The 7 October attacks: what you can state, and what to leave alone

Before you read the sources

Two warnings, and then you can start.

The material behind this page includes forensic description and footage taken by the attackers. Decide now how long you will spend with it, and stop at that point rather than at the end of the document. Nothing here requires you to watch anything.

And the one that matters more in public. A figure you cannot source is worse than no figure at all. That is not caution, it is the whole method here.

You need nobody’s permission to do this. One evening with the two UN documents below, and a page of notes, puts you ahead of most people who will argue with you.

The rule this page is built on

The record of 7 October that is documented beyond serious dispute is already enormous. It does not need help.

In the weeks afterwards a number of specific atrocity claims circulated, were repeated by senior people, and were later retracted or never substantiated. Each one is now a free win for anybody who wants to change the subject. Worse, it lets them argue from your one bad claim to the whole body of evidence. That move has been made against this case repeatedly since.

So: use only what is documented, and say plainly that you are doing so. Declining to repeat a discredited claim buys you more credibility in a room than any amount of force.

What you can state, with a source

Israel’s official death toll is around 1,200. It was revised down from about 1,400 in November 2023, after forensic work established that some of the unidentified bodies were attackers rather than victims. Say the 1,200 figure and say that it was revised, because a correction offered before it is demanded reads as confidence.

The UN Commission of Inquiry, which is not a body friendly to Israel, recorded the following. More than 1,200 people killed, at least 809 of them civilians. Among the civilian dead, at least 280 women, 40 children and 25 people aged 80 or over. Sixty-eight were foreign nationals. At least 314 Israeli military personnel were killed.

Almost 15,000 people were injured and taken to hospital. At least 252 were abducted into Gaza as hostages, among them 90 women and 36 children. At least 1,000 attackers crossed into Israel.

At the Nova music festival, the Israeli military’s investigation found 364 people murdered and 40 more taken into Gaza.

The foreign dead matter to the argument and are usually left out of it. In early November 2023 Human Rights Watch recorded 32 Thai nationals killed and 22 taken hostage. Also ten Nepali agricultural trainees, four Filipino carers and one Cambodian student. Later Thai government counts were higher. These were farm workers and students, and no account of that day as an act against occupation explains them.

Why the core record holds: they filmed it themselves

This is the most important evidential fact about 7 October, and most people arguing about it have never had it explained to them.

The attackers recorded their own conduct. The UN Special Representative’s team reviewed more than 5,000 photographs and around 50 hours of footage. It came from the militants’ own body cameras and dashcams, from mobile phones, from CCTV and from traffic cameras. Some was broadcast live, in at least one case from a victim’s own social media account.

The Commission of Inquiry put it directly. Hamas issued statements denying that its fighters targeted civilians. The Commission noted that those statements contradict hundreds of items of visual material it had documented. CCTV and bodycam footage, it found, clearly show deliberate killings, mistreatment and abduction of civilians in every attacked community and at the Nova site. In one verified bodycam video, the men identify a residential kibbutz, note that the people look like civilians, and go in.

That is why the core account is not seriously disputable. It does not rest on Israeli testimony, or on journalism, or on your good faith. It rests on what the perpetrators chose to film.

Sexual violence: state it exactly, not maximally

This is where accuracy earns the most and overstatement costs the most. Learn these findings in the words they were written in.

Pramila Patten, the UN Special Representative on Sexual Violence in Conflict, led a mission to Israel from 29 January to 14 February 2024, reporting on 4 March. What it found:

• Reasonable grounds to believe that conflict-related sexual violence occurred in multiple locations, including rape and gang rape in at least three: the Nova festival site and its surroundings, Road 232, and kibbutz Re’im.
• A pattern, across locations, of bodies found naked or naked from the waist down, mostly women, hands tied, shot repeatedly and often in the head. The report calls this circumstantial, and indicative of some forms of sexual violence.
• Clear and convincing information that some hostages taken into Gaza were subjected to sexual violence including rape and sexualised torture, with reasonable grounds to believe it may have been continuing.

And what the same report says it could not establish, which you state in the same breath:

• It could not establish the prevalence of sexual violence. The mission ran two and a half weeks and was not investigative in nature.
• At kibbutz Be’eri it found at least two widely reported media allegations to be unfounded, and was unable to establish whether sexual violence occurred there at all.
• At Kfar Aza and at the Nahal Oz base, reported incidents could not be verified.
• No discernible pattern of genital mutilation could be established.

Say all of it. A sentence that carries the findings and the limits together is one nobody briefed can take apart. The maximal version is.

What to leave alone, and why

Name these before somebody names them at you.

“Forty beheaded babies.” This began with a broadcast report on 10 October 2023 relaying what soldiers had said. It was repeated at the highest political level, then walked back. An Israeli official told CNN they could not confirm the age, sex or status of decapitated victims. Decapitation and mutilation of bodies are documented. That is a different claim, and the difference is the point.

“A pregnant woman’s womb was cut open and her foetus stabbed.” Patten’s team identified this as one of the Be’eri allegations that was unfounded. A colleague who was at the scene contradicted the volunteer who told it.

“Fighters carried written orders to commit rape.” Patten’s team was unable to substantiate any of the claimed pamphlets. The Commission of Inquiry could not obtain them or verify their authenticity.

“Fourteen hundred dead.” Superseded in November 2023. Using it now says you stopped reading in the first month.

A body found with clothing displaced, offered as proof of assault. One such widely repeated account was explained by soldiers having moved the body to check for booby traps. Circumstantial indications belong in the argument as circumstantial. The UN reports treat them that way.

The cost of repeating any of these is not the correction. It is that everything true you said before it goes with it.

Genuine dispute, and the other kind

You will be told that “even the UN couldn’t verify the rape claims”. Some of that is a real disagreement and some is not. Telling them apart is a skill worth building.

The real one: the Commission of Inquiry stated it could not independently verify rape allegations. It cited lack of access to victims, witnesses and crime scenes, and obstruction of its investigation by the Israeli authorities. That is an evidentiary limit, not a finding that nothing happened. Patten’s mission, which did have access, reached findings the Commission could not.

The bad-faith version treats every unestablished detail as proof that the established ones are fabricated. The answer is not volume. It is to ask which specific finding they dispute, and then to read it out.

When the day is framed as a consequence

You will meet “this didn’t happen in a vacuum”. Do not accept the framing and do not refuse to engage, because refusing looks like you have no answer.

Separate causation from justification, out loud, and then hold them to the record. Attacks have histories. Intentionally killing civilians has no context that makes it defensible.

Then the specifics. The attackers’ own footage shows them identifying civilians as civilians and going in anyway. The commander of the al-Qassam Brigades announced the operation as a call to liberate all of Israel by killing and burning. Those are the perpetrators’ words about their own purpose.

Where this backfires: delivered as a debating manoeuvre, it sounds like you will not discuss Palestinian suffering at all. Say the history is real and that you will discuss it, then say why it does not do the work being asked of it.

Four lines you can say

“I’ll only use what’s documented. I don’t need the rest, and neither does the case.”

“They filmed it themselves. That’s why this isn’t a matter of whose word you take.”

“The UN Special Representative found reasonable grounds to believe rape and gang rape happened at three named sites, and said she couldn’t establish how widespread it was. Both halves are hers.”

“That claim didn’t hold up and I’m not going to use it. Which of the findings that did hold up do you dispute?”

Tonight, with the documents open

Fill in the blanks178 words

Two questions from memory first. Which figures on this page can you state with the document they came from, and which can you only state? And which widely repeated claim about that day did not hold up, so that you never carry it into a room by accident?

Then open the two UN documents in the sources list and write down five figures with the paragraph each came from. That is the whole preparation, and it is what lets you answer a follow-up question instead of repeating yourself.

Then: if somebody puts 7 October to me in public, then I will state the toll, the hostage figure and the fact that the attackers filmed it, before I say anything about sexual violence.

And a coping plan, because the likeliest obstacle is not a counter-argument. It is somebody throwing a discredited claim at you as though it were yours: if that happens, then I will say that the claim did not hold up, that I have never used it, and ask them which documented finding they are actually disputing.