CS Paper Radar

Strategic Technical Debt: A Real Options Approach to Early-Stage Software Experimentation

Rashid Azarang, Mohammad Reza Azarang Esfandiari

cs.SE ·submitted 2026-08-17 05:05 UTC · v1
theory

Abstract

Technical debt is treated almost universally as an engineering pathology. This paper argues that under the conditions defining early-stage software work (high hypothesis uncertainty, cheap experiments, and the freedom to abandon), deliberately incurred technical debt is a rationally priced financial instrument: a call option on the validated product, purchased at a discount that is largest exactly when uncertainty is highest. We make three contributions. First, a demarcation: debt is strategic when its expected cost loads on the success branch of the venture (repaid only if the hypothesis validates) and toxic when it imposes unconditional cost while held (security exposure, data loss, corrupted experimental signal), a boundary stated formally that renders the popular "prudent vs. reckless" intuition testable. Second, a sequential model: a finite-horizon dynamic program over belief and debt stock yielding four results: a shadow price of debt equal to the risk-discounted probability of repayment; a technical-debt overhang (the belief threshold for scaling rises with the debt stock, proved via an envelope Lipschitz bound); a refactoring-pivot theorem (optimal repayment concentrates at the commitment boundary, predicting the practitioner-reported refactoring burst at product-market fit, registered here as a falsifiable prediction); and a volatility result under risk-neutral valuation. A pivot-salvage correction shows the folk rule "maximum debt at maximum uncertainty" fails whenever failure redirects rather than terminates the venture and the salvage differential clears the discounted cost premium. Third, a two-test primary empirical program (validation-event refactoring timing; the first repository measurement of pivot salvage), pre-registered with frozen analyzers before any data contact. Calibrations are illustrative, not estimates.

Comment
25 pages, 4 figures. Pre-registered empirical program: OSF bs3cr (EP3'), rvx5t (EP-Pi)

Closest papers

Nearest neighbours by meaning across the whole index.

The Data Debt of Stockouts: Partial Identification and Minimax-Regret Inventory Control

cs.AI eess.SY ·2026-11-14 ·Jinbiao Chen, Shuang Jin, Hanzhang Qin
similarity 0.57
Abstract

Problem definition: Inventory decisions select which demand enters sales logs. Unrecorded forecasts can leave demand frequencies unidentified even when stocking levels are recorded. We study which multistage ordering decisions these logs can justify and what finite samples can guarantee. Methodology/results: Nested stockout thresholds yield exact sets of compatible demand laws and finite representations for policy evaluation. We choose one deterministic inventory rule to minimize its largest expected cost gap from the informed optimum under the same law. This population value is data debt for the specified information and policy class. Under independent demand, carryover can create positive debt despite common optimal actions in every isolated period. Finite samples add learning error above debt. The required sample count grows linearly with the selection bound, with upper and lower orders matching up to logarithmic factors. This rate requires fixed demand and cost bounds, sufficient revelation, and a uniformly positive probability of clearing all feasible stock across the confidence sets. Controlled and semi-synthetic studies separate policy improvements from tighter guarantees. Managerial implications: More records can tighten guarantees while debt persists. Deeper revelation can establish common optimality without fully identifying demand. Using these guarantees requires validated logging, a credible selection range, and a regret tolerance in cost units. Information acquisition must distinguish tighter assurance from operating savings.

0.04

Restraint-Induced Loss in Software-Intensive Enterprises: Large A Verified Technical Economic Impact Model of Operational Downtime and Delayed Commercialization, with Tail-Risk Security Sizing and Three Enterprise Case Studies

cs.CE ·2026-10-20 ·Mahdi Eslamimehr
theory similarity 0.53
Abstract · 53 pages

When an enterprise must stop using a software component (for example, because of an interim court restraint in a technology patent dispute), the economic loss passes through a chain of technical mechanisms. Withdrawing the component exposes dependent capabilities. Fallback paths degrade throughput, correctness or feature coverage. Incident rates and recovery backlogs rise. Customers respond. Remediation programmes compete against the restraint clock, and planned launches are pushed back. This paper develops and verifies a \emph{technical--economic impact model} (TEIM) that quantifies this chain end to end as a probability distribution of present-value loss rather than as a point estimate. The technical layer combines (i)~a lattice-valued, fixed-point semantics of service dependencies with degraded fallbacks, (ii)~continuous-time Markov reliability models evaluated by uniformization and specified in continuous stochastic logic, and (iii)~a fluid queueing model of post-outage backlog. These feed (iv)~convex acute-outage loss functions with regulatory and contractual thresholds, (v)~compound-Poisson excess-incident processes, (vi)~survival-weighted functional, attrition and funding channels that provably avoid double counting, (vii)~a real-options valuation of forcibly delayed launches with an explicit closed form, and (viii)~a COCOMO~II--driven stochastic activity network for design-around and substitution strategies. We prove seven theorems and eight propositions.

0.29

How Far Can an Agent Push? Trust-Safety Margins in Live Language-Based Economic Games

cs.MA cs.GT ·2026-10-18 ·Dima Galat, Marian-Andrei Rizoiu
benchmark similarity 0.52
Abstract

Game theory identifies when a strategic action should be accepted, but not how long a live, language-based counterpart will continue to accept it. We entered the GLEE competition in its final week with one protected agent, Milton, and four instrument agents that ran deterministic, per-game randomized trials of policy changes; Milton received changes only through pre-specified gates or registered, rollback-guarded production rules. In negotiation, learned pricing improves adjusted percentile by 0.071 (n=21,168); in bargaining, a continuation model conditioned on the rejected offer did not improve on a recency-weighted incumbent and was retired. In persuasion, a preregistered five-dose experiment (7,391 assigned games, blinded until a fixed stop) shows that recommending every unit raises sales by 0.194 of the maximum (95% CI [0.147, 0.241]) where the receiver's slack makes obedience rational (52% of the Bayesian benchmark), lowers them where it does not (-0.067), and produces a larger early-to-late purchase decline than honest recommendations (-0.123). A rule fixed before unblinding now routes Milton's dose per cell. Milton stood 11th of 460 agents in bargaining at the fixed stop; we also report the close, where the final day's pool shift moved all five agents' bargaining ratings together while their play was unchanged, and three dispatch defects were found by checking delivered behavior against the design.

0.13

JUBILEE: A Benchmark Environment for Off-Policy Evaluation in Debt Collections

cs.LG q-fin.RMstat.ML ·2026-11-07 ·Diego Silva Prado, Tiago Amaral Penteado de Jesus, Bruno Caldas de Souza, Cristiano Bevitori Maffia De Oliveira
benchmarkcode/data similarity 0.52
Abstract · 28 pages, 3 figures, 35 tables. Code: https://github.com/alpha-synthesis/jubilee. Data: https://doi.org/10.5281/zenodo.24008544

Off-policy evaluation (OPE) estimates the value a candidate decision policy would achieve, using only data logged by the policy already in production. In debt collections, unrecorded factors can influence both actions and outcomes. This paper presents JUBILEE, a benchmark environment built around a debt-collections simulator that generates million-loan panels with competing risks, rare recovery events, and hidden factors driving both outcomes and assignment. The evaluation includes an 80-run comparison of tabular and gradient-boosted estimators under two synthetic outcome mechanisms and a 190-run parameter grid. Independent policy simulations provide reference values with estimated Monte Carlo uncertainty for a synthetic reward score and for simulated cash. With hidden factors influencing action assignment, several tabular methods agree closely while underestimating the field-visit policy's score, an error that also appears with boosted models. Intervals from tabular doubly robust estimation usually contain the reference without hidden assignment dependence but mostly miss it when that dependence is present. Estimates from randomized decisions do not show the same large error, although their limited precision leaves policy choice uncertain. The decision consequences depend on how policy value is defined. The field-visit policy ranks last under the score and first under simulated cash, which excludes channel costs. Under cash, log-based estimators miss this best candidate in nearly every run with hidden assignment dependence in the paired study. JUBILEE supports reproducible study of these failure modes under specified collections mechanisms.

0.60

Architectural Debt Enforcement under Incomplete Observation: A Controlled Study of Ratchets and Evidence Qualification

cs.SE ·2026-10-30 ·Zhiheng Zhang
similarity 0.52
Abstract · 15 pages, 9 tables. Includes appendices

Architectural debt gates must distinguish an allowed change from an incompletely observed one. We study a decision contract separating observation qualification, prohibited relation identity, inherited debt and authorized baseline updates. Across 240 fixed files from four repositories, the current observer and programmatic reference jointly qualify 166 files (69.17%), all with exact reference-relative edge agreement; 74 remain reference-UNKNOWN. Coverage varies from 13/60 to 59/60 by repository. On 480 constructed state scenarios sharing 60 source anchors, both the candidate ratchet and mature dependency-cruiser baseline match every expected outcome, whereas static and count-only baselines match 420 and 360. These results do not establish candidate superiority. Preserved failures include nine additional false-PASS probes and 120 unqualified full-tree timing attempts. A subsequent generic binding layer rejects 54 specified mismatches, accepts 54 current controls and permits legal repetition. Two bounded package scopes support 40 scenarios and 48 fresh-process functional checks; their costs remain tied to the measured implementations. Diagnostics explain residual uncertainty through type/runtime relations, installation context and generated-artifact provenance. TypeScript 5.9.3 and 6.0.2 yield identical outputs on ten paired emission witnesses, but neither that agreement nor verified release provenance changes the frozen UNKNOWN labels. The contribution is a controlled account of evidence-qualified enforcement and its limits, not language completeness, natural maintenance effectiveness or a new shrinking-baseline algorithm.

0.34

Strategic Repayment and Latent Fragility in Financial Networks: Why Banks Withhold Payable Liquidity

cs.GT q-fin.RMecon.THq-fin.MF ·2026-11-16 ·Hicham El Kaissi
similarity 0.51
Abstract

A network in which every bank pays in full can nevertheless admit a self-fulfilling settlement freeze under the same balance-sheet fundamentals. Thus, observing full repayment does not establish the uniqueness or robustness of the payment equilibrium. We derive this result from a precautionary-liquidity choice: banks may delay otherwise feasible payments to insure an interim funding need, trading a positive settlement-discipline cost against the marginal value of retained cash. The optimal reserve is a critical quantile of future liquidity requirements, and the repayment rule nests Eisenberg-Noe maximal feasible clearing exactly. Counterparty underpayment both reduces current cash and raises desired reserves. We characterise network amplification and the conditions for coexistence, including a homogeneous two-bank GPD frontier in which interbank exposure and maximum funding pass-through enter through one dimensionless index. A two-date settlement accounting rule distinguishes payment timing from ultimate default. The main result concerns equilibrium coexistence; the conditional best-response adjustment diagnostic does not assign selection probabilities.

0.39

Building AI-Intensive Software with AI: Early Results and a Cautionary Tale on Measuring Development Cost

cs.SE cs.AIcs.LG ·2026-08-13 ·Victor Barros de Miranda Neves, Kiev Santos da Gama, Vinicius Cardoso Garcia
published similarity 0.50
Abstract · 4 pages, 2 figures, Accepted for publication at the International Workshop on Intelligent Software Engineering (ISE 2026) @ CBSoft 2026

Empirical reports on the true cost of AI-intensive software development remain scarce, and the few that exist are easy to get wrong in ways that never surface in the final number. We report early results from an ongoing case study: a six-person student team built a full conversational onboarding assistant -- RAG-based code chat, guided tours, dependency graphs, technical-debt analysis -- over one academic term using pervasive AI assistance. We instrumented development with a three-layer cost model (real AI spend, self-reported human effort, human counterfactual) and initially reported a 19.4x cost ratio. A follow-up pass revealed two independent errors -- inferring per-token cost under a flat-rate subscription, and pricing the counterfactual with the wrong regional labor rates -- that together had inflated the ratio by roughly 2x; the corrected figure is ~9.9x. We present this correction as an early, generalizable finding in its own right: both errors are easy to make, invisible in the final number, and plausibly common in similar reports. We outline next steps toward a more robust, replicable costing methodology for AI-intensive development.

0.23

Anytime Contracts for AI Services

cs.GT cs.AIcs.LG ·2026-10-16 ·Boris Velasevic, Eden Saig, Eric Mazumdar
similarity 0.50
Abstract · 19 pages, 3 figures

AI services are prone to moral hazard, as model providers can silently route requests to cheaper, lower-quality models, and the inherent stochasticity of model outputs makes this degradation hard to detect. Contract theory addresses moral hazard by designing payment rules that reward high-quality outputs and prevent strategic overcharging. However, existing frameworks assume interactions of fixed duration, whereas AI services are typically consumed through subscriptions and pay-as-you-go APIs that users can terminate at any time based on observed performance. We show that this flexibility can break standard contracts, opening them up to abuse by either the providers or users. To address this, we extend contract theory to account for unknown exit rules. We formalize the problem as a Principal-Agent game, and show that budget-optimal contracts in this setting are proportional to anytime-valid hypothesis tests based on likelihood-ratio e-values. We give an efficient algorithm for computing anytime contracts and show how betting e-processes can be used to approximate them when the models are unknown. Finally, we evaluate our findings on synthetic data and data from deployed LLMs.

0.59