In 2002, the most rigorous armed conflict dataset then available — the UCDP, covering 225 conflicts from 1946 to 2001 — showed conflict declining sharply after the Cold War. The evidence was strong enough that the world's leading conflict scholars, SIPRI, and the UN embraced cautious optimism.
UCDP v25.1 (updated to 2024) and verified 2025–2026 sources show that optimism was premature. The 2022 spike to approximately 237,000 battle deaths — driven by Ethiopia and Ukraine — was 48% higher than any year recorded between 1993 and 2019. By March 2026, a new interstate war had opened in the Middle East.
Raw death counts systematically deceive in asymmetric conflicts. The Human Cost Index (HCI = total casualties ÷ population × 10,000) reveals that Gaza's violence — 75,200 dead from a population of 2.3 million — represents 327 deaths per 10,000 people. That is second only to Rwanda among modern conflicts, and 1,500 times the rate recorded in Iran from the same 2026 war.
The Casualty Rate (HCI ÷ conflict years) measures the velocity at which a conflict consumes a population. Rwanda's rate was approximately 4,568 per 10,000 per year. Gaza's is 131. Lebanon in the 2026 Iran-Israel war reached 26.5 in its first 31 days. The metric translates atrocity from totals into urgency.
The UCDP methodological framework explicitly identified the need for better human cost accounting. HCI and Casualty Rate are not replacements for that framework — they are its logical completion, finishing the work of precision that the field's founders began.
Gleditsch, N. P., Wallensteen, P., Eriksson, M., Sollenberg, M., & Strand, H. (2002). Armed conflict 1946–2001: A new dataset. Journal of Peace Research, 39(5), 615–637. https://doi.org/10.1177/0022343302039005007
Human Security Centre. (2005). Human security report 2005: War and peace in the 21st century. Oxford University Press.
Russett, B., & Oneal, J. R. (2001). Triangulating peace: Democracy, interdependence, and international organizations. Norton.
Uppsala Conflict Data Program. (2025). UCDP/PRIO armed conflict dataset version 25.1. Uppsala University. https://ucdp.uu.se/
Obermeyer, Z., Murray, C. J. L., & Gakidou, E. (2008). Fifty years of violent war deaths from Vietnam to Bosnia. BMJ, 336(7659), 1482–1486.
Rwanda killed 800,000 people in 100 days. Gaza has killed more than 75,000 in 2.5 years. The numbers are not the same — but the Casualty Rate metric reveals that the mechanism is. This post introduces the velocity of killing as the field's most urgent missing variable.
On January 5, 2025, the Lancet estimated that 75,200 people had been violently killed in Gaza since October 7, 2023. The UCDP ranked that conflict eighth in its 2024 deadliest table. Both numbers are correct. The problem is what they hide.
UCDP v25.1 covers armed conflict through 2024. Supplemented with verified data from 2025 and 2026, it tells a story the optimists of 2002 did not foresee: the S-curve has reversed, and the mechanisms they trusted have proven insufficient.
In 2002, the most comprehensive armed conflict dataset ever built said war was in decline. SIPRI agreed. The UN agreed. The evidence was genuinely compelling — which is what makes what came next so important to understand.