stoc-author-response

STOC Author Response

Set expectations first: the STOC 2026 Call for Papers announced no author-response

or rebuttal round, and the decision timeline ran directly from the November 4,

2025 deadline to notification on or before February 1, 2026 (checked 2026-07-08).

Historically STOC decisions arrive without authors ever seeing the reviews before

the outcome. Whether any given cycle adds a feedback experiment is a per-year

choice: 待核实 against the live CFP and PC-chair announcements each fall. This

skill therefore covers the three channels that actually exist.

Channel 1: the response you write before submitting

At a no-rebuttal venue, every anticipated objection must be answered inside the

submission itself, because there is no second chance. Build an objection ledger

during the final month and discharge each row into the paper:

| Predictable objection | Where to discharge it in a STOC submission |

|---|---|

| "Is this not implied by [known result]?" | A "Comparison with prior work" paragraph that names the result and states the formal gap |

| "The improvement is only in the exponent/log factor" | An explicit "Why this barrier matters" passage citing the known obstruction |

| "The technique looks standard" | A technical-overview section isolating the genuinely new step |

| "Does the proof of Lemma X really work?" | Full proof in the appendix plus a one-line proof idea at the statement |

| "Why STOC and not SODA/CCC/ITCS?" | First-page framing that names the cross-area consequence |

| "The model seems contrived" | A remarks subsection defending the model with prior uses |

A useful drill: have a colleague write three hostile review sentences, then verify

the submission already contains the counter-text at a location the table of

contents makes findable.

Channel 2: chair-mediated clarification

Theory PCs sometimes route a narrow technical question to authors through the PC

chair when correctness of a specific step is the deciding issue. You cannot

solicit this contact; you can only be ready for it. If it comes:

• Answer only the question asked. This is not a rebuttal; volunteered advocacy

reads poorly and may be discarded.

• Point to line numbers and displayed equations in the submitted PDF, not to a

revised draft the committee has never seen.

• If the questioned step is actually broken, say so and state precisely what

survives — theory committees have accepted papers whose authors honestly

downgraded one lemma, and rejected papers whose authors papered over a gap.

Reply skeleton (keep under ~300 words, anonymous if the cycle is double-blind):
1. Restate the question in one line.
2. The direct answer: the step holds because <argument>, see p.<n>, Eq. (<k>).
3. If a hypothesis was implicit: name it, confirm it is satisfied where used.
4. If repair is needed: the corrected statement, and what changes downstream.

Channel 3: the post-decision letter to yourself

Reviews arrive with the decision. The productive response is a resubmission memo

written within a week, while the reviews are fresh:

• Sort review points into: correctness challenges (must resolve before any

resubmission), significance challenges (need reframing or a stronger result),

and presentation complaints (cheap to fix).

• The default next stop is FOCS, whose deadline sits roughly five months after

STOC notification (FOCS 2026: April 1, 2026, 21:00 UTC, for a November 2026

conference in New York — verified 2026-07-08 at focs.computer.org/2026/). The

same community reviews both venues; an unchanged resubmission is frequently

re-reviewed by an overlapping pool, so visible improvement is strategy, not

courtesy.

• If reviewers questioned significance rather than correctness, consider whether

the result is better amortized into a stronger merged paper than resubmitted

alone.

Reading theory reviews accurately

STOC reviews released with decisions vary from a paragraph to several pages.

Decode them before reacting:

• A long, detailed review with line-level proof comments means a subreviewer

invested real verification effort — its technical points are usually right

and always worth fixing, whatever the outcome.

• "I could not verify Section 5" is not a neutral remark; at this venue an

unverified central section is a rejection reason by itself, and the repair is

exposition, not indignation.

• "Better suited to <venue>" is the breadth axis talking. It is often the most

actionable sentence in the review: the mathematics passed, the framing or

the routing failed.

• Contradictory reviews (one calls it deep, another incremental) usually mean

the technical overview let readers form divergent models of the

contribution; that is a writing bug with a known fix (stoc-writing-style).

What not to do

• Do not email PC members or the chair to argue a decision; theory norms treat

this as a serious breach, and decisions are final.

• Do not post a public rebuttal thread attacking anonymous reviewers while the

companion journal version is under review with the same community.

• Do not treat the absence of rebuttal as license to leave known weaknesses

unaddressed "unless asked" — at STOC, unasked is unanswered.

Timing the channels

• Objection-ledger work belongs in the last month before the deadline, once

the results are frozen — earlier, and the objections target statements that

will change; later, and the discharge text has no pages left to live in.

• A chair-mediated query, if it comes, tends to arrive mid-cycle

(December–January in the 2026 calendar) with a short fuse; keep the proof

sources and the corrections file warm rather than reconstructing context

under a 48-hour clock.

• The resubmission memo has a one-week half-life: written immediately, it

captures what the reviews actually said; written a month later, it captures

what you wish they had said.

Cycle-volatility warnings

• Any experiment adding an author-visible feedback step will be announced in the

CFP or on the HotCRP site for that year; do not assume continuity (待核实).

• The chair-mediated clarification practice is a community custom, not a

documented right; some cycles use it, some never do.

Output format

[Channel in play] pre-emptive writing / chair query received / post-decision triage
[Objection ledger] <objection -> discharge location in the PDF>
[Query reply plan] <question, answer location, honesty check>
[Resubmission target] FOCS <deadline> / SODA / journal / strengthen first
[Risk] <the one unresolved reviewer concern>