How does the Butthurt Meter work?

A single-needle grievance gauge reads 0-100 butthurt units. Four times a day a research apparatus (Grok 4.5, wired to live web and X search) collects the day's stories as structured data, and audited, unit-tested code computes the needle from unique outrage clusters, a peak-relative reading, and a trailing-365-day historical percentile. The model proposes the stories; it never touches the needle.

How often is the reading updated?

Four times per day, at 00:00, 06:00, 12:00 and 18:00 UTC. The needle has never once been early.

Is the score decided by an AI?

No. The language model proposes and describes the stories, but the overall needle score and the all-time peak are computed entirely by deterministic, unit-tested code. The robot writes the jokes; the math does the feelings.

Is the Butthurt Meter real news?

No. It is satire and commentary on public freakouts, not reporting. Scores are jokes with calibration. Every card is labeled real story or satire, and a deterministic filter blocks fabricated quotes and limits coverage to public figures on an allowlist.

What is a unique outrage cluster?

The fundamental unit of the science: one real-world event, no matter how many people are yelling about it. Forty identical hot takes about the same event count as one outrage. Deduplication is performed at the event level.

BUREAU OF GRIEVANCE METROLOGY · TECHNICAL BULLETIN Nº 001 · PUBLIC EDITION

The Science^™

Q: Is the Butthurt Meter politically biased?

Yes, and it says so on every page. The Bureau acknowledges exactly one systematic bias: the meter only points one direction politically. That disclosure is more than most instruments make.

A complete and unnecessarily rigorous account of how one needle knows, four times a day, exactly how upset Democrats are.

Certificate of Calibration

CURRENT CERTIFIED READING: —/100 BUTTHURT UNITS

This instrument is certified accurate to ±0.0 vibes at a 95% confidence interval (n = the discourse). Certification renews automatically at 00:00 and 12:00 UTC, whether or not anyone deserves it.

MOTTO: DOLOR MENSURATUR, NON IUDICATUR¹

§1 · THE INSTRUMENT

The Butthurt Meter is a single-needle grievance gauge, graduated from CHILL to BUTTHURT across 100 standard units of upset. The needle is real. The mathematics are real. The numbers are extremely real. What they measure is a feeling — but so is temperature, if you think about it, which we have, at length, in this document.

§2 · DATA ACQUISITION

Four times per day, a research apparatus (Grok 4.5, wired to live web and 𝕏 search) reads the recent news and posts so that you are not required to. It returns the day's outrage inventory as structured data. It is not permitted to invent events, quote anyone, or editorialize beyond the legally mandated level of sarcasm.²

§3 · CLUSTER THEORY

The fundamental particle of this science is the unique outrage cluster (K): one real-world event, no matter how many people are yelling about it. The raw intensity R takes K as its spine and adjusts for how loudly the board is actually screaming:

R = K + 3.5·(I − I₀) + lead-story bonus I is the measured heat of the day's stories; I₀ = 0.78 is a certified Normal Amount of Upset. A loud board pushes R above K; a listless one drags it below. Calibrated, audited, and clamped, in that order.

40 identical hot takes = 1 (one) outrage First Law of Grievance Metrology. Corollary: a ratio'd post is not an outrage; it is a cry for help.

Deduplication is performed at the event level. A ceasefire covered by forty outlets, screamed about in nine hundred quote-posts, remains a single cluster. Science requires discipline.

§4 · THE NEEDLE EQUATION (scoring_version v2-peak-history)

The needle S is computed from R by blending two independent estimators — a peak-relative reading and a historical percentile — because one number pretending to be objective is amateur hour, and we use two.

H = 100 · (R / P)^ρ, ρ = 1.2 The Escalation Exponent: near the record, every additional simultaneous scandal counts extra. Grievance compounds. We have the exponent to prove it.

Q = smoothed percentile(R | trailing 365 days) How does today rank against every day we have measured? Kernel-smoothed, because history has texture and cliffs are for geology.

S = clamp₀¹⁰⁰( α·H + (1−α)·Q ), α = 0.6 The certified blend. 60% "how close to the worst day ever," 40% "how unusual is this, historically."

P is the all-time peak — the hottest reading ever recorded, currently anchored at a calibration ceiling of seven (7) standard outrages. That ceiling is our Kelvin. A new peak is certified only when a reading clears the old one by a full margin — near-misses are logged as candidates and watched with suspicion. Peaks can only rise; the ceiling of human upset has never once come down, and we see no reason to model otherwise.³

⚠ REGULATORY NOTICE: THE MODEL PROPOSES STORIES. IT DOES NOT TOUCH THE NEEDLE. S AND P ARE COMPUTED BY AUDITED, UNIT-TESTED CODE ONLY. THE ROBOT WRITES THE JOKES; THE MATH DOES THE FEELINGS.

§5 · LEGACY TAG SPECTROMETRY (per-story readings)

Each individual story card carries its own intensity reading, informed by the classical tag-weight spectrum (retained for instrument redundancy, like the second altimeter in a cockpit):

SPECTRAL LINE	CONTRIBUTION	SATURATION CAP
trump	+18	3 clusters
musk	+12	2 clusters
politics · media · tech · economy	+9	3 each
culture	+8	3 clusters

legacy = 100 − 60·e^{−(raw−40)/45} The saturation curve: outrage approaches 100 asymptotically. It never arrives. There is always room to be more upset. This is the most empirically verified statement on this page.

§6 · QUALITY ASSURANCE

Every candidate story passes a deterministic policy filter before it may approach the needle:

CHECK	ENFORCEMENT
No fabricated quotes, ever	regex, not vibes
Public figures only	private individuals barred; the model discloses every name; a block list excludes specific figures
Real story vs. satire	labeled on every card, in public
Too few survivors	fail closed — if the science cannot be done safely, no science is emitted, and yesterday's science remains in effect
SEARCH X chip	optional deep-link to an X search for the cluster — a docket tool, not a verified source post or corroborating exhibit

§7 · PEER REVIEW

This infrastructure was adversarially reviewed by 45 independent AI agents instructed to destroy it.⁴ They confirmed 21 defects. All 21 were remediated. A robotic canary now loads this site every hour and personally verifies that the needle still moves; if it does not, a human is woken by email, which is its own form of outrage and is not currently counted in R (see Limitations).

§8 · LIMITATIONS

The instrument measures discourse about feelings, which are themselves feelings. Precision: absolute. Accuracy: philosophical. The Bureau acknowledges exactly one systematic bias — the meter only points one direction politically — and has published this fact on every page, which is one more disclosure than most instruments make.

§9 · PUBLIC RECORDS

The Bureau's complete records, as required by absolutely no one. Every reading is published as public, crawlable JSON — the same endpoints this site reads. No key, no signup, CORS open.

ENDPOINT	CONTENTS
/data/gauges.json	the current reading: needle, headline, story cards (updated 4×/day)
/data/history.json	the time series: every reading's score + trend
/feed.xml	RSS: subscribe in any reader and the needle comes to you

The JSON shape is a stable contract (schema version 1). Prefer a spreadsheet? The Bureau will render the full time series to CSV, in your browser, on demand:

¹ "Grievance is measured, not judged." The Bureau's founding motto, adopted unanimously by both of us.
² Sarcasm compliance is verified per 29 C.F.R. § never; the blurbs are jokes and are labeled as jokes.
³ Longitudinal studies pending. Funding: none. Confidence: total.
⁴ True story. The review transcripts exist. The robots were not gentle.

← RETURN TO THE CERTIFIED NEEDLE

The Science™