W. This is a claim about the structure of the payoff space itself, rather than computational difficulty or data availability. The ordinary payoff space records the parties to a transaction, and W is not a function of those parties' payoffs. No algorithm operating on party utilities and issue weights can recover it. The blind spot is therefore a property of the space the platform is designed to explore, not a defect in the platform's code. The Information-Exclusion paper extends this result to the ledger¹¹: W is absent from the disclosed records firms and markets report, by the same construction that keeps W out of the payoff space. The off-ledger status of system welfare creates a specific informational friction. Because β_W is part of no mandatory or customary disclosure, it is not reflected in equilibrium prices, and no ex-ante premium forms for system-welfare risk.

The independent W-monitoring channel is a central design change. Current platforms derive every displayed value from party inputs. A W-aware platform must instead accept a W-signal from a source external to the negotiating parties, a signal that is not derived from their payoffs and not correlated with them by construction. For financial benchmarks, the independent channel could be a benchmark administrator's rate-submission process. For environmental commons, it could be a third-party environmental monitoring service. For labor markets, it could be an independent workforce welfare index. For public health, it could be an independent epidemiological monitor of mortality and morbidity. For financial stability, it could be a supervisor's systemic-risk indicator. For digital platforms, it could be an independent data-protection and online-harm monitor. For consumer credit, it could be an independent borrower-welfare measure of default and re-borrowing harm. Decision Accounting makes the signal reportable through Field 17 of the decision record. The report-incentive condition pL ≥ κ ensures that the signal can be truthful when the detection probability is set by a multi-audience Conflictoring protocol rather than by a single, predictable evaluator. The point of externality here is precise: if the W-signal came from the parties, they could adjust it the way they adjust any other input.

R2 is the pre-decision Decision Accounting record that carries the system-welfare consequence alongside the parties' outcomes. The 17-field record gives the decision a place to state alternatives, make a scoreable prediction, and record the system consequence before commitment: Field 15 is ALTERNATIVES, Field 16 is PREDICTION, and Field 17 is SYSTEM WELFARE. The record names the system boundary, evidence, uncertainty, time horizon, and review trigger. A system-adjusted payoff display can then show Π_sa = Π − μ · ΔW, where μ converts welfare units into the private-payoff numeraire (the common unit—here dollars—that other values are converted into for comparison) and ΔW is the measured change in system welfare. The display makes the consequence visible; it does not prove the prediction or prescribe a choice. Decision Accounting makes the record auditable so a later reader can compare the prediction with the outcome and inspect how the decision was made.

Trajectory detection adds a temporal dimension. R3 computes T^∗, the model-derived crossover horizon at which accumulated system loss catches up with private gain under stated parameters. In the linear model, T^∗ = δ/(ηλ): δ is initial private surplus in dollars, λ is the annual system-loss rate in dollars per year, and η is the fraction of that loss that feeds back into the private payoff. Write the units beside the parameters before calculating; δ > 0, λ > 0, and 0 < η ≤ 1 are required for a finite positive result. If η = 0, this linear formula has no finite crossover. A Hollow Win can persist for a period before failing, so the platform should display the calibrated horizon and its uncertainty alongside the current classification. Resilience, decay, or a sudden shock requires the generalized Appendix D equation. T^∗ is a model-derived horizon for a stated calibration, not a calendar date or a collapse guarantee.

The claim that the fix is a software change rests on two premises. First, the platform already has the computational infrastructure to accept inputs and display outputs, so adding a W-signal input field and a system-adjusted payoff display row is a user-interface change rather than a new platform architecture. Second, the W-signal does not need to be computed by the platform. It can be accepted from an external source, which means the platform's role is to carry the coordinate, not to derive it. Decision Accounting makes the signal reportable and verifiable, and Conflictoring supplies the incentive structure for truthful reporting. Neither requires new legislation. Both require a design change to the platform's input and output space. The minimum viable product, accept a W-signal, compute the system-adjusted payoff, display the result, and flag agreements where the system-adjusted payoff is negative, can be built within a normal software release cycle rather than a regulatory one.

Pigou, Coase, and Ostrom are three restorations, each an attempt to bring the system coordinate back into the analysis through taxes, property rights, or common-pool governance¹², but each is partial: it addresses one mechanism of system degradation without changing the underlying structure of the payoff space (see the theorem chapters for the full relationship to Nash, CAPM, and the welfare theorems). The platform fix described in this chapter is a structural intervention, because it changes the payoff space itself by adding the W coordinate. That is different from a Pigouvian tax, which changes prices within the existing space¹³, and from a Coasian bargain, which reassigns property rights within the existing space¹⁴.

A boundary case arises when the negotiation platform itself is part of the system that could be degraded. A platform that facilitates high-frequency trading in financial markets, for example, may be the infrastructure through which system welfare is lowered. In that case the platform's own design choices, such as latency, order types, and fee structures, affect W. The platform cannot accept a W-signal from an external source, because the platform is the source. This case requires a different intervention: the platform must compute its own β_W and display it to users. The Decision Accounting framework still applies, since the platform's own decision record must include Field 17. The Information-Exclusion paper notes that the same friction that excludes W from the ledger also excludes it from the platform's own design space. The fix requires the platform to treat itself as a studied domain rather than a neutral conduit.

RECALL — see the theorem chapter: W is not a function of the parties' payoffs. A standard group decision-support system (GDSS) transforms party utilities and issue weights; it cannot derive W from those inputs. A W-aware platform therefore has three separately testable requirements: - R1 — independent system-welfare monitoring channel. Obtain the system-welfare signal through a measurement process outside the parties' payoff reports. Name the source, boundary, period, uncertainty, and audit path. A benchmark-integrity series, emissions monitor, public-health measure, or infrastructure-reliability record can serve as the signal; a party's self-report cannot. - R2 — pre-decision Decision Accounting record. Before commitment, write the system-welfare consequence alongside the parties' outcomes in the 17-field record. Field 15 is ALTERNATIVES, Field 16 is a scoreable PREDICTION, and Field 17 is SYSTEM WELFARE. Record the boundary, evidence, uncertainty, time horizon, and review trigger. A system-adjusted-payoff display such as Π_sa = Π − μΔW can make the record legible. Here μ is the stated conversion weight that expresses the welfare-loss measure in the private-payoff numeraire; it is a modeling choice that must be declared and calibrated. The display is an implementation detail, not R2 itself. Recall link: the Decision Accounting chapter owns the field-by-field record. - Notation boundary: in this display, Π is the private payoff or return term being adjusted, while ΔW is the system-welfare change. the measurement chapter uses Π for annual industry revenue in average β_W; check the declared boundary before transferring the symbol. - R3 — trajectory detection. Estimate when accumulated system loss overtakes private surplus. In the linear model, T^∗ = δ/(ηλ), where δ is initial private surplus, λ is the annual system-loss rate, and η is the fraction of system loss that feeds back into the private payoff. Under full feedback η=1, T^∗=δ/λ. For an illustrative case with δ=$20M, λ=$5M/year, and η=1, T^∗=4 years: at that horizon, ηλT^∗ = 1 × ($5M/year) × 4 years = $20M, so the accumulated feedback equals the initial surplus. When η=0, the linear model has no finite crossover because system loss does not feed back into private surplus. Each parameter must be calibrated and carry units before the result is used. With resilience or decay, solve the generalized crossover equation from the formal route. T^∗ is a parameter-dependent horizon, not an observed date or a guarantee. The three requirements answer different questions: R1 measures the omitted system state through an independent channel, R2 records its consequence beside the parties' outcomes before commitment, and R3 computes how the calibrated trajectory changes the private/system balance over time. The record and display inform a decision; they do not prove the prediction or compel a choice. Worked (illustrative). The deal offers $100M in private gain (Π=$100M) but destroys $250M of system welfare (ΔW=$250M). At μ=1, the system-adjusted payoff is Π_sa = $100M − $250M = −$150M. The negative result is the decision: the deal takes $150M more out of the shared system than it delivers to the parties, so on these numbers it should not go forward. Computing Π_sa does not create the harm — the $250M loss was already there; it only puts that loss in front of the decision-maker at approval, where a payoff-only screen would have shown all green. The constructive response is to reduce the system-welfare damage — redesign the deal, add safeguards, or compensate the harm — until Π_sa is no longer negative, and only then renegotiate and proceed.