Skip to content
Shoresh Wiki

Are 83% of Gaza's dead civilians, per 'the IDF's own data'?

“Israel’s own intelligence shows 83% of the people it killed in Gaza were civilians.”

You have seen this number presented as a confession, “even the IDF’s own database admits it.” That framing is the whole trick. Here is what the database is, what the division actually divided, and how far the result travelled.

Where the number comes from

Section titled “Where the number comes from”

In August 2025, journalist Yuval Abraham of +972 Magazine and Local Call, with the Guardian, reported on a leaked Israeli military intelligence database. As of May 2025 the database listed roughly 8,900 named Hamas and Palestinian Islamic Jihad fighters as confirmed or “probably” dead. The journalists set that against the Gaza health ministry’s total of about 53,000 dead, called the named fighters 17 percent, and published the remainder as the headline: 83 percent civilians.

What the division actually divided

Section titled “What the division actually divided”

The numerator is a list of the dead the army had individually identified by name as fighters, from two organizations. It excludes, by its own construction:

  • Unidentified fighters. The investigation’s own text concedes “the total number of militants killed is likely higher than the number recorded.” Men killed in airstrikes on tunnels, in collapsed buildings, or in units that stopped reporting do not get named in an intelligence file. In every war, naming the enemy’s dead one by one lags killing them; treating the named subset as the complete count guarantees the “civilian” remainder is inflated.
  • Every other armed faction. The PFLP, the Popular Resistance Committees, the al-Aqsa Martyrs’ Brigades, the DFLP, the PFLP-GC and the Palestinian Mujahideen Movement all fought and died in Gaza. None of their dead are in a Hamas-and-PIJ file, so all of them land on the “civilian” side of the ledger.
  • The military-adjacent apparatus. Political-wing officials, police who fought, logistics and rocket-storage personnel: not named as fighters, counted as civilians.

The denominator has the opposite problem. The health ministry’s list is a raw total that never distinguishes combatants from civilians, and a December 2024 Henry Jackson Society audit by Major (ret.) Andrew Fox documented age and gender misclassifications, pre-war and natural deaths, and deaths caused by Hamas’s own misfired rockets inside the count. Divide a padded total by a deliberately partial subset and the output is not a civilian share. It is a floor for fighters dressed up as a ceiling.

By mid-2025 Israeli assessments and independent analysts put combatant deaths at 20,000 or more, which against the same denominator yields a civilian share in the range of 55 to 65 percent, in line with, or below, the historical norm for urban warfare. You do not have to take the IDF’s estimate on faith to see the point: 8,900 named dead fighters is evidence for at least that many, not evidence for at most that many.

How far the lie travelled

Section titled “How far the lie travelled”

The headline did exactly what it was built to do. In the three days after publication, August 21 to 23, 2025, posts pairing “83%” with Gaza generated roughly 96,700 posts and 759,000 interactions with a potential reach in the billions, and the figure entered the debate circuit as “the IDF’s own data,” a phrase the reporting itself never earned: the IDF said the figures were incorrect, and the database was leaked, not published. The correction, the part where the investigation admits the fighter count is incomplete, appears in paragraph after the headline and travelled nowhere. That asymmetry, a viral number and a buried caveat, is the signature of every claim on this shelf.

A floor is not a census

Section titled “A floor is not a census”

Keep the whole thing in one sentence: the investigation counted the fighters Israel could name, subtracted them from a list that names no fighters at all, and called everything left over civilian. When you hear “83%, from Israel’s own database,” you now know the database counted two factions by name and its own authors conceded the count was low. The number that survives scrutiny is the one the method actually supports: the named dead are a minimum, the real combatant toll sits far above it, and the civilian share of Gaza’s dead, grievous as it is, is a fraction the 83% figure was engineered to overstate.

Section titled “Related pages”

Sources

Section titled “Sources”