ijoc-rebuttal
GitHub用于起草INFORMS Journal on Computing (IJOC)的修改回复信。针对审稿人关于实验公平性、规模和可复现性的质疑，提供逐点回应策略与修订计划。强调以证据而非修辞应对，指导如何整合新实验数据并诚实陈述局限，不直接运行分析。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill ijoc-rebuttal -g -y
SKILL.md
Frontmatter
{
"name": "ijoc-rebuttal",
"description": "Use when drafting the response to an INFORMS Journal on Computing (IJOC) decision letter — turning referee objections (especially about experiments and reproducibility) into a point-by-point response with re-run evidence. Builds the rebuttal and revision plan; it does not run new analysis on its own (route to ijoc-data-analysis)."
}
Rebuttal Strategy (ijoc-rebuttal)
When to trigger
- An IJOC major/minor revision arrived and you need a response-letter strategy
- Referees challenge the fairness, scale, or reproducibility of the computational experiments
- You disagree with a referee on a method point and must push back without alienating the associate editor
- You need to convert a long, mixed decision letter into an ordered, checkable revision plan
The IJOC rebuttal mindset
IJOC revisions are won with evidence, not rhetoric. Most consequential objections are computational ("the comparison is unfair," "instances too small," "cannot reproduce"), and the strongest response is a re-run that settles the question: more standard instances, a stronger baseline, more seeds, a performance profile, an updated deposit. Because the associate editor is undisclosed and single-blind, write to the substance of each report for a technical reader; never speculate about who wrote what. The AE is reading for whether you took the experiments seriously.
Structure of the response
- Open with the changes that matter. A short summary: the main new experiments, the strengthened baselines, the reproducibility fixes — so the AE sees the substantive work first.
- Point-by-point, quoting each comment. Reproduce each referee point verbatim, then respond. Use a consistent tag: [Done — with evidence], [Done — text], [Respectfully disagree — with reason].
- Show the evidence inline. For computational objections, include the new table/profile or its key number and a pointer to where it now lives in the paper and the deposit. "We re-ran on the full MIPLIB-class set; the performance profile (new Fig. 5) shows we dominate beyond τ=1.4; updated in
results/(illustrative)." - Be honest about limits. Where a competitor still wins a regime, say so and scope the claim rather than overclaiming — IJOC referees reward calibrated honesty.
Mapping common IJOC objections to the response
| Objection | Winning response | Owning skill |
|---|---|---|
| "baseline is weak / outdated" | re-run vs. current solver/SOTA, symmetric tuning; show the new comparison | ijoc-data-analysis |
| "instances cherry-picked / too small" | full standard set + scaling plot; report losses too | ijoc-data-analysis |
| "results not statistically supported" | add Wilcoxon + performance profile; multiple-seed dispersion | ijoc-data-analysis |
| "no guarantee / correctness unclear" | add validity/complexity proof; reconcile theory with plots | ijoc-theory-development |
| "cannot reproduce your numbers" | fix and re-tag the deposit; align results/ to tables; confirm it runs clean |
ijoc-data-analysis |
| "contribution incremental" | sharpen claim + positioning vs. recent work | ijoc-contribution-framing / ijoc-literature-positioning |
| "out of scope" | re-argue computing-first fit, or reconsider venue | ijoc-topic-selection |
Disagreeing well
You may push back, but with a reason a technical reviewer accepts: a misread of the method, a baseline that is not actually comparable, a request outside the paper's scope. Concede the framing ("the reviewer is right that X would strengthen the paper") even while declining ("but it changes the research question; we instead add Y, which addresses the underlying concern"). Flat refusal with no alternative is what loses AEs.
Checklist
- Response opens with the substantive new experiments and reproducibility fixes
- Every referee point quoted and answered with a clear status tag
- Computational objections answered with re-run evidence, not prose
- New/updated exhibits pointed to in both the paper and the deposit
- The deposit re-tagged and confirmed to reproduce the revised results
- Disagreements carry a technical reason and an alternative
- Claims rescoped honestly where a competitor still wins
- Tone is professional; no speculation about the undisclosed AE/referees
Anti-patterns
- Answering an experimental objection with words instead of a re-run
- Marking points "Done" without changing the paper or the deposit
- Updating the manuscript but leaving the GitHub deposit stale (a "cannot reproduce" relapse)
- Flat refusal of a reviewer request with no reason or alternative
- Overclaiming after revision instead of scoping to the regimes you actually win
- Guessing the AE's or reviewer's identity and writing to it
Output format
【Journal】INFORMS Journal on Computing
【Skill】ijoc-rebuttal
【Summary of changes】new experiments / baselines / reproducibility fixes
【Point-by-point】each comment quoted + status tag + evidence pointer
【Re-run evidence】profiles/tests/seeds added where challenged
【Deposit】re-tagged and reproduces revised results? [Y/N]
【Honest scoping】claims rescoped where a rival wins
【Next step】resubmit via ScholarOne → ijoc-review-process to track the new round
Version History
- 1839142 Current 2026-07-05 13:21
