Audit sunk cost escalation

Audit whether cumulative sunk cost predicts aggregate project continuation after project fixed effects, checkpoint time, forward value, success probability, remaining cost, and future irreversibility, with project-cluster bootstrap uncertainty.

Statistical audit & measurement Audit CTOs Tech CEOs IT managers Team leads Investors

What it's for

Makes portfolio-governance drift visible when continuation becomes increasingly associated with money already spent rather than the remaining forward case, while refusing simplistic blame or automatic cancellation.

What you give it

Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

Field	Type	Role	Required
bootstrap_draws	integer ≥ 200, ≤ 20000	Numerical control	Optional
checkpoints	array of objects (9 fields) ≥ 20 items	Evidence	Yes
confidence_level	number ≥ 0.8, ≤ 0.999	Your calibration	Optional
max_detail_rows	integer ≥ 1, ≤ 500	Numerical control	Optional
minimum_material_probability_effect	number ≥ 0, ≤ 1	Your calibration	Optional
minimum_project_count	integer ≥ 5, ≤ 1000	Your calibration	Optional
ridge_penalty	number ≥ 0, ≤ 1000	Your calibration	Optional
seed	integer ≥ 0, ≤ 4294967295	Numerical control	Optional

Each `checkpoints` record

Field	Type	Required
checkpoint_order	integer (≥ 0)	Yes
continued	boolean	Yes
cumulative_sunk_cost	number (≥ 0)	Yes
expected_future_benefit	number (≥ 0)	Yes
forward_success_probability	number (≥ 0, ≤ 1)	Yes
future_irreversibility_cost	number (≥ 0)	Yes
id	string (non-empty)	Yes
project_id	string (non-empty)	Yes
remaining_cost	number (≥ 0)	Yes

Example input

{
  "bootstrap_draws": 500,
  "checkpoints": [
    {
      "checkpoint_order": 0,
      "continued": false,
      "cumulative_sunk_cost": 20,
      "expected_future_benefit": 160,
      "forward_success_probability": 0.75,
      "future_irreversibility_cost": 10,
      "id": "portfolio-checkpoint-0-0",
      "project_id": "portfolio-project-0",
      "remaining_cost": 100
    },
    {
      "checkpoint_order": 1,
      "continued": true,
      "cumulative_sunk_cost": 35,
      "expected_future_benefit": 150,
      "forward_success_probability": 0.65,
      "future_irreversibility_cost": 15,
      "id": "portfolio-checkpoint-0-1",
      "project_id": "portfolio-project-0",
      "remaining_cost": 90
    },
    {
      "checkpoint_order": 0,
      "continued": true,
      "cumulative_sunk_cost": 25,
      "expected_future_benefit": 160,
      "forward_success_probability": 0.75,
      "future_irreversibility_cost": 10,
      "id": "portfolio-checkpoint-1-0",
      "project_id": "portfolio-project-1",
      "remaining_cost": 100
    },
    {
      "checkpoint_order": 1,
      "continued": true,
      "cumulative_sunk_cost": 40,
      "expected_future_benefit": 150,
      "forward_success_probability": 0.65,
      "future_irreversibility_cost": 15,
      "id": "portfolio-checkpoint-1-1",

Truncated for display — the full payload is 226 lines.

What you get back

This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.

Example output

{
  "assumptions": [
    "Checkpoints are immutable pre-decision records with comparable continue/stop semantics; future benefit, success probability, remaining cost and irreversibility cost were available at that checkpoint and exclude sunk expenditure.",
    "The coefficient is a controlled within-project association in continuation probability per one standard deviation of log sunk cost, not a causal effect, proof of bias, intent, misconduct or an instruction to stop a project.",
    "Forward-economics controls, checkpoint fraction and project fixed effects do not rule out omitted time-varying information, option value, dependencies, contractual constraints, learning, strategic commitments or selection into later checkpoints.",
    "This is aggregate decision-process governance and must not be attributed to named executives, teams or employees; review the decision record and omitted forward considerations before drawing any conclusion."
  ],
  "configuration": {
    "bootstrap_draws": 500,
    "confidence_level": 0.95,
    "controls": [
      "project fixed effects by within-project centering",
      "forward expected economic margin",
      "forward success probability",
      "remaining cost",
      "within-project checkpoint fraction"
    ],
    "log_sunk_cost_mean": 3.8746,
    "log_sunk_cost_standard_deviation": 0.3521,
    "ridge_penalty": 0,
    "seed": 84,
    "sunk_cost_transform": "log1p_then_global_standardization"
  },
  "decision": "no_precision_supported_sunk_cost_escalation_signal",
  "method": "project_fixed_effect_forward_economics_sunk_cost_audit_v1",
  "project_diagnostics": [
    {
      "checkpoint_count": 2,
      "continuation_rate": 0.5,
      "first_forward_margin": 10,
      "first_sunk_cost": 65,
      "last_forward_margin": -7.5,
      "last_sunk_cost": 80,
      "project_id": "portfolio-project-9"
    },
    {
      "checkpoint_count": 2,
      "continuation_rate": 0.5,
      "first_forward_margin": 10,
      "first_sunk_cost": 60,
      "last_forward_margin": -7.5,
      "last_sunk_cost": 75,
      "project_id": "portfolio-project-8"
    },

Truncated for display — the full payload is 134 lines.

How it works

Statistical audit & measurement — Check whether a number is fit to decide on: coverage, timing, reconciliation, and the gaps a dashboard hides.

1 Freeze immutable project checkpoints containing the actual continue/stop decision and only the forward benefit, success probability, remaining cost, future irreversibility and cumulative spend known before that decision.
2 Transform and standardize sunk cost, remove project fixed effects and declared time-varying forward-economics controls with Frisch-Waugh-Lovell residualization, and estimate the remaining within-project continuation-probability association.
3 Bootstrap whole projects, check leave-one-project-out sign stability, require a governed material effect plus a confidence interval above zero, and return only an aggregate governance-review signal rather than attributing intent or ordering cancellation.

Before you trust it

Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.

Assumptions & guardrails

Metric definitions, weights, aggregate grain, sampling, missingness, dependence, and comparison windows correspond to the management claim being audited.
Decision checkpoints are comparable, immutable and not selected according to later outcomes; forward values and constraints are complete, measured before action and exclude already-spent cost; omitted option value or commitments can explain the residual association.
Association, instability, or measurement quality is not a causal effect and must not be converted directly into an individual employment decision.
A positive coefficient is a controlled association, not causal proof of sunk-cost bias, irrationality, misconduct or any named leader's intent, and it cannot automatically stop a project or drive an employment decision.

Minimum evidence

checkpoints: at least 20 rows/items

How to validate it

Validate on future periods or held-out aggregate units, compare with a simple baseline, and require stability across plausible metric definitions and decision thresholds.

Calibrating it to your org

Same for everyone

The mathematical kernel, validation rules, method version, and JSON output semantics are organization-independent; no tenant-trained coefficients or company benchmark is embedded in the function.

Specific to you

point-in-time forward economic margin and comparable within-project checkpoint panel preserving every eligible decision, including stops and unresolved projects, without rewriting forecasts from later outcomes
project and checkpoint eligibility, continuation semantics, finance perimeter and currency, forward-value horizon, probability vintage, option/dependency/contract treatment, material probability effect, confidence, ridge, bootstrap, privacy boundary, and review ownership

Calibration workflow

1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.

Call it from your AI

You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.

gitrevio_capabilities_search
  { "q": "audit whether cumulative sunk cost predicts" }
  → finds "audit_sunk_cost_escalation"

gitrevio_capability_describe
  { "capability_id": "audit_sunk_cost_escalation" }
  → returns the input schema and agent guidance shown on this page

gitrevio_capability_run
  { "capability_id": "audit_sunk_cost_escalation", "arguments": { ... } }
  → returns the result shown above

Works in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.

Related tools

Calculate earned value forecast

Turn period-level planned value, accepted earned value, and actual cost into a correlated Bayesian CPI/SPI distribution for final cost, completion period, budget overrun, and deadline miss, with classical EAC cross-checks and an early-progress abstention gate.

Sequential Bayesian & bandits

Optimize reliability investment frontier

Construct a dependency- and exclusion-feasible Pareto frontier across investment cost, expected residual loss, CVaR loss, and expected downtime under coherent scenarios and explicit pair interactions, then select the least-cost evaluated portfolio clearing governed reliability targets.

Forecasting & survival

Audit commercial technical commitment integrity

Audit signed commercial promises against explicitly allocated technical scope, dependency order, funded capacity, acceptance criteria and evidence; expose orphan scope, double allocation, cycles, late plans and maximum contractual penalty without interpreting legal rights from engineering activity.

Constrained optimization

Audit decision execution fidelity

Audit whether approved decisions actually became verified implementation at the promised aggregate-unit and component grain, with whole-unit bootstrap uncertainty and simultaneous gates for fidelity, overdue scope, unverifiable evidence, exceptions and critical gaps.

Statistical audit & measurement

Audit KPI threshold bunching

Detect a post-target excess concentration immediately above a governed KPI threshold: compare within-unit pre/post local mass and above-versus-below mirror asymmetry, bootstrap whole units, report density bins and a smoothed log-density jump, and explicitly refuse to equate bunching with individual gaming or intent.