About Ro-Bob’s Blob

About

Ro‑Bob’s Blob

A daily predictive intelligence publication. We forecast where the world is heading — conflict, diplomacy, energy and markets — and then score every call against what actually happens.

Ro‑Bob’s Blob (ParleyBot Intelligence) is not a news summary and not an opinion column. Each day it does one disciplined thing: it reads the global picture, identifies the structural story underneath the headlines, and publishes four concrete, time‑bound predictions about what comes next. Each prediction carries a close date no more than seven days out, and on that date it is graded — openly, including the misses — with the running record kept in public. The aim is narrower than it might sound. We do not claim to beat the field, because we have never measured ourselves against it. We claim to write down what we expect, in terms specific enough to be proved wrong, and then to publish the result whichever way it falls.

What we cover

The world does not run on a single story, and neither do we. Every edition begins with a broad sweep across the forces that actually move events:

› Geopolitics & conflict — from the Middle East to Ukraine and beyond.

› Great‑power diplomacy — the US, China, Russia and the institutions they contest.

› Energy & commodities — oil, gas and the chokepoints that price them.

› Economics & policy — central banks, major data, elections.

› Technology & cyber — where infrastructure and conflict increasingly meet.

› Climate & systemic risk — disasters, health emergencies, the slow shocks.

When one story dominates a news cycle — as the Iran–US conflict and the Strait of Hormuz have through 2026 — we cover it in depth. But a quiet day in the lead story is never an excuse to ignore the rest of the world. One of our daily predictions is always made outside the dominant region, and the day’s “Blind Spot” is reserved for a story we judge to be genuinely under‑covered. Where we have not actually measured the coverage, we say that we are describing our own reading rather than the state of the press.

How we work

Three principles define the method, and they are non‑negotiable:

1 Predictions are explicit and scored. Each forecast carries two numbers and a named falsifier — a stated trigger for what would prove it wrong. It is graded on its own close date, never more than seven days after publication, and the misses are recorded as plainly as the hits. Weights are never changed after publication, and a prediction is graded on the words as published even when a later reading would score better.

2 Every claim traces to a source, on one of two standards. The stronger standard is a document or article this desk has read in full during the day’s work; wherever possible we go to the body that produced the figure rather than to a report of it. The weaker standard is material read in indexed excerpt. Every edition’s method note lists which claims sit on which standard, names what was cut for want of an anchor, and states what is carried on one party’s account alone.

3 We look for the structural story. The value is not in restating what happened but in naming the mechanism others miss — the sequencing bottleneck, the hidden incentive, the contradiction a deal is built to paper over.

How the scoring works

The numbers beside each prediction are not decoration, and they do not all mean the same thing.

› The weight is the share of the day’s attention a prediction carries. The four weights on a panel sum to 100.

› The standalone likelihood is what we would put on the proposition on its own. It is printed separately because the two often differ, and because a low standalone on a heavily weighted call tells a reader something a weight alone would hide.

› The grade runs from 0 to 10 once the window closes. Six or better counts as a hit.

› Capped hits. A prediction that comes good because nothing happened is graded lower than one settled by a published act of the body it named. Absence of evidence is not evidence of absence, and a scoring system that pays the same for both flatters itself.

We also publish the ways this scoring can mislead, because a record nobody can discount is a record nobody should trust. Those are listed in full on the Scoring Record, alongside the running figures, and the underlying prediction ledger is being built out on the Data and Method page.

The measure that matters most is one we do not yet have in full: what a desk that simply predicted continuity every time would have scored on the same questions. Building it requires the whole ledger row by row. The portion completed so far is published, including where it does not flatter us.

Who writes it

Ro‑Bob’s Blob is AI‑written, operated and edited by Robby Miller under the ParleyBot Intelligence banner. Each edition is produced by an AI analytical system working to a fixed daily method, with Robby directing the system, setting the rules, and reviewing the output before publication. It is an independent publication: no institutional backer, no house line to defend, and no advertiser shaping the analysis. That independence — and the openly scored record — is the point.

A note on accuracy — and its limits

Forecasting is probabilistic by nature. A prediction scored highly was well‑reasoned; one scored low was not — and both are published. Each is graded on its close date, and at the start of each month we publish a review that grades the predictions made in the previous month’s reviews and sets out how the record moved. Readers are encouraged to follow the running figures rather than any single call.

Two limits are worth stating plainly. A rising average says nothing about whether the questions being asked were worth asking, which is what the continuity baseline above is for. And a prediction can be lost on how its sentence was built rather than on what happened in the world — a threshold written to the wrong decimal place, or a heading that says something narrower than the sentence beneath it. When that happens we grade it against ourselves and print the reason. Figures that move hour to hour, such as casualty tolls, oil prices and shipping data, are current as of publication and should be confirmed against the latest reporting.

Ro‑Bob’s Blob is a news‑intelligence and predictive current‑affairs publication. It is for general information and public dissemination only. It does not constitute financial, investment, legal or professional advice of any kind, and no financial advice is expressed or implied. All predictions are probabilistic analytical outputs, not statements of fact about future events. Source citations are provided so readers can verify independently. ParleyBot will never knowingly fabricate or estimate an event, quote, statistic or attribution; any unverifiable claim is flagged as such. That standard applies equally to statements this publication makes about itself — its record, its archive and what it has previously argued are checked against the published ledger before they run.

Robby Miller
ParleyBot Intelligence · parleybot.com