brycewang-stanford Contribution Framing (ijoc-contribution-framing)
When to trigger
- You cannot state in one sentence what is computationally new and by how much
- The intro frames an application win when the contribution should be the method
- A coauthor or referee says "interesting but what is the actual contribution?"
- The experiments are solid (
ijoc-data-analysis) but the paper does not claim them crisply
The IJOC contribution sentence
A strong IJOC contribution sentence names the method, the axis of improvement, the magnitude, and the evidence regime — and is falsifiable. Template:
"We propose [method] for [OR problem]; it [wins on axis] by [magnitude] over [SOTA baseline] on [benchmark set], with [guarantee/property]."
Worked, illustrative: "We propose a Benders-with-lazy-cuts scheme for two-stage stochastic network design that solves instances 6× larger than the best published exact method on standard library instances within the same time limit, with a proof of cut validity and a 31% root-gap reduction." That sentence tells the Area Editor exactly what to verify.
Types of contribution IJOC rewards
Frame the contribution as one (or a clear combination) of these, and make the primary one unambiguous:
| Contribution type | What makes it land at IJOC | Common failure |
|---|---|---|
| New exact method / formulation | larger solvable instances, tighter bounds, with validity proven | "new model" with no computational edge |
| New heuristic/matheuristic | better quality or time, fairly compared, ideally with a bound | win over a strawman baseline |
| ML-for-OR method | beats both OR and prior learning baselines; feasibility safe | a black box with no OR baseline |
| New simulation/estimation method | lower variance/cost at equal accuracy, with theory | "we simulated it" with no methodological news |
| Computational tooling/methodology | enables computations not previously feasible, reproducibly | software existence without a methodological claim |
Application is the motivation, not the contribution
The most common framing error at IJOC is leading with the application ("we study airline crew scheduling") rather than the computing ("we design a column-generation acceleration that…"). State the application in one or two sentences as motivation, then pivot hard to the methodological news. If you remove the application and the paper still has a contribution, you are framed correctly; if it collapses, the computing is not yet the contribution — return to ijoc-topic-selection.
Calibrating the claim to the evidence
Overclaiming triggers referee distrust; underclaiming wastes a good result. Match the verb to the evidence: "proves," "guarantees" only with proofs; "consistently outperforms" only with a performance profile and a statistical test; "scales to" only with a scaling plot. Scope honestly by regime ("on sparse, large instances") rather than blanket dominance. The claim in the abstract, the intro, and the conclusion must be the same claim.
Checklist
- One sentence states method + axis + magnitude + baseline + benchmark + property
- The primary contribution type is unambiguous
- The claim is computing-first; application is motivation only
- Verbs match the evidence (prove/guarantee/outperform/scale) and nothing stronger
- The claim is scoped by regime where a competitor still wins
- Abstract, intro, and conclusion state the identical claim
- Removing the application leaves a standing contribution
Anti-patterns
- Leading with the application domain and burying the method
- "We are faster/better" with no magnitude, baseline, or benchmark named
- A contribution sentence that would fit Operations Research (model news) rather than IJOC (computing news)
- Claiming guarantees the theory section does not prove
- Different contribution wording in the abstract vs. the conclusion
- Listing five small contributions with no headline — referees cannot tell what to evaluate
Output format
【Journal】INFORMS Journal on Computing
【Skill】ijoc-contribution-framing
【Contribution sentence】method + axis + magnitude + baseline + benchmark + property
【Primary type】exact / heuristic / ML-for-OR / simulation / tooling
【Computing-first?】application is motivation only? [Y/N]
【Claim–evidence match】verbs justified by proofs/profiles/scaling? [Y/N]
【Regime scoping】where the claim holds / where a rival wins
【Next skill】ijoc-tables-figures