The Forecasting MachineRead report

Loading shared forecast

The Forecasting MachineOpen interactive report

Will direct fighting between the armed forces of Ethiopia and Eritrea result in over 500 battle deaths between November 10, 2026 and February 10, 2027?

Resolves based on the cumulative number of fatalities resulting from direct clashes, battles, or remote violence between the recognized armed forces of Ethiopia (e.g., ENDF) and Eritrea (e.g., EDF) occurring between November 10, 2026, and May 10, 2027. Fatalities are used as the standard measurable metric for 'casualties' as tracked by major conflict monitors. The Armed Conflict Location & Event Data Project (ACLED) will be the primary source for fatality counts. If ACLED data is delayed or unavailable by May 24, 2027, the resolution will rely on official UN reports or a consensus of major international news agencies (e.g., Reuters, AP, AFP).

Forecasted on

Forecast probabilities

  • Less than 500 casualties: 64.5%
  • 500 to 1000 casualties: 13.8%
  • More than 1000 casualties: 21.8%

Reasoning

The most likely outcome is fewer than 500 qualifying deaths (64.5%): serious direct fighting is possible, but the evidence points more strongly to localized clashes and proxy conflict than to sustained war. The risk of a rapid escalation remains substantial, putting more than 1,000 deaths at 21.8%; a limited but costly round of fighting, producing 500–1,000 deaths, is least likely at 13.8%.

Below 500: The likeliest path is containment

The crisis has moved well beyond rhetoric: diplomatic ties were severed, Eritrean troops were reported inside Tigray, Ethiopian drones reportedly struck Eritrean military vehicles, and Ethiopian forces advanced to the border [e2, e32, e34, e20, e21]. Yet these reports do not establish sustained combat between the two countries’ armed forces, and some Eritrean forces reportedly withdrew [e30]. The key distinction is that severe fighting in Tigray—including the TPLF’s claim of hundreds of Ethiopian government soldiers killed near Aksum—does not count toward this forecast unless deaths result from direct Ethiopia–Eritrea fighting [e55].

Both governments also face reasons to avoid opening a full interstate war. Ethiopia is already contending with internal conflict, while U.S. and U.K. officials have urged restraint and a return to dialogue [e1, e9, e11, e15]. Continued proxy-heavy conflict, threats, and occasional localized incidents could therefore remain below the threshold.

500–1,000: Possible, but a narrow middle ground

This outcome would likely require a significant but contained exchange: for example, several major border battles followed by a ceasefire. The forces’ proximity around Zalambessa, reported incursions, and drone strikes make such clashes plausible [e20, e21, e32]. But this bin is less likely than either containment or a much larger escalation: current reporting does not establish a substantial direct bilateral death toll, while a sustained war could push losses through this range quickly.

Above 1,000: A serious escalation tail risk

The path to this outcome is a breakdown of restraint and sustained conventional fighting. Ethiopia’s leadership has repeatedly emphasized gaining Red Sea access, with particularly stark statements concerning Assab; Eritrea has framed those ambitions as a threat and sought UN action [e18, e22, e23]. The two states’ previous war killed more than 100,000 people, underscoring how quickly deaths could accumulate if a full-scale conflict begins [a15, a16]. Reported troop deployments and military contact make that risk real, even if they do not prove that war is inevitable [e17, e34].

Why the probabilities favor containment

The evidence supports a live risk of direct combat, but not yet a reliable count—or clear trajectory—toward 500 bilateral battle deaths. Diplomatic breakdown, military contact, and the dispute over Assab raise the chance of a dangerous escalation; withdrawals, international pressure, Ethiopia’s internal conflicts, and the distinction between civil-war deaths and direct bilateral deaths all weigh against treating that outcome as the default. Accordingly, the forecast assigns 64.5% to fewer than 500 deaths, 13.8% to 500–1,000, and 21.8% to more than 1,000.

Cited evidence