Electional timing decisions, without hidden scoring weights

Cosmic Ephemeris turns sampled electional repair evidence into an exact hardening decision audit. It preserves competing tradeoffs, measures worst-case regret across uncertain failure budgets, and returns every tie.

See how the audit works Browse API documentation

Availability: Live beta
Production: 78 routes
Evidence: Exact sampled integers
New licenses: None

Direct answer

What is an electional hardening regret audit?

It is a deterministic comparison of how much protection each criterion-repair hardening profile leaves on the table when the number of simultaneous repair failures is unknown. At each analyzed failure budget, Cosmic Ephemeris finds the best observed avoided loss, subtracts every profile's avoided loss, and reports that exact difference as regret. The profiles with the smallest worst regret across all budgets are the minimax-regret ties.

Live request boundary

Call the R231 endpoint from a trusted backend

Send POST https://api.cosmicephemeris.com/api/astro/electional-criterion-repair-hardening-regret-audit from Node.js 22+ or another server runtime. Keep COSMICEPHEMERIS_KEY private; never put it in browser code, URLs, prompts, or analytics. Civil dates must stay within 1900-01-01 through 2050-12-31.

curl --fail --request POST \
  --url https://api.cosmicephemeris.com/api/astro/electional-criterion-repair-hardening-regret-audit \
  --header "Authorization: Bearer $COSMICEPHEMERIS_KEY" \
  --header "Content-Type: application/json" \
  --data '{"start_utc":"2026-10-04T00:00:00Z","end_utc":"2026-10-05T00:00:00Z","location":{"lat":40.7128,"lon":-74.006},"profile":{"zodiac":"tropical"},"filters":{"moon_signs":["Aquarius","Pisces"],"mercury_motion":"direct"}}'

Transparent by construction

A reliability audit, not another opaque score

Question	Typical ranked output	R231 live contract
How are tradeoffs combined?	Configured or undocumented weights	No weights; each failure budget remains visible
What happens to ties?	Often reduced to one displayed winner	Every tied benchmark and minimax profile is returned
Can the result be checked?	A final score may hide intermediate evidence	Exact integer regret, budget witnesses, and profile IDs
Is it advice?	May be presented as a recommended window	No ranking of times, prediction, interpretation, or recommendation

Evidence pipeline

How the live audit reaches minimax regret

01
Sample the declared criteria
Reuse the strict seven-grid, at-most-seven-day electional evidence boundary.
02
Prove minimal repairs
Retain exact cross-grid criterion relaxations and their recovered observations.
03
Model hardening profiles
Protect one repair at a time and preserve every unique avoided-loss vector.
04
Minimize worst regret
Compare profiles budget by budget and return all least-maximum-regret ties.

Bounded and auditable

The exact analysis ceiling

70maximum unique hardening profiles

4failure budgets: zero through three

280maximum profile-budget regret cells

19,600maximum benchmark references

The regret layer reuses the preceding hardening hierarchy and performs no additional chart calculation. A maximum-tie synthetic acceptance case emitted all 19,600 references and 70 minimax ties in about 0.7 ms on the acceptance host.

Answer-ready response

What a consumer can inspect

The source contract exposes separate budget benchmarks and per-profile regret certificates. This abbreviated shape is illustrative; it is not a live request example.

{
  "regret_status": "positive_minimax_regret_hardening_profiles_observed",
  "hardening_budget_benchmarks": [
    {
      "unavailable_repair_count": 2,
      "maximum_avoided_worst_case_observation_count": 7,
      "best_hardening_profile_ids": ["hardening_profile_003"]
    }
  ],
  "hardening_profile_regrets": [
    {
      "hardening_profile_id": "hardening_profile_002",
      "maximum_regret_observation_count": 1,
      "maximum_regret_failure_budgets": [1, 2],
      "minimax_regret": true
    }
  ],
  "minimax_regret_hardening_profile_ids": ["hardening_profile_002"]
}

Truthful availability

What is—and is not—available today

Built and accepted in source: engine, gateway, OpenAPI, migration, Node, Python, n8n, and 78-route loopback verification.

Not completed: migration 064, client publication, production deployment, live public endpoint verification, search indexing, ranking, or answer-engine citation.

Method limits

What the evidence does not claim

It does not prove the best continuous instant between sampled times.
It does not assign probabilities to failure budgets or preferences to criteria.
It does not establish that hardening causes an outcome outside the returned model.
It does not interpret astrology, predict events, or recommend a date.
It does not claim universal market uniqueness; the competitive review is limited to cited public product material.

Questions and answers

Electional reliability API FAQ

Is the electional hardening regret endpoint live?

Yes. R231 is live with 78 calculation endpoints, migrations 039 through 064 applied, public authentication boundaries verified, and a receipt-signed R231 calculation accepted with exact one-unit accounting and cleanup.

What does minimax regret mean in this audit?

For each analyzed failure budget, the audit compares a hardening profile with the best observed profile at that same budget. Regret is the exact difference in avoided worst-case sampled observation loss. Minimax profiles have the smallest maximum regret across all analyzed budgets, and every tie is returned.

Does the audit rank electional times or recommend a date?

No. It audits the robustness of sampled criterion-repair hardening profiles. It applies no probabilities, preference weights, quality score, interpretation, prediction, or recommendation.

Does this research add a licensed dataset or dependency?

No. The regret layer is original Cosmic Ephemeris source using standard-library integer comparisons and existing bounded evidence. It adds no dataset, package, catalog license, third-party code, or request-time network source.