The Ratio of Civilian to Combatant Deaths in Gaza is Probably ≈3:1 or Worse.

A Demographics-Based Estimate

Introduction

This post attempts to estimate the ratio of civilians to combatants Israel has killed in Gaza. I aim to come up with a lower-bound estimate of the ratio. If my approach is successful, we will be able to say that the ratio is at least as large—that is to say, at least as bad—as I estimate it to be.

The basis for my piece is a conservative demographic analysis of who has been killed in the current war. My analysis is based on killed Gazans fully identified by the Gaza Health Ministry (GHM), which has published detailed data containing the names, ages, and genders for all 24,686 fully identified killed Gazans. (The analysis excludes about 10,000 killed Gazans who have not been fully identified, and about 10,000 more Gazans who are reportedly buried under the rubble.)

I establish six demographic categories for killed Gazans: Female Children and Teenagers (0-18), non-Elderly Women (19-59), Elderly Women (60 and older), Male Children and Teenagers (0-18), non-Elderly Men (19-59), and Elderly Men (60 and up). I calculate what percentage of the overall killed Gazans each group constitutes, and then estimate the proportions of each group who are combatants vs civilians. This allows me to estimate the overall proportions of civilians and combatants among Gazans killed in this war.

I proceed by consulting demographic data from the two major previous wars in Gaza—Operation Cast Lead (2008-2009) and Operation Protective Edge (2014)—to estimate what percentages of the killed persons in the six aforementioned categories were civilians vs combatants. Then I look at what percentages of the total, fully identified Gazans killed in this war belong to each of the six categories. By the end, I arrive at a calculation that the civilians to combatants ratio is at least about 3:1, and quite possibly worse to much worse. This would mean that Israel’s civilians to combatants ratio is perversely high,¹ as bad as or worse than the civilians to combatants ratio in the Battle of Aleppo carried out by Bashar al-Assad and Vladimir Putin.

Two Core Assumptions

My piece stands or falls on the truth or falsity of two assumptions. While of course one or both of them might turn out to be false, both of them are epistemically justified, and probably true.

My first assumption is that the demographic data of fully identified killed Gazans are either representative of the demographics for killed Gazans as a whole; or are unrepresentative in a way that actually underestimates the percentage of civilians and civilian-heavy groups, such as women, elderly, and children.

My second assumption is that, in the current war, the percentages of civilians among the six demographic groups of killed Gazans I define in this piece—Female Children and Teenagers, non-Elderly Women, Elderly Women, Male Children and Teenagers, non-Elderly Men, and Elderly Men—are about the same or higher in the current war as compared to Cast Lead and Protective Edge.

To avoid disrupting the flow of this piece, I will confine to an Appendix (click here to jump to it) my arguments as to why the two aforementioned assumptions are most likely true, and why the Gaza Health Ministry is a reliable source, at least insofar as fully identified deaths are concerned. The serious reader will want to consult this Appendix to evaluate my assumptions.

Let us now turn to the core issue of the post: an attempt to infer the IDF’s civilians-to-combatants ratio in Gaza from the demographics of the identified dead in Gaza. This will be, to reiterate, a conservative or low-ball estimate, insofar as identified deaths (which contain a substantially lower percentage of civilians) are used as a representative sample, rather than overall reported deaths or unidentified deaths (both of which contain much higher percentages of women and children compared to fully identified deaths).

The Demographics of Identified Killed Palestinians

The Gaza Health Ministry (“GHM”) recently published data on the demographic distributions of fully-identified killed Gazans. The GHM broke down the distribution of the 24,686 killed Gazans into four non-overlapping categories:

The GHM’s four categories (“Children,” “Women,” “Elderly,” and Men) are not clearly defined. However, the GHM has also published lists of all fully identified killed Gazans, which contain gender and age breakdowns. This will allow us to break the demographics down into more clearly defined and useful categories. Below, for example, is a screenshot from the GHM’s database for all killed children

One thing to notice—which I highlight in the above image— is that in the GHM’s taxonomy 18-year-old (but not 19-year-old) teenagers are counted as “children.” This definition is misleading insofar as it departs from the common understanding of “children.” It also misleads because the term “child” evokes a non-combatant, yet an 18-year old man is quite capable of fighting in war. For our purposes, therefore, we will rename the GHM’s “children” categories “Children and Teenagers.” Our category excludes nineteen year olds, so as to conform to the GHM’s “child” category.

It should also be noted that the GHM’s definition of elderly comprises all persons 60 and older. We will also follow this definition for our elderly categories.

The Six Demographic Categories

By examining all the data published by the GHM I have been able construct six, non-overlapping categories of fully identified deceased Gazans killed in the current war, according to the following demographics.

Male Children and Teenagers (aged 0-18): account for 17.4% of deaths in the current war.

Non-Elderly Men (aged 19-59):  account for 40.5% of deaths in the current war.

Elderly Men (aged 60 and over): account for 4.9% of deaths in the current war.

Female Children and Teenagers (aged 0-18): accounts for 14.2% of deaths in the current war.

Non-Elderly Women (aged 19-59): account for 20.1% of deaths in the current war.

Elderly Women (aged 60 and over): account for 2.9% of deaths in the current war.

Now that we have defined and calculated the six aforementioned demographic categories, we will calculate what percentages of these categories were civilians versus combatants in Israel’s recent major wars in Gaza: Cast Lead (2008-2009) and Protective Edge (2014). By assuming that the percentages of civilians and combatants across our six demographic groups are the same in the current war as they were in Cast Lead and Protective Edge, we can estimate the civilians to combatants ratio for the current war.

Comparison with Past Gaza Wars

The extraordinary database of the leading Israeli human rights organization B'Tselem contains highly granular demographic information on the Palestinian casualties of all past Israeli wars in Gaza: one can easily see how many Palestinians were killed, whether they were civilians or combatants, and what their age and sex was.

Using the B’Tselem database, I will estimate the percentages of civilians and combatants killed in the current war in Gaza. I proceed by assuming that the percentages of killed Palestinians in the current war who were civilians and combatants—across the six aforementioned demographic groups, Male Children and Teenagers, Non-Elderly Men, etc—are equal to the corresponding percentages of civilians and combatants killed across the six demographic groups in Cast Lead and Protective Edge,

Defining “Civilians” versus “Combatants”

In general, I defer to B’Tselem’s definition of civilians vs combatants to determine the percentages of various categories of killed Palestinians who were civilians vs combatants. One exception relates to police officers killed in Operation Cast Lead, which B’Tselem counts neither as civilians or combatants. Consistent with the Law of Armed Conflict (LOAC), I assume all police officers killed in Operation Cast Lead are civilians until evidence to the contrary is produced.²

However, for 137/248, or 55% of the police killed in Cast Lead evidence can be produced that they were combatants for LOAC purposes. There is no such evidence implicating the other 45% of police. All of these should be considered civilians until proven otherwise.

Consistent with LOAC, I also count persons of unknown affiliation as non-combatants and therefore as civilians. I also exclude from my analysis Palestinians killed in the West Bank and Israel during Cast Lead and Protective Edge; my data are confined to hostilities in Gaza. Finally, only Palestinians classified by B’Tselem as having been killed by Israeli forces are counted.

Data and Calculations

Finally, I am ready to begin calculating the civilians to combatant ratios. I will proceed as follows. First, I will analyse Gazans killed in Operation Cast Lead and Operation Protective Edge, in terms of the six demographic groups I described above. For each of my six groups, I will calculate what percentages of Gazans killed in Cast Lead and Protective Edge were civilians vs combatants.

Then I will apply these percentages to the GHM data of identified Gazan fatalities in the current war. Because we know the exact proportion of each of my six groups among the total number of identified Gazan dead, defining how many (or what percentages) of each group are civilians or combatants³ will allow us to estimate the percentages of civilians and combatants among killed Gazans in the current war.

Let us start with the easiest part. Three of my six demographic groups are female, and for all of these, percentages of civilians and combatants can be easily calculated. Since a remarkable 100% of females—including Female Children and Teenagers, non-Elderly Women, and Elderly Women—killed in Protective Edge and Cast Lead were civilians, we will assume that virtually all women in these three categories are also civilians in the current war. That is to say, before we begin to analyse the three male groups, we already assume that 37.2% of Palestinians killed in the current war (those who are female) were civilians.

That leaves three groups of killed Palestinians: Male Children and Teenagers (17.4% of total killed in the current war), non-Elderly Men (40.5% of the total killed in the current war), and Elderly Men (4.9% of total killed in the current war). To estimate the percentages of persons in these categories who are civilians and combatants in the current war, we will calculate the percentages of killed Gazans belonging to these three categories who were civilians vs combatants in Cast Lead and Protective Edge.

Male Children and Teenagers

In Cast Lead, 234 out of 272 of killed Male Children and Teenagers, or 86.0%, were civilians. 38⁴, or 14.0%, were combatants:

In Protective Edge, of killed Male Children and Teenagers, 371 out of 404, or 91.8%, were civilians. 33, or 8.2%, were combatants:

Non-Elderly Men

In Cast Lead, 426 out of 875 killed non-Elderly Men, or 48.7%, were civilians. 449, or 51.3%, were combatants:

In Protective Edge, 491 out of 1226 non-Elderly Men Killed, or 40.0%, were civilians. 735/1226, or 60.0%, were combatants:

Elderly Men

In Cast Lead, all 34 Elderly Men killed were civilians, so 100% were Civilians and 0% were combatants:

In Protective Edge, 63/64 Elderly Men Killed, or 98.4%, were civilians. 1/64 Elderly Men killed was a combatant, for a total of 1.6% combatants:

Assessing The Percentages of Combatants and Civilians

Now that we know what percentages of our six demographic categories were civilians and combatants in Cast Lead and Protective Edge, we can formally estimate the civilians to combatants ratio for the current war. We will proceed by multiplying each of our six categories of fully identified killed Gazans—defined in terms of the respective percentages of the Gazan dead they constitute—by the percentages of them who were civilians and combatants in Cast Lead and Protective Edge. Then we will add up the percentage points and divide the total percentage points for civilians by the total percentage points for combatants, thereby calculating civilians to combatants ratios.

Remember that 100% of women—inclusive of all three of our female categories: Female Children and Teenagers; non-Elderly Women; and Elderly Women—were civilians in Cast Lead and Protective Edge. The three female categories, or the total women of all ages among all deaths, add up to 37.2 percentage points. Thus, we will add 37.2 percentage points to reach our final estimate for percentage of civilians among killed Gazans in the current war.

Assuming Cast Lead Demographics

Civilians Killed:

Male Children and Teenagers=17.4% of total deceased, multiplied by 86.0% (the rate at which killed Males and Teenagers were civilians in Cast Lead)=15.0% of total deaths in 2024 Gaza war.

Non-Elderly Men (aged 19-59)=40.5% of total deceased, multiplied by 48.7% (the rate at which killed non-Elderly Men were civilians in Cast Lead)=19.7% of total deaths in 2024 Gaza war.

Elderly Men (60 and up)=4.9% multiplied by 100% (the rate at which killed Elderly Men were civilians in Cast Lead)= 4.9% of total deaths in 2024 Gaza War.

Combatants Killed:

Male Children and Teenagers =17.4% of total deceased, multiplied by 14.0% (the rate at which killed Male Children and Teenagers were combatants in Cast Lead)=2.4% of total deaths in 2024 Gaza war.

Non-Elderly Men (aged 19-59)=40.5% of total deceased, multiplied by 51.3% (the rate at which killed non-Elderly Men were combatants in Cast Lead)=20.8% of total deaths in 2024 Gaza war.

Elderly Men (60 and up)=4.9% multiplied by 0% (the rate at which deceased Elderly Men were combatants in Cast Lead)= 0.0% of total deaths in the 2024 Gaza war.

Final Calculation:

Civilians= 37.2 (All Females) plus 15.0 (Male Children and Teenagers) plus 19.7 (non-Elderly Men) plus 4.9 (Elderly Men)==77% of total deaths.

Combatants=0 (All Females) plus 2.4 (Male Children and Teenagers) plus 20.8 (non-Elderly Men) plus 0 (Elderly Men) =23% of total deaths.

77÷23=

3.35, the civilians to combatants ratio for the current war (assuming Cast-Lead level concentrations of civilians and combatants in my six demographic groups). This can be rounded down, to the nearest whole number as 3:1.

Assuming Protective Edge Demographics

Civilians:

Male Children and Teenagers=17.4% of total deceased in current war, multiplied by 91.8% (the rate at which killed Male Children and Teenagers were civilians in Protective Edge)=16.0% of total deaths in 2024 Gaza war.

Non-Elderly Men=40.5% of total deceased in current war, multiplied by 40.0% (the rate at which killed non-Elderly Men were civilians in Protective Edge)=16.2% of total deaths in 2024 Gaza war.

Elderly Men =4.9% of total deceased in current war, multiplied by 98.4% (the rate at which killed Elderly Men were civilians in Protective Edge)= 4.8%

Civilians= 37.2 (Females)+16.0 (Male Children and Teenagers)+ 16.2 (non-Elderly Men)+4.8 (Elderly Men) =74% of total in current war.

Combatants:

Male Children and Teenagers=17.4% of total deceased in current war, multiplied by 8.2% (the rate at which killed Male Children and Teenagers were Combatants in Protective Edge)=1.4% of total deaths in 2024 Gaza war.

Non-Elderly Men=40.5% of total deceased in current war, multiplied by 60.0% (the rate at which killed non-Elderly Men were combatants in Protective Edge)=24.3% of total deaths in 2024 Gaza war.

Elderly Men =4.9% of total deceased in current war, multiplied by 1.6% (the rate at which killed Elderly Men were combatants in Protective Edge)=0.1% of total deaths in the 2024 Gaza war.

Combatants=0.0 (All Females)+1.4 (Male Children and Teenagers) +24.5 (non-Elderly Men) plus 0.1 (Elderly Men)= 26% of total in current war.

74÷26=

2.85:1, the civilians to combatants ratio for the current war (assuming Protective-Edge level concentrations of civilians and combatants in my six demographic groups). This can be rounded up, to the nearest whole number, as 3:1.

Conclusion

Based on the two ratios I calculated above, I estimate that the IDF’s civilians-to-combatants ratio in Gaza is at least ≈3:1. As noted above, this is a lower-bound estimate. The ratio may well be larger or even much larger.

3:1 would be a hideous ratio. Contrary to the popular myth that the typical civilians to combatants ratio is 9:1 in war—promoted by John Spencer, a leading propagandist for Israel’s war effort⁵—a 3:1 ratio would be extremely high. Indeed, 3:1 was the approximate ratio compiled by Russian and Syrian armed forces in the siege of Aleppo, widely recognized as one of the great atrocities of the 21st century.

The accuracy of this estimate hinges on the two empirical assumptions I mentioned at the outset of the piece. Both of these core assumptions will be defended in the Appendix that follows this piece.

Ultimately, we can’t know for sure how many civilians have been killed in Gaza, nor what the IDF’s civilians to combatants ratio is. But while the fog of an ongoing war prevents us from calculating these figures with academic precision, we must use the information we have at hand to make empirical and moral assessments as to the conduct of the IDF in its so-called “war” Gaza. If we wait until the epistemic landscape settles to make such assessments—that is, wait until after the war—it will be far too late.

Appendix

As noted above, the credibility of my estimates for Israel’s civilians to combatants killed ratio in Gaza rests on two assumptions. The first assumption is that that the proportion of civilians among the killed Gazans fully identified by the Gaza Health Ministry—as suggested by the relative share of civilian-heavy demographics, such as women and children, among the identified dead—is either roughly equal to, or else less than the actual proportion of civilians among the total dead Gazans.

Secondly, I assume that the proportion of civilians among my six demographic groups— Female Children and Teenagers, non-Elderly Women, Elderly Women, Male Children and Teenagers, non-Elderly Men, and Elderly Women—is about the same or higher in the current war as compared to Cast Lead and Protective Edge.

The first assumption turns on the credibility of the Gaza Health Ministry (“GHM”). And indeed, the GHM boasts a strong track record of general and demographic accuracy. Its estimates of death for Operation Cast Lead (2008-2009), Protective Edge (2014), as well as the 2021 clashes between Israel and Hamas were extremely accurate (each well within 5% of post-war estimates conducted by the United Nations). As for the current war in Gaza, both American and Israeli Intelligence, as well as the United Nations, have confidence in the GHM’s estimates of fully identified deaths.

It should be noted that the data collection procedures of the GHM have changed during the current war, because of the destruction of much of their infrastructure and capacity to individually autopsy and verify specific deaths. The GHM now has two categories of reported deaths.

The first category is identified deaths, with full identifying information, including names, addresses, ages, and (Israeli-issued) ID numbers. The GHM methodology for recording identified deaths is the same as it has been for past wars, in which the GHM was proven essentially accurate.

The second category is unidentified deaths. GHM appears to draw these largely from media reports on airstrikes.

There are major demographic gaps between identified and non-identified deaths. Most striking in the data is the big gap between the percentage of dead women and children among identified dead—where women and children are a clear but modest majority—and the unidentified dead, where women and children are over 90% (!) of the total. This gap has provoked allegations from defenders of Israel that the GHM is fabricating its death-count for the unidentified category.

While no evidence has been provided of GHM fabrication of fatalities—and indeed, there evidence to the contrary⁶—it is more than plausible that these data suffer from sampling bias, and under-count killed Hamas militants.⁷ However, even if we assume that many thousands of unidentified dead Hamas militants have been missed in the GHM “unidentified deaths” data, unidentified deaths are at present so overwhelmingly (over 90%) comprised of women and children that including thousands of hypothetical killed Hamas militants in these data would still result in a more civilian-heavy distribution for unidentified deaths compared to identified deaths.   

At the end of the day, the GHM has given us two sets of data: fully identified deaths; and non-identified deaths. The latter has a much higher concentration of civilians than does the former. Hence the First Assumption—that the percentage of killed Gazans who are civilians is about as high or higher than that suggested by the demographic breakdown of fully identified deaths—is consistent with the data we have at hand, and probably true.

Now I will address my second assumption, the assumption that the percentages of civilians among the killed Gazans in my six demographic groups are about as high or higher in this war than it was in Cast Lead and Protective Edge. This assumption is supported by the nature of the current war in Gaza. Israel is waging the current war much more indiscriminately than it waged previous wars in Gaza—adopting (even) looser rules of engagement, arbitrarily killing Gazans in free-fire zones, and showing much greater tolerance for collateral damage. Thus, one would expect that the percentage of the dead who are civilians, across various demographics groups, would be at least as high as it was in the previous wars, and likely substantially higher.⁸

Hence, the proportion of various demographics of killed Gazans who are civilians—or more specifically, the proportion of killed Gazans in my six demographic groups who are civilians—is probably about the same or higher than it was in Cast Lead and Protective Edge.

1. For a rebuttal of the popular myth that the typical civilians to combatants ratio is 8:1 or 9:1, see footnote 5.

2. I compiled this evidence through the entries of “martyrs” on the Al-Qassam website. See Martyrs of al-Qassam: Izz al-Din al-Qassam Brigades (alqassam.ps).

These entries contain 137 police officers killed in Cast Lead, along with detailed biographical information describing their experience in Al-Qassam.
In a minority of cases, it appears likely that a “martyr” was not actually a member of Al-Qassam at the time he was killed, but rather had left A-Q to become a police officer. See, for example, the entry for forty-two year old “martyr” Ibrahim Yousef Nofal.  إبراهيم يوسف نوفل : كتائب الشهيد عز الدين القسام (alqassam.ps)  However, I include all these persons—as well as the minority of entries whose authenticity is open to serious doubt—as combatants for statistical purposes here.

This is because—consistent with my attempt to establish a bare-minimum estimate for the civilians to combatants ratio—I want to err on the side of over-counting militants and correspondingly under-estimating the civilians to combatant ratio.

3. Again, these percentages are defined in terms of their corresponding percentages in Cast Lead and Protective Edge.

4. Including two 18-year old police officers.

5. The 9:1 myth—which has been repeated by bureaucrats at the EU and UN—has been discredited in the academic literature. The below screen shot is taken from a paper by Adam Roberts, which comprehensively discredits the myth. Even wars involving heavy urban warfare are almost always nowhere close to 9:1 or 8:1.

It is worth emphasizing that all the examples of post-WWII wars Roberts lists with likely 9:1 or worse ratios involved literal, established genocides (Cambodia, Rwanda, Second Congo War, Darfur), while the other one is a disputed genocide (Northern Uganda).

6. The fact that several thousand deaths previously unidentified deaths reported by the GHM have been subsequently identified is evidence that unidentified deaths are not being faked. Although it is quite plausible that the demographics of unidentified deaths are distorted by sampling bias, on which see footnote 7.

7. The existence of sampling bias—even severe sampling bias—is not proof of bad faith on the part of GHM. As Avi Bitterman has suggested, it is plausible that media reports of Palestinian casualties tend to disproportionately cover airstrikes where women and children were the primary victims. And hence relying on media reports—the GHM methodology for counting unidentified deaths—would lead one to over-estimate the percentage of women and children among the overall dead from the war; though would not lead one to cite fake deaths.

8. Obviously, the proportion of killed females who are civilians couldn’t increase relative to Cast Lead and Protective Edge, since 100% of them were civilians in these wars.