The account of 7 October has fallen apart, and what is left is Israeli propaganda

The fuller answer

Start with the fact the whole argument turns on. Most of the record of 7 October was made by the people who carried out the attack.

They wore body cameras and helmet cameras, and filmed on their phones. They broadcast live from inside houses they had broken into, in at least one case from the victim’s own Facebook account while her relatives watched the feed. The military wing’s Telegram channels published footage the same day. A cameraman whose stream carried a Saudi channel’s watermark filmed abductions at Nir Oz as they happened, and can be heard telling the attackers to keep a captive alive.

That changes what a denial has to do. An account resting on Israeli testimony can be met by doubting Israeli testimony. An account resting on what the accused filmed and published cannot. Anybody arguing the record was manufactured has to say who manufactured it, why the attacking side distributed it in real time, and how it then survived forensic examination by people with no interest in helping Israel.

This entry is about how the day is known. For what happened, and the claims to leave alone, use the guide on the 7 October attacks. For what Hamas is, use the entry on resistance.

Who examined the material, and under what conditions

The UN Commission of Inquiry started collecting on the morning of 7 October. It preserved thousands of open-source items to the standards used for digital evidence, cross-referenced them against other sources, and where needed applied visual media authentication, alteration detection, metadata extraction and facial recognition. It geolocated footage against satellite imagery and dated it by analysing shadows and weather.

Here is the part to say out loud. It did all of that while Israel refused to cooperate with it, reportedly instructed medical professionals not to speak to it, and denied it entry to the country. Whatever else that refusal cost, it means the Commission’s conclusions about the visual record cannot be attributed to Israeli management of the evidence.

Those conclusions are blunt. Hamas published its own account in January 2024, claiming its fighters targeted only soldiers and armed civilians. The Commission recorded that this contradicts hundreds of individual items of visual material it had documented, and that CCTV and bodycam footage clearly show deliberate killings, mistreatment and abduction of civilians in every attacked community and at the Nova site.

Then the point most people have never been given. Al Jazeera’s investigative unit made its own film on that day from seven hours of footage taken from CCTV, dash cams, phones and the head cameras of dead Hamas fighters. The network your opponent is most likely to trust worked from the perpetrators’ own recordings and treated them as authentic.

And judges outside Israel have looked at it. On 21 November 2024 Pre-Trial Chamber I of the International Criminal Court found reasonable grounds to believe Mohammed Deif was responsible for murder, extermination, torture, hostage-taking, and rape and other forms of sexual violence.

The distinction that decides this: a correction is not a collapse

Concede the corrections first, and name who made them, because that is the part your opponent expects to have to drag out of you.

Israel revised its own death toll down from about 1,400 to about 1,200. The UN Special Representative’s mission team examined the claim that a pregnant woman’s womb had been cut open and her foetus stabbed, and found it unfounded. It established that a girl’s body found naked from the waist down had been moved by a bomb squad, which explained the scene. The two ZAKA volunteers whose accounts travelled furthest were contradicted by their own colleagues, by Haaretz and other Israeli outlets, and by UN investigators. The Special Representative’s report also notes that witnesses became more cautious over time, in some cases retracting statements they had made.

Every one of those withdrawals was made by the UN’s own team, by Israeli media or by Israeli first responders correcting each other. None was extracted by a critic. Say that plainly, because it is the answer to the whole move.

The move is this. Take the early claims that did not survive, which exist after every mass atrocity in every conflict, and use them to retire the body of evidence that did survive. It works because the two are never held apart. Hold them apart out loud and it stops working. A record that corrects itself is a record under test. The account being offered against it has not been tested by anybody, least of all by the party that wrote it.

Sexual violence: exactly what was found, and exactly what was not

This is where overstating costs most, because the correction comes back at you holding the report you cited.

Pramila Patten’s mission ran from 29 January to 14 February 2024. Her team, including a forensic pathologist and a digital analyst, reviewed over 5,000 photographs and around 50 hours of footage from militants’ bodycams and dashcams, phones, CCTV and traffic cameras, held 33 meetings with Israeli institutions and conducted 34 confidential interviews.

It found reasonable grounds to believe that conflict-related sexual violence occurred at several locations across the Gaza periphery, including rape and gang rape, and named the Nova site and its surroundings, Road 232 and kibbutz Re’im. It found clear and convincing information that hostages taken into Gaza were subjected to sexual violence, with reasonable grounds to believe it might be continuing.

Now the limits, in the same breath, every time. The mission was expressly not investigative and was not a substitute for one. It could not establish prevalence, and said the true extent may never be fully known. It could not substantiate any of the alleged pamphlets instructing fighters to commit rape. It could not attribute violations to specific groups. And it recorded that much of its information came from Israeli national institutions, because Israel had refused access to the UN bodies with an investigative mandate.

The Commission of Inquiry, denied that access, could not independently verify the rape allegations at all, and said why: no access to victims, witnesses or crime sites, and obstruction by the Israeli authorities. It also noted the absence of forensic evidence, since the priority afterwards was identifying and burying bodies. That is an evidentiary limit, not a finding that nothing happened, and the same Commission verified the deliberate targeting, abduction and abuse of women and the desecration of women’s bodies.

Carry both halves or do not carry it.

The Hannibal claim, honestly

Do not dismiss this one and do not swallow it. Learn where the line falls.

What is documented. The Commission of Inquiry verified that in at least two incidents Israeli fire likely killed up to fourteen Israeli civilians: a woman killed by helicopter fire while being taken from Nir Oz towards Gaza, and thirteen hostages in the Pessi Cohen house at Be’eri, killed by tank shelling and crossfire. It viewed a televised statement in which a tank crew describe knowingly applying the Hannibal Directive against two vehicles near Nir Oz. Haaretz reported in July 2024 that Hannibal-type orders were issued at the Erez crossing, at Re’im and at Nahal Oz, and that an order went out that no vehicle should return to Gaza. An Israeli army review reported by N12 acknowledged multiple incidents of Israeli forces firing on Israeli forces. Israel’s own Be’eri probe conceded that a tank shell killed the hostage Adi Dagan, while concluding that most of the hostages in that house were murdered by the attackers and that the commander’s decisions were reasonable. That disagreement, between the Commission and the Israeli army, is unresolved, and you gain nothing by pretending otherwise.

What is not documented. There is no verified count in the hundreds. The Commission confirmed attack helicopters were over the area but was unable to confirm they fired at anything at the Nova site. The figure claiming all 364 Nova dead were killed by Israeli aircraft was asserted by the Palestinian Authority, rested on an unnamed Israeli police source quoted by Haaretz saying a small number of people might have been harmed by helicopter fire, was denied by the police, and was deleted by the PA. The larger totals in circulation are built by inference from shot-up vehicles and unexplained deaths, not from any body count.

And what it would change if every maximal version were true. Deaths caused by Israeli fire are Israel’s to answer for, and the families at Be’eri are entitled to an account they have not been given. What those deaths do not do is convert an attack into something other than an attack. Nothing in the Hannibal file touches the footage.

“It was a prison break that got out of hand”

The perpetrators’ own material answers this faster than argument does. Hamas’s spokesman said the operation was first proposed in 2021. Planning ran to thousands of fighters. On the morning itself Deif broadcast a call to Palestinians across Israel to torch the earth, kill, burn and destroy. Hamas figures said within days that taking hostages was an objective. And in bodycam footage verified by the Commission, fighters driving past kibbutz Sufa recognise it, note that the people look like civilians, and go in.

When somebody produces a genuine retraction

They will, and some will be right. Do this.

Agree, immediately and without hedging. Name it first if you can, because a correction you supply is worth more than one you concede. Then ask the only question that matters: what does that change about the material that was not retracted? Make them answer about the bodycam footage, the livestreams, the hostages who came back and testified, and the nineteen Thai and Nepalese farm workers killed at Alumim.

Lines you can say

Nobody follows these mechanically. Take the shape and use your own words.

They filmed it themselves and published it. Who do you think faked that?

Al Jazeera built its own film out of the attackers’ head-cam footage. If that’s fabricated, tell me who fabricated it.

The UN commission that authenticated the footage was the one Israel refused to let in. That’s why its findings are worth something.

The pregnant woman claim was withdrawn by the UN’s own team. I don’t use it. Which of the findings that stood do you dispute?

Israeli fire did kill Israelis that day. The verified count is fourteen, not hundreds. What does either number change about the footage?

Where this backfires. Reciting evidentiary standards at somebody who has spent a year watching Gaza coverage reads as bloodlessness, and you will lose the room before you finish the sentence about metadata. Say what you feel about the day in your own words first. And never claim more than the reports say to win a moment, because the person across from you may have read them, and being corrected with your own citation costs you every true thing you said before it.