Call Me Back - with Dan Senor - Data Science Vs Hamas Math - with Abraham Wyner

Episode Date: April 10, 2024

30,000. You hear that number and you already know exactly what we are referring to. It’s 30,000 casualties. That’s the number of Palestinians that have been killed in Gaza as a result of the IDF response to the October 7th invasion of Israel, according to the Gaza Health Ministry. Of course, we don’t know how the Gaza Health Ministry arrived at that number. How does it collect this data, analyze it, and how does it account for civilian casualties versus Hamas terrorists? It’s a big round number that everyone - from news reporters, to aid organizations to governments - mindlessly repeat. Well, a data scientist at the Wharton School at the University of Pennsylvania has taken the time to try to understand how these numbers are computed. He published his study in a piece in Tablet Magazine — it’s called “How the Gaza Ministry of Health Fakes Casualty Numbers”. You can find it here: https://www.tabletmag.com/sections/news/articles/how-gaza-health-ministry-fakes-casualty-numbers Abraham Wyner is Professor of Statistics and Data Science at The Wharton School of the University of Pennsylvania and Faculty Co-Director of the Wharton Sports Analytics and Business Initiative. Professor Wyner received his Bachelor's degrees in Mathematics from Yale University, where he graduated Magna Cum Laude with distinction in his major. He was the recipient of the Stanley Prize for excellence in Mathematics. His PhD in Statistics is from Stanford University.

If you just take Hamas's number of casualties, 30,000, and call them all civilians, or almost all civilians, then you're looking at an extreme imbalance of civilians to fighters, and that's the public perception. If you just take Hamas's number, 6,000, you still have what we call the missing men problem. You either have essentially no civilian men dying, or the idea that all civilian men are Hamas. And neither of those are true. So to start with, without even doing an analysis of looking at the data, you've got a problem. 30,000. You hear that number and you already know exactly what I'm referring to. It's 30,000 casualties. That's the number of Palestinians that have been killed in Gaza as a result of

the IDF response to the October 7th invasion of Israel. So we're told. Of course, we don't know how the Gaza Ministry of Health has arrived at that 30,000 number? How does the health ministry collect this data, analyze it? How does it account for civilian casualties versus Hamas terrorists? 30,000 is a big round number that everyone from news reporters to aid organizations to governments mindlessly repeat. And I say mindlessly because it also accounts for 13,000 Hamas terrorists that have been killed. And while the remaining 17,000 civilian deaths are extremely tragic, a ratio of 1 to 1.4, especially in a battlefield designed to optimize the number of civilian deaths by Hamas, is actually much lower than other counterinsurgency wars of this kind

throughout history. Now, we have been skeptical of both the overall number and the composition of the casualties. For the longest time, we weren't the only ones skeptical. Here was President Biden in late October. What they say to me is I have no notion that the Palestinians are telling the truth about how many people are killed. I'm sure innocents have been killed and it's the price of waging a war. But I have no confidence in the number that the Palestinians are using. And here was White House National Security Council spokesman John Kirby from November. We don't have the capacity or the capability to provide you an exact figure.

We still don't believe that taking the Ministry of Health's numbers at face value is wise, but we can't provide you with an alternative number. However, recently, ever since President Biden made an about-face and began citing the Gaza Health Ministry numbers without any caveat or qualification at the State of the Union address, the numbers are now cited everywhere, also without caveat or qualification. Well, one data scientist at the Wharton School at the University of Pennsylvania has taken the time to try to understand how these numbers are actually computed. He published his study in a piece in Tablet Magazine, which we'll post in the show notes. It's called How the Gaza Ministry of Health Fakes Casualty Numbers. Abraham Weiner is his name.

He's a professor of statistics and data science at the Wharton School, as I said, and he's the faculty co-director of the Wharton Sports Analytics and Business Initiative. Professor Weiner received his bachelor degree in mathematics from Yale University, where he graduated, and he received his PhD in statistics from Stanford University. Abraham Weiner on data science versus Hamas math. This is Call Me Back. And I'm pleased to welcome to this podcast for the first time, Professor Abraham Weiner, Professor of Statistics and Data Science at the Wharton School at the University of Pennsylvania. Abraham, thanks for being here. It is great to be here. He joins us from his office in Philadelphia at Penn. So, as I mentioned in the intro, Abraham, you wrote this very thought-provoking essay,

piece of analysis for Tablet Magazine, and because your data science skills are probably stronger than mine, I maybe read it a third time to try to understand all the calculations and analysis you were presenting. And then through our mutual friend, Alana Newhouse, I got a hold of you and said, I am eager to have you on this podcast to dig deeper because this question of statistics coming out of the Gaza Ministry of Health is one. It's like a character that the numbers, obviously there's the human catastrophe that the numbers represent, but then there's just the story of the numbers and then how the numbers are used and are the numbers legitimate or are they not legitimate and how do we make sense of the numbers? And then you come out with this piece that in many respects raises a lot of questions and in other respects is clarifying. So the number of civilian casualties in Gaza has been at the center of international attention since the start of the war. And it's been the key metric in which the

international community has been relying on to take action against Israel. Claims of genocide in the International Court of Justice, decisions on a possible weapons embargo against Israel. And they just repeatedly were heard over and over, just come down to this one number. We hear it all the time, somewhere in the approximate area of 30,000 Palestinians, now 31,000 or 32,000 that have been killed in Gaza since Israel, since October 7th, since Israel's response. And at first, the numbers were discredited by the Biden administration. And the more recently, they've been removing their caveats that has had all sorts of implications. And anyways, the numbers are presented on a silver platter to the media, all by the Hamas-controlled Gaza

Ministry of Health. So you decided to tackle this issue. And I guess my first question is, when did you start tracking these numbers and why? Okay. So let me start with the why. I'll start by pointing out that the numbers that come out of the Gaza Minister of Health are the total number of casualties. They don't make any effort to separate fighters from civilians. And that is routinely skipped when you see it written about in the media. Sometimes it's carefully acknowledged, this is a mixture of civilian and fighters, and other times it's completely skipped. And I think the important question is, why does that matter? And I think that is important,

especially important, because Israel is fighting a unique battle where it's an urban environment. Hamas is buried underground. It uses humans, civilians as shields, and none of this is new. And the real question is, Israel has to take the war and is taking the war to Hamas and its fighters. And along the way, civilians are going to die. I don't think anybody disagrees or expects anything else. And the question is, in what proportion? And there's something called proportionality in military terms. And it's not literal statistical proportionality, like one to one or a ratio of that nature. That's often misunderstood.

Proportionality means, do the numbers of civilians who are dying, which of course, every one is a tragic occurrence, but is it balanced by the military objective? It's not the numbers. It's not a proportionality game of deaths. It's a military objective. So Israel must clean out Hamas. That's a stated military objective. In a war, your goal is to, of course, defeat the enemy. And the question is, is the number of civilians justified by the loss of the military gain in fighting fighters? So why this is important, Hamas's objective is to get the world to be convinced that Israel

is not attacking its fighters, but it's just indiscriminately bombing. And as a result, its actions are not obtaining a military, justify a military objective, but are just destroying a population, destroying its buildings, cutting off its supplies, killing its civilian population. And if there's no military justification, then why is Israel engaged in this way? So the absolute numbers are one thing, but far, far more important is the division into civilians and fighters. So that's the thing that got me interested.

The thing that got me interested is not the total, which people have suspected might be wrong, might be right. There's been a lot of reference to previous encounters with Israeli army and Hamas. And the acknowledgement has been that the total numbers of casualties have often been in the approximate ballpark when they're straightened out much later. And so there's this assumption that the total number of casualties are probably about right, but very little conversation about the division into fighters and civilians. And if you just go back historically and just dig into the historical record, I'll take the 2009, I believe that was Kastled, I forget.

Late 2008, early 2009. Hamas, I think, reported just under 50 of its fighters having been lost in that battle during the war. And 22 months later, they admitted between 600 and 700, while Israel said it was 700 the whole time. So here you have an extreme discrepancy in a previous encounter. And in that encounter, did they also change the total of non-combatants that had been killed? So they don't do that. So they just release the number of combatants during the war. And then sometime later, a different arm of either the media office or the ministry says,

oh, we've lost this number of fighters. It's almost like two voices coming at each other. There's no backwards updating. One of the reasons why this is hard is they don't wear uniforms, right? So how do you differentiate between a civilian and a fighter? There's no way to do that in a natural way that you would have in a normal war. So that's the issue that got me interested in this. Not the total.

The total is almost too hard to disentangle. I mean, you can try. And some of that, of course, is in my analysis. But the real issue is the division into the civilian and fighters. And I'll give you one number that helps you really understand from the very beginning that there's something wrong. Hamas's Ministry of Health has been reporting almost from the beginning, somewhere between 68 and 72% are women and children. Now, I'm going to ask you, what do you think the percentage in the population is women

and children? It's not a trick question. It's right around 70%. It almost mirrors it to the decimal place. So what we're getting here is a number of total deaths that matches the civilian population almost exactly. And so that immediately has to raise a red flag. And it raised a red flag with me. Now, when I started working on this in early January, I was actually on my way to a faculty mission to Israel with University of Pennsylvania faculty. And I was very bothered by the numbers and their usage in so many conversations, particularly at a college campus. I mean, you probably have some idea what's happening, but

there are enormous numbers of protests going on on campus. And the word is genocide, genocide, indiscriminate bombing. And as I mentioned earlier, this is a war and war produces horrible casualties. But Israel's position is this isn't genocide. This isn't indiscriminate bombing. It's careful, and some would say too careful, war being waged against an entrenched fighter population inside a civilian community buried underneath, which is even worse and even more difficult. Now, I'm no expert on military, of course, but I would suspect that the ratio of fighters to civilians is probably going to be very different than in a normal war, and that you should be looking for something in the order of, I don't even know what it is, but looking to other wars, which I did.

And when I published my article, I discovered that America, when it took on the ISIS fighters in Mosul, their ratio of civilians to fighters was something like four to five to one. And we wonder what that number is for Israel. So if it's all civilians, wonder what that number is for Israel. So if it's all civilians, you have a real problem for Israel. We're waging an ineffective war that is simply rooting out a population and therefore ceasefire and genocide

are not totally inappropriate descriptions. If on the other hand, the casualty ratios are closer to one to one or two to one, then you have a completely different impression of the war that's being waged. And that's what got me interested. Okay. So let me ask you just a process question. How does the Gaza Ministry of Health today, I've read different reports on this, how does the Gaza Ministry of Health today explain how they actually gather this data? If we were to ask the Gaza Ministry of Health,

you say you're gathering this data, how do you actually gather it? How's it done? Okay, so I can tell you what I know, and I would venture that this is a better question for others. So what I know is that in the beginning, they were talking about getting it from hospitals. So they were getting essentially, the bodies were being delivered to hospitals, and they were being identified, and then it was being released publicly. There has been some articles written about it. None of the public data has any information on the time, place, and manner of the death. They are being communicated in some mysterious process.

Certainly mysterious to me. I will say that recently they have a telegram channel, which comes out in Arabic. I don't have telegram and I don't speak Arabic, but other people have written about it. And they actually indicate two separate sources of information. One is the hospitals and the other is what they call the Gaza media office. This is not in my article and it's somewhat new. And I was reading about this online. They are completely different.

The descriptions coming, the casualty counts obviously are different totals because they represent different groups. But the proportions of children and women in the two separate sources are incredibly different. And that's as of today. So I don't know how that actually happens. There's an organization called OCHA, which everybody refers to. And so every article that's been written out about casualties, they were two in the

British medical journal, The Lancet. And there's so many people appearing over this and have found strange contradictions like OC OCHA runs a report talking about women's deaths, and they refer to a UN women's report. And if you go to the UN women's report, it refers to OCHA. It's like this circular, everybody's pointing at each other as you're providing the data, no, you. So there's a real ambiguity about what we're actually getting to see and how. And all this, of course, is nothing to do with my analysis. It's just the background. But I'll even throw out one other thing.

You have this idea that there's somehow this kind of grassroots, scrappy organization doing its best as it can in absolutely horrible circumstances. And that's one story. The other story, of course, is that Hamas is very sophisticated. I mean, what they were able to accomplish on October 7th did not suffer from lack of planning. And I don't understand why there's this, the propaganda arm of this is extremely important. And this notion that somehow they wouldn't be prepared to do this, I think is sort of naive. That's not a data argument, but it's a background argument. Now I'll throw in one other piece of information before we get to any specifics. Finally, sometime in the mid-February, this was actually after I had done my analysis,

but before I had written the article, Hamas admitted that it had lost 6,000 of its fighters. And Israel started, I don't know whether it'd been doing it all along, you might know better than I, but by the end of February, Israel was describing the total number of fighters that it had killed as 13,000. So here you have three sets of numbers. You have the 6,000 admitted to by Hamas. You have the 13,000 that Israel has claimed. They have the 30,000 number that we'll just stipulate is approximately right. I forgot the fourth fact that the percentage of women and children matches the population. They all can't coexist. Something has to be wrong. Okay. I want to come back to that in a moment. But before we do, I just want to give just another example because I am struck by the process piece of this. On October 18th,

there was this bombing of the Al-Hilla Hospital in Gaza City. And it seemed to me like within minutes, the international press led by the New York Times was reporting that 500 Palestinians were killed in this bombing. And by the way, they sourced it to the Gaza Ministry of Health. Right. So did anyone stop to say, how do you know that? How do you know within minutes? First of all, we now know the bombing wasn't the actual hospital. It was a bombing was the parking lot next to the hospital. And obviously it was not an Israeli bomb. It was an errant Palestinian Islamic Jihad bomb that had inadvertently landed there. We now know a lot of things. But even if a hospital was bombed, a building was bombed, that means these hundreds

of casualties were in the building when it collapsed. The idea within minutes they could dig through the rubble and do an accurate count. I was just thinking it through at a practical level. How do they come up with this number so quickly? And everyone just ran with that number, the press, and no one paused to say, how do you know that? Before we even get into disputing and analyzing, like literally, practically, how do you know that? They didn't know that. And they had no idea. What you're describing is a process by which they can say just about anything. And if it makes Israel look bad, it'll get run with. A hospital parking lot was hit, yet they claimed that it was the hospital itself.

Now, if you imagine a whole hospital was destroyed, you can kind of estimate how many people died. That's not even what happened. But I have to say, I remember before it was disputed, the Israeli Jewish community that I speak to was horrified by the bombing. And they're like, how could we be doing this? And my reaction to the people who were asking me about this was, can't you just wait just 12 hours, 24 hours to get a reaction from Israel, to get some more information? What's with the quick trigger? And this goes not only to the New York Times and the mainstream media, but also people who should know better. Data is part of the war machine. And you got to recognize that data is war in the information age. Okay. So now let's go to your findings.

What about the reported numbers? You talked about how they're describing the composition of those casualties, meaning women and children relative to the rest of the population of casualties tracks the actual population perfectly. So was that the primary detail or methodology or red flag that defied mathematical logic? That was it. So I will say that's really important because when you look at circumstantial evidence, circumstantial evidence has to be viewed in the context of sort of what you might call your posterior, or start with a prior understanding of our beliefs, and then you have some hard facts and that gets adjusted and you get what's called a posterior distribution.

And what I'm asking, given all the data that we have, what is my best understanding of the validity of the data? So as I mentioned before, you have a number, say 30,000 total deaths. You have the percentage of women and children, 70% matching the total population. You have Hamas's number of 6,000 fighters and Israel's number of 13,000 fighters. So if you take Israel's number of fighters and Hamas's total number of casualties as accurate, stipulate to that, then you are looking at an overestimation of the number of women and children by a factor of two, which is a huge, huge difference. If you just take Hamas's number of casualties,

30,000, and call them all civilians or almost all civilians, then you're looking at an extreme imbalance of civilians to fighters. And that's the public perception. If you just take Hamas's number, 6,000, you still have what we call the missing men problem. You either have essentially no civilian men dying or the idea that all civilian men are Hamas. And neither of those are true. So to start with, without even doing an analysis of looking at the data, you've got a problem. And so that is the guiding principle. Because if you start with the principle that we may or may not have a problem or what we

call agnostic or what in mathematical terms would be sort of a flat prior or may not have a problem or what we call agnostic or what in mathematical terms would be sort of a flat prior or we'd have no prior understanding of what might be, then circumstantial evidence takes on a very different character. So I started with the supposition that something must be wrong and it must be wrong in the women and children in particular. So that drove my analysis. So let me start by saying, well, what I look at. All right.

So OCHA produces these summaries. And I found summaries online from October 26th through November 10th, I think it is. 16 days in total. And one of the criticisms of my article, which is fair, is it's a very short period of time. The riposte to that is also, I think, fair, which is, but that's all I could get. You know, you can only do what you can do. They were releasing total casualty counts fairly regularly before that period at odd intervals after, and they still release breakdowns, but not daily data. So if you're looking for

problems, my goal was to try to find it in the daily figures. That seemed to me in the effort to push out these numbers quickly, as we described with the hospital fiasco, that that's where we might be able to find something wrong. If you aggregate over a week, then it's easy to just drop them down and then you'll have no evidence. The data science sort of objective is to find something wrong on the data that's coming out daily because that's the best place to look. I mean, this is something that I've done. So you might call this forensic data analysis. And I did a legal case somewhat recently where someone was accused of faking their numbers. And there was an anomaly and strange

anomaly in the digits and the decimal places in this banking data. And I was able to clearly show that they had faked it. So I looked at this data from what I had, which was from the 26th through the 10th. It wasn't perfect data. There was a couple of repeats, but it was pretty good. And that was my starting point. And I was in particular looking to see if there were anomalies in the children and women totals that are reporting. So that's where I started to look at that group. Okay. So tell me what you found. All right. So the first thing that I found, which is something actually I had already noticed in other graphics that had been produced. So up until about October 26th, the total casualty count, which grows every day, obviously, grew and looked like a sort of spotty, jerky way, which you would expect to see with big

days of casualties and followed by small days of casualties. But over this period of time that I was looking from October 26 to November 10, the variability in the total casualties is represented by the total casualties graph, just marched very straight with very little deviation from day to day. Now, there was deviation from day to day, but not the kind of deviation that you would expect to see in a war, where you'd see some days where not very many women and children dying, and other days, large numbers, large numbers of men dying, and depending on what's going on on the battlefield, you just didn't see that variation. It was very small for the entire period, 16-day period.

And that was the first tip-off to me. Now, this, of course, isn't proof. And many people have attacked this argument in particular and coming up for other explanations for it. But it was the first thing that got me thinking, wow, there's something in the daily data that doesn't make sense. It marches too elegantly and too consistently upwards in time. The totals, by the way, the totals do that.

The individual children, women, and men numbers, they bounce around a lot more. But the totals seemed to march up very regularly. And this is not what's expected in a war zone. You should expect way more variability on a day-to-day basis. And this data set that I have has that data. It's inferred because they give you the totals and the number of children, the number of women, then you can infer, well, the number of men from that. But it just was extremely consistent. And meaning there's variability because no two days in a war are the same. No, I mean, if you just talk to what's going on, I mean, the bombings are different, the strategies are different. Well, it's not just that, it's that they've been doing these

mini pauses in the war fighting, Israel, the IDF has. So, you know, more recently, they have certain parts of the day where they're not conducting military operations or at least highly kinetic operations that could result in a lot of casualties. So it doesn't make sense that on those days, that reality is not reflected in the data. The data just marches on in a very linear way as though Israel's fighting the exact same war every single day. Right. The variation was about 15% a day, the standard deviation. I took a lot of shit. I'll just be straight up for the graph that I used, which looks at the total count. And it's very, very linear in the total count.

But if you look at it carefully, you can see the variation. But the argument was people aren't going to look that carefully. But I actually explained what the variation was in the article itself. And this was the graph that was used. In fact, it was the graph that originally tipped me off to there being a potential problem here because it was marching so linear. And in actual contrast to the way it looked prior to this, where it wasn't marching so linearly at all. And that's the period of time that they weren't producing the individual daily casualty statistics broken down by women and children. Again, I'm concentrating on the data

that I had. That was the first finding that was, in some sense, not new, because actually, other people had noticed this too, but in sort of casual conversation rather than in an actual analysis. And that led me to the next question. Well, do you see the same kind of variability? Or do you see sensible variability in the numbers of women and children. And the next thing that I observed, which is, again, completely baffling, is the completely disconnectedness between the number of children and the number of women dying on average on a given day. And the argument would be, again, depending on what Israeli's actions are, and what their battle plans are, and where people are distributed, on days where you don't have very many children death, you should also have very few women death. And days we have a lot of women, you should see a lot of men. That's what we call positive correlation. I didn't see anything. There was no correlation at all. Explain that again. On days where you have fewer women. You should have fewer children because women and children are generally found together.

Right, right. That's one piece. The other piece of it is, again, the Israeli operations. This should be driving the casualty figures. On days where there's pauses and not that many activities, you should see very few. On days when they're very aggressive and very big bombs in big areas, you might see a lot more. And of course, all this produces an enormous amount of variability. And then when you dig down and ask, well, how does that split up? Well, some days you should have very few women and children, and some days a lot women and children. And that's the kind of thing that varied a lot. And even though the overall

total casualties didn't vary so much, but you saw incredible sort of weirdness where some days a lot of children, a lot more children than women, some days a lot more women than children, a lot. And that didn't make sense either. So, and again, I don't know how this happens and you can speculate how it happens, but it's not sensible if these are reporting actual casualty numbers as they happen. And if you were to speculate how it happens? Oh God, that's where you really, when you start to speculate how it happens, that's where you get into trouble because you can't really know. So one way to speculate is that they know the numbers of

deaths and they really just have to divide them. And there are probably clear cases that sort of remains the backbone. But one issue is you probably have children who are also fighters. Where are you putting them? You have men who are not fighters. You need to put them somewhere, right? So how that actually happens, I can't really tell you. But the division across this period of time just lock steps into 70%. And by the way, this, of course, is the problem we observed at the macro level. So you have two things that have to exist simultaneously, the problem at the macro level and the micro level, and they match. And so that's the concern. Now, there's more. They just really emphasize the same thing. The men and the women are very negatively correlated.

And the reason why they have to be is that everything kind of bounces around somewhat in the individual categories. But the overall numbers are really quite steady. And so you have to have this negative correlation, the mathematical formula for variance. Variance should add as you add things together. Children, men, women, when you sum them up to get the total variation, you should get bigger, particularly if they're what I would have expected, positively correlated, you should get a lot of variance total, even if you have smaller variances in the individual

components. But you have larger variances in the individual components and unexpectedly small variance in the total. So that's where the negative has to come in. And that's possibly where the issue is. Again, you cannot know this. And this is why the critics have responded, people who don't want to see this. Well, first of all, they're not even starting with the macro problem. They're not even addressing it. They're just ignoring it. And if you're willing to ignore that, then you can come up with explanations for why

you have this. And most of them are procedural. We don't know what's actually happening on the ground there. And no doubt, it's difficult. So how do we know what's causing the individual numbers? Okay. So now I want to take a step back, Abraham, and just in places like Iraq, during the war in Iraq, I worked for the US government at the early stage of the second Gulf War or the war against ISIS in Mosul, US led war against ISIS. Were there any numbers coming out of those places that were considered credible? Yeah, there were. I mean, I've done some research, not a lot. I've

certainly not made this a subject. I've actually tried to get war data in the interim weeks before I published. I did the best I could. I couldn't get anything and I didn't want to delay publication further. I remember in Afghanistan reports of 10 times as many civilians as fighters dying when the US war in Afghanistan. And the Mosul data, which I cited in my article, was something like four to five to one. Urban battles are terrifying to the civilian population. I think Israel has claimed that it does double knocks on the buildings when they're going to take out something of a tunnel,

tries to clear the civilians, urge them to move. These people will say that's just propaganda. I can't distinguish reality from a claim, but that makes sense. But just to get to what the data is saying, if you stipulate 30,000 dead on the Hamas side and you stipulate 13,000 fighters on the Israeli side, you get a ratio of about 1.4 to 1, civilians to fighters. And while, again, to reiterate, having the infrastructure damage, the lack of food, the deaths and the missingness is terrifying, awful, but it is well, well within accepted boundaries for operations in a war zone of this type. And that doesn't get to the responsibility issue.

The philosophy department will come in and the ethics and morals will come in and they'll weigh in on, well, whose fault is it, right? You ask yourself, well, this is happening because of the way Hamas wages war and they started this. So Israel has tried to make that claim rigorously, but has more or less fallen down because it hasn't been able to make the case that the military proportionality justifies the civilian death. And that's what these figures I think are showing. So since the fraudulent nature of this data is seemingly so obvious, are these numbers being

examined by non-affiliated statisticians in academia besides you? Like, have you seen anyone else saying, wait a minute, something's up here? Okay, so that's a great question because it's very hard to make academic level argument here. I don't have data that's rich enough to do this. It's all circumstantial, all of it. Each piece is circumstantial. You have to look at it together as a whole to make the claim that I make that this isn't right and that there's something fake about

it. Because every piece of it can be explained away. There were two articles in medical journals. One of them I thought was embarrassing in the sense that it claimed that there was no evidence of overcount in Gaza. And it came out in early November and it looked at unreal workers and compared their death rates over time to the overall casualty count over time. And the immediate reading this is that it's a principle in science, which is absence of evidence is an evidence of absence. They just started with that. Say, we went and looked at this data and we didn't find anything. And all they did was show two charts.

They showed a chart of the rates of unruh deaths and a chart of the overall deaths. And they said, look, they're on the same rough order of magnitude. And the unruh deaths, which are very much, much smaller, have been audited, presumably, and therefore we know they're correct. And therefore things don't look untoward. The obvious response to that is, first of all, it's not the total casualty count that we're concerned about. It's the breakdown. So immediately that has to be acknowledged, that what matters is the breakdown. And they didn't even address that point. And the second question is, the breakdown, and these were presumably UNRWA employees

are mostly adult male. And according to the data, we're just not seeing adult males. So there actually should be a huge discrepancy. They didn't even mention that. And then finally, they didn't even mention the possibility that UNRWA workers might be overly represented as Hamas fighters, which eventually Israel pointed out they found 12 of them running around Israel on October 7th. And they claimed that a sizable percentage are actually affiliates of Hamas and potentially are simply dying because they're just

fighters. So all these possibilities were essentially ignored by that. So a good analysis that potentially would be interesting in the level of detail that could warrant a scientific publication might compare other wars, but that data would have to emerge and I don't have it. So if someone could get that, that would be fascinating. But until then, it's really just at the level of speculative statistical analysis. Those who don't want to believe it will just say there's lots of reasons why there can be major errors in the data. There were a couple of obvious errors in the data that I didn't address and other people have pointed them out to me. But even those errors, when you speculate how they can be fixed, lead to more errors, more problems. It's almost like just an

unwinding. Fix one thing, something else happens. Have you been approached at all since the piece was published by any journalists or mainstream publications saying, whoa, we may have been getting this wrong, we need to rethink this, or has it been crickets, whoa, we may have been getting this wrong, we need to rethink this, or has it been crickets? Well, I've certainly been approached by many people, many congratulations from obviously the pro-Israel community, many, many, many hate emails from those who are claiming that I'm minimizing the scope of the tragedy in Gaza, which I am not doing. I tried to be clear about that. I'm not trying to minimize deaths. I'm not trying to acknowledge that there isn't huge damage and that this isn't a terrible war by any measure. I'm making the claim that Israel is fighting

an entrenched enemy in the best way it can and that the civilian lives are part of the war. And fundamentally, war is not a war crime. And that's the point that I'm trying to emphasize. So I have not heard from mainstream media. I've heard from a lot of foreign media who have been very critical of what I've done, mostly by arguing that I'm interpreting the data one way and you can interpret it another. They still don't look square in the eye, the problem of the proportionality. I don't know why it doesn't raise red flags, just about anyone, but it clearly doesn't. If you have ran with this number for a long time, I just don't see how people are going to change their minds.

I have found that people don't change their minds and it's almost impossible to reason them out of something that they haven't been reasoned into. So if you're the kind of group who are willing to run with that Gaza hospital false casualty statistic without checking, then that's someone who's publishing data without really thinking, without reasoning, and my reasoning will never make an impact. So I have not really heard from that. It's been reprinted probably in the conservative press. I don't think that this is a conservative liberal issue, but of course, it has been made into one. Mark Twain once said it's easier to fool people than to convince them that they have been fooled. So even if the actual data could be verified, even if we could get to

some number that all the, let's call them responsible stakeholders, could accept? Does it matter? Does it matter? Well, I guess that's more your bailiwick than mine. The mattering is really the experts in the field. I would argue that if we concentrated on the military proportionality, the numbers of civilians relative to the number of fighters, which is a proxy on the military proportionality, the numbers of civilians relative to the number of fighters, which is a proxy for the military justification, then I think it matters a ton. I think Hamas is trying to claim and its allies in the West that this is an indiscriminate war, which is almost de facto war criminal, war crime, and therefore bordering, if not crossing over into genocide.

Because if you're just bombing for the sake of bombing, I think that's almost textbook, right? But if on the other hand, although it really isn't, I mean, Dresden was bombed indiscriminately, and that was done to bring them to their knees. So even that argument, I think, is a pretty bad faith argument. But I think it matters. I really think it does. And that's, in some level, why I got involved, just because I think this issue matters and people like me should be heard from. Have I been heard from enough? I'd like to think maybe I have. Maybe I've made a difference in some way. I'm certainly taking a lot across the chin for getting involved. Well, I must say, at least to this consumer of your piece, I am grateful that you got involved. I hope you stay involved. I hope you continue to track it, because there will be more and more periods now that will just have even more variances, right? We're going to have, if there is a hostage deal where there'll be a ceasefire, then there'll be probably a period in RAFA of intense fighting. So you'll have real extreme variability. And if the data also does not reflect that, that'll be...

I don't think they're releasing data with the detail that they did over the period of time that I looked at it. We haven't seen it since the middle of November. It comes out sporadically now. All right. I'm grateful for this. It's incredibly informative. We will post the piece in our show notes.

And as you learn new things and see more data, if there is no data that is released, I hope you'll come back on. I'd love to. This has been a wonderful experience. All right, great. Thank you, Professor Weiner. That's our show for today. To keep up with Abraham Weiner, you can find him on X, at Adi Weiner, that's A-D-I, and then his last name, W-Y-N-E-R. And we will post a link to the show notes of his piece from Tablet. I highly recommend it. Call Me Back is produced and edited by Ilan Benatar.

Our media manager is Rebecca Strom. Additional editing by Martin Huergo. Until next time, I'm your host, Dan Senor.