What's New

October 2026

A page for every counter

Every counter, and every group of counters, now has a page of its own at its own address — for example, Deaths in ICE Custody or the Iran war. Click any counter's name on the home page to get there. These pages replace the old Info pop-up, so they can be bookmarked, linked to and shared.

About, FAQ, Sources and this page have their own addresses too.

Charts you can take with you

The chart button on any counter opens a larger chart that can be customized and downloaded.

The style editor offers presets (black and white, dark, slide, scientific) or your own colors and fonts, and you can download any chart as a PNG or SVG for an article, slide or post.

Charts of groups of counters, and of the overall total, offer two additional views that display the breakdown of contributions to the total: Today's breakdown as a pie chart, and the breakdown over time as stacked bands.

The style is yours to change, but some things aren't: every chart keeps the counter's own title, says whether its numbers are reported or modeled, marks projections as projections, flags a value axis that doesn't start at zero, and labels a chart built from your own Model Explorer choices as a custom scenario.

Reported or modeled?

Every count on the site is now labeled as one of two kinds. Reported counts add up deaths that someone documented — ICE custody deaths, measles deaths, strikes on boats — and each says what was reported. Modeled estimates come from published research models, like the disease counters for the USAID and PEPFAR cuts. The two kinds deserve different levels of confidence, and you shouldn't have to dig to find out which one you're looking at.

More sources for the Iran war

HRANA, whose counts anchored our Iran war counters, stopped publishing cumulative totals after July 17. Since then, the civilian counter adds deaths from individual incidents documented by Airwars, a monitor of civilian harm from military action, using the lowest reported figure for each. The Iranian military counter adds only deaths that Iranian state media, the IRGC or wire services explicitly report, each reviewed by hand. Both are floors, not estimates: Airwars publishes incidents days to weeks after they happen, and Iran does not announce most of its military losses.

Privacy Improvements

We removed Google Analytics and PostHog. The only visit counting left is Cloudflare's, which uses no cookies; the About page now has a privacy statement.

Smaller changes

• The measles page now spells out which deaths count: those a coroner, medical examiner or state health department attributes to measles or one of its complications. A positive test alone isn't enough.
• A new logo and a cleaner header.
• Layouts, charts and counter rows work better on phones.

September 2026

Better mortality models

Three of our disease counters are now anchored to the best published models we could find, rather than extrapolating forever from early estimates:

• Malaria now follows the Malaria Atlas Project's 2025 analysis of the President's Malaria Initiative — the field's most detailed geospatial model of what PMI funding was preventing.
• AIDS deaths are now derived from UNAIDS' projections of the impact of the PEPFAR cuts. Because UNAIDS models all PEPFAR-supported programs, these counters now cover the whole world, not just sub-Saharan Africa.
• Tuberculosis now follows a peer-reviewed epidemiological model published in PLOS Global Public Health, which projects the effect of the U.S. funding cuts under three explicit recovery scenarios. Our counter defaults to the no-recovery scenario — the one that matches what has actually happened so far — and the more optimistic scenarios remain a click away.

Some of these counts changed substantially as a result, in both directions. The models we used before, from the Impact Counter project, have not been discarded: each affected counter keeps them as a selectable alternative, so you can compare the old and new estimates yourself.

Each counter's Info page (now the counter's own page) documents exactly how its current model works and where every number comes from.

A more honest culpability model

We have always said that responsibility for these deaths could shift away from the U.S. over time, as other donors and agencies had a chance to take over — and our old model expressed that with an admittedly arbitrary rule: a straight-line decline to zero over two years.

The published transition plan for handing off PEPFAR programs schedules annual funding step-downs, and our culpability model follows them, rescaled so culpability reaches zero as the scheduled handover completes (October 2025 through September 2029). The AIDS counters step down along that schedule.

The other USAID counters decline along a smooth curve over the same window. This models how long a reasonable, orderly handover would have taken — not how the actual transition is going. If you weigh that judgment differently, you can change it (see below).

The Model Explorer

Every counter backed by a model now has an Explore Model button on its Info page (now its counter page). It shows you the whole machine: the source data, the mortality model built on it, and the culpability schedule applied on top, each with charts and data tables you can inspect.

And it isn't just for looking. If you disagree with our judgment calls, make your own: choose a different mortality model where alternatives exist, pick a different culpability schedule, or cap the share of these deaths you attribute to the President at all. Apply your changes and the whole page — every counter and total — recalculates with them, and remembers your choices the next time you visit. One Reset All button puts everything back.

We built this because these numbers rest on judgment calls, and we would rather show you the judgments than ask you to take our word for them.

Fresher data

We now check our sources more frequently. Counts built from individual reports — ICE deaths, epidemic deaths, military actions — are updated as new reports appear, and our model-based counters are revisited whenever the organizations behind them publish new estimates. The Sources page lists everything we rely on, and each counter's page has been rewritten to document its current methodology.