IEEE S&P Topic Selection

Use this before any writing starts: is the project an Oakland paper, and if so, as a research paper or as an SoK? S&P has been the field's flagship since 1980 (first held at the Claremont Resort in the Oakland hills — hence the nickname), and its reviewer pool reads for adversarial rigor above topical fashion.

The two-question fit test

1. Is there a real adversary? State who attacks, with what access, and what they win. If the paper has no defensible threat model — or the "attack" needs an adversary who could already do worse things more cheaply — Oakland reviewers will say so in round one. Papers where security is a motivating anecdote rather than the analyzed object usually belong at a systems, ML, or networking venue instead.
2. Does the evidence close the loop? An attack needs an end-to-end demonstration against a realistic target; a defense needs evaluation against an adaptive adversary who knows the defense exists; a measurement study needs a sampling story and an ethics story. A promising idea with simulation-only evidence is a workshop paper, not yet an S&P submission.

SoK is a first-class lane

Both the 2026 and 2027 CFPs solicit Systematization of Knowledge papers, marked with the SoK: title prefix and a submission-form checkbox (sp2027.ieee-security.org/cfpapers.html, checked 2026-07-08). An SoK is judged on its treatment of existing work — a new viewpoint on a major area, evidence that challenges a long-held belief, or a genuinely new taxonomy — not on new results. If your "survey" cannot name the belief it overturns or the structure it reveals, it is a literature review, and S&P rejects those.

Routing among the security venues

Route by contribution type and by calendar, not prestige: the four flagships alternate deadlines across the year, and S&P's own two cycles (June and November for the 2027 symposium) may simply not be the next open door.

Embargo math changes the answer

Two S&P-specific rules constrain venue strategy (2027 CFP, checked 2026-07-08):

• A rejected paper may not return to S&P for one year from the original submission date — so a weak submission burns the venue, not just the cycle.
• A new submission counts as a resubmission when roughly 40% or more overlaps a previously rejected one; splitting a rejected paper into two does not reset the clock.

If the evidence will not be ready, targeting a sibling venue first and S&P with the mature version is often the higher-expected-value order.

Quick self-interrogation

Oakland fit worksheet (answer in one line each):
1. Adversary + access + payoff: ..............................
2. Evidence type that closes the loop (exploit / adaptive eval /
   measurement + ethics / systematization): ..................
3. If SoK: which belief or structure does it change? .........
4. Embargo status: any ≥40%-overlap rejection at S&P in the
   last 12 months? ...........................................
5. Which cycle is realistically reachable with finished
   evidence (not finished writing)? ..........................
→ Any blank line = not ready to pick S&P yet.

Signals you are choosing S&P for the wrong reason

• The main argument for Oakland is "it is ranked highest" — the four flagships are close peers; wasted cycles are the real cost.
• The threat model was written after the experiments, to justify them.
• The contribution is a faster/cheaper version of a published system with the same security argument — reviewers will ask what the security delta is.
• The plan depends on the Revise outcome as a safety net; Revise is offered to a limited set of papers reviewers already believe in, not to drafts.

Output format

[S&P fit] Strong / Plausible / Re-route (one-line reason)
[Paper type] research / SoK
[Threat-model status] stated & realistic / stated but strained / missing
[Evidence gap] <the one experiment or analysis that closes the loop>
[Embargo check] clear / blocked until <date> / 40%-overlap risk
[Alternative order] <e.g., USENIX Sec first, S&P with mature version>