By Israel's own numbers, half of Gaza's ruins had no target and two in three of its dead were not fighters

Israel keeps two counts of its war in Gaza. The army says how many "targets" it has struck: 29,000 by February 2024, 40,300 a year into the war, 19,500 more in 2025, 65,000 in all by this month. And it says how many of the dead were combatants: 17,000 by October 2024, 22,000 by the ceasefire of October 2025, more than 24,000 now. Two other counts exist. The United Nations counts Gaza's buildings from satellites: by October 2025, 123,464 had been destroyed and 198,273 destroyed or damaged to some degree, four in five. Gaza's health ministry lists the dead by name, age and sex: 74,250 by this month. In January a senior Israeli officer told reporters that the army believes the ministry's toll is about right; the army's spokesman said this was not official data, and the army has published none of its own. Set Israel's counts against the other two and the war divides in half. One half is the war Israel describes: targets struck, fighters killed. The other half is destruction and death for which Israel has given no military reason at all.

Arguments about Gaza run on anecdote: a video, a general's quote, a tunnel under a hospital, a number repeated until it sounds counted. We did the dull thing. We took Israel's own figures for what it struck and whom it killed, gave Israel the benefit of every doubt, and set them against the satellite count of destroyed buildings and the ministry's count of the dead, which a senior Israeli officer has called largely accurate. Then we built Israel's own explanation for the rest, that Hamas was in the buildings, as a simulation, and let it run with every lever on Israel's side. Even so, half of the buildings and two in three of the dead are left without a reason.

The buildings

Take the army's count at its word, and take every target to be a different building, which flatters it: a target can be a person, a vehicle or a rocket launcher in a field, none of which is a building. By October 2025 the army had reported 55,000 to 61,000 targets in Gaza (its figure for the last quarter of 2024 is unreported; its total of 65,000 by this month sets the most it can be). That covers 45% to 49% of the 123,464 destroyed buildings. The other half, 63,000 to 68,000 buildings, were destroyed without being counted as a target by the army that destroyed them. The latest figures say the same: 65,000 targets against the 134,422 buildings destroyed by June 2026, or 48%.

Could they have had a military reason the army did not count? The usual claim is that Hamas was in the buildings. Grant it in full: every Hamas fighter in a different building, each one a legitimate target, each destroyed for that reason. Nobody knows how many fighters Hamas had. Published estimates of its strength before the war run from 20,000 (American intelligence) to 30,000, and some Israeli reports say 40,000; none can be checked. We take 15,000 to 40,000, to cover them all. At 40,000, a fighter in every building explains at most 32% of the destroyed buildings; at 15,000, 12%. The army's claims about infrastructure are smaller still: 4,700 tunnel shafts located by October 2024, which is 3.8% of the destroyed buildings; 270 weapons depots and 14,000 undefined "terrorist infrastructures" in 2025, 11%. Count every one of them as a separate building that no strike had already covered, which is more than generous, and the army's account reaches 60% to 64% of the destroyed buildings. Half of what Israel destroyed has no target in its own count. At least a third has no reason of any kind on record.

When the other half fell

The satellites say when. In the war's first four months the army reported more targets than the satellites found destroyed buildings: 29,000 against 26,000, or 1.12 targets for every building. Whatever one thinks of those strikes, the army's count can cover them. Then the counts part. From February to October 2024 the army added 11,300 targets while 29,000 buildings were destroyed. From October 2024 to the end of 2025 it added at most 25,000 targets while 73,000 buildings fell: 0.34 targets for every newly destroyed building. And the war was getting quieter by every other measure. ACLED, which logs violent events from press reports, counts 3,800 deaths a month in Gaza in the war's first year and 1,500 a month in the fifteen months after it. The destruction did not slow: 4,600 buildings a month in the first year, 4,900 a month after it.

What levels buildings at that rate with so few strikes is demolition, and the evidence for it comes from Israel as much as from Gaza. The army has said since the war's first winter that it demolishes buildings that sit over tunnel shafts or are rigged with explosives; its own summary of 2025 lists 19,500 targets struck and, separately, 14,000 "terrorist infrastructures" destroyed. The Centre for Information Resilience counted 17,916 buildings in eastern Rafah: 22% were destroyed by 1 February 2025, during a ceasefire, and 68% by 11 May, after the army cut its Morag corridor across the city, with excavators and timed detonations on film. The kilometre-wide belt along the border held some 2,850 buildings, by an Israeli geographer's count for the Associated Press, and was being cleared from the war's first months. Israeli soldiers have described the routine. One guarded four or five bulldozers that flattened 60 homes a day. A bulldozer operator boasted of 50 buildings a week: "Instead of sending soldiers to clear houses and risk their lives, we destroy the houses from outside."

A demolition can be lawful. A house rigged with explosives, or one that sits on a tunnel shaft, may be destroyed. But the army has never said which of the buildings it levelled were those, has published no count of them, and does not count them as targets. The half of Gaza's destruction that Israel cannot account for is, by the timing, the half that came by bulldozer.

The people

The same arithmetic applies to the dead, and again Israel's own figure does the work. The army says it had killed 22,000 combatants by the October 2025 ceasefire, when the ministry's toll stood at 67,000. It now says more than 24,000, against a toll of 74,250. By its own count, then, 50,000 of the dead, two in three, were not fighters. The health ministry's lists give an age and a sex: among 74,250 dead there are 20,179 children and 12,500 women, 44% of the total, who were not fighters on any definition. Of the 41,571 men, if every one of the army's 24,000 combatants was among them, more than four in ten were not fighters either.

"Human shields" does a great deal of work in opinion pages and none at all in the arithmetic: the army's own count of combatants is the same number whoever was standing in front of them.

Even the 24,000 is generous. The army counts "operatives", a word it does not define. Under the laws of war, belonging to Hamas's civil administration does not make a person a target; fighting does. If the count includes police, clerks or officials of the Hamas-run administration, as the word allows, the share of the dead who were fighters is lower still, and the share who were not is higher than two in three.

The chase that does not add up

"Hamas was everywhere" is usually offered as a sentence. We offered it a grid.

Could the half without a target be the price of chasing fighters who move? A fighter who uses five buildings over the war might seem to justify five. We tested the claim with a simulation: Gaza as a grid of 245,000 buildings; fighters placed in part of it; each month the real number of buildings destroyed that month, from the satellite count, handed out first to any fighter the army has found; a found fighter's building destroyed, the fighter dead with some probability; survivors moving, recruits arriving. The unknowns, how many fighters, how much of Gaza they held, how often the army found one and how often a strike killed him, were not chosen. We drew 3,000 combinations at random and kept the 150 that best match three things reported about the war: how many fighters Hamas still had, an American intelligence assessment, reported by Reuters, that recruitment had roughly replaced losses by early 2025, and a toll of dead fighters that fits the age and sex of the dead. On the first the reports disagree, so we ran it twice: once with the 20,000 to 23,000 fighters that Israeli television reported at the start of 2025, and once with ACLED's judgment that a year into the war Hamas kept at most half its men.

The chase explains 4% to 53% of the destroyed buildings, most likely 14% to 17%. The floor is firm; the figure in the middle leans on an assumption, and the floor does not. The top depends on one thing no data settles: how much of Gaza Hamas held. And the model carries a lesson for the army's defenders. Make the army good at its job, finding most fighters quickly and killing them, and it explains less, not more: 63% of the first month's destruction, then less and less, because by the ninth month four-fifths of the fighters are dead and the 88,000 buildings destroyed after that have almost nobody left to chase. By mid-2026 the share is 9%. An army that was precise and a war that had to go on for three years cannot both be true of the same rubble.

Is it genocide?

That is not a question this publication can answer, or tries to. Genocide is a crime of intent, and intent is for courts and for the international bodies that gather evidence of it. An independent commission of inquiry set up by the UN's Human Rights Council concluded in September 2025 that Israel has committed genocide in Gaza, a finding Israel rejects; the International Court of Justice has South Africa's case before it and has not ruled. Anyone else who tells you the numbers settle it, either way, is selling something. What the numbers here can say is narrower. With every lever set Israel's way, every target a different building and every one of the army's 24,000 combatants a fighter, half of Gaza's destroyed buildings have no target in the army's count, at least a third have no reason of any kind on record, and two in three of the dead are not fighters by the army's own count. The explanations Israel has offered are tested against Israel's own figures, and they fail by half. What the remainder was, and why, is for others to judge.

Israel's own numbers draw the line. On one side are 65,000 targets and 24,000 fighters: the war the army describes. On the other are almost 70,000 buildings for which it has named no target, at least 40,000 for which it has given no reason of any kind, and 50,000 people it does not call fighters. The second war is the larger one. It is the one Israel has never explained.

Every number in this article, every assumption behind the simulation and every version of its specification is set out in the full report, with the methods as an appendix: download the report (PDF). The data behind every chart is free to reuse with credit.

This article was corrected on October 11th 2026 after a check of its sources. An earlier version said seven in ten of the dead were not fighters by the army's own count; with the army's latest count of combatants the share is two in three, as it was at the ceasefire. It said the army accepts a toll of about 70,000; that figure came from an anonymous briefing which the army's spokesman said was not official data. It credited an estimate of 20,000 to 23,000 remaining fighters to ACLED; the figure is Israeli television's, and the simulation has been run again on both that reading and ACLED's own (4% to 53%, most likely 14% to 17%; it was 5% to 54%, most likely 14%). The army's newer total of 65,000 targets moves one ratio from 0.36 to 0.34. The finding that half of the destroyed buildings have no target in the army's count did not change. Details are on our corrections page.