The Battle of Bakhmut: Postmortem

On May 20th, PMC Wagner forced Ukrainian troops out of their last remaining position within the city limits of Bakhmut, consequentially bringing about the nominal end of the largest battle of the 21st century (so far). Bakhmut has been the most substantial locus of military operations in Ukraine for most of the past nine months. Combat there took on a frustrating tempo, with progress often measured in single city blocks. This was a battle that was extremely violent and bloody, but at times agonizingly slow and seemingly indecisive. After countless updates in which nothing of note seemed to have happened, many people were surely beginning to roll their eyes at the very mention of Bakhmut. Consequentially, the abrupt capture of the city by Wagner in May (rather predictably, the final 25% of the city fell very quickly relative to the rest) seemed a bit surreal. To many it likely seemed that Bakhmut would never end - and then, suddenly, it did. 

Bakhmut, like most high-intensity urban battles, exemplifies the apocalyptic potential of modern combat. Intense bombardment reduced large portions of the city to rubble, lending the impression that Wagner and the AFU were not so much fighting over the city as its carcass. 

The slow pace and extreme destruction has made this battle a rather difficult one to parse out. It all seems so senseless - even within the unique paradigm of war-making. In the absence of an obvious operational logic, observers on both sides have been eager to construct theories of how the battle was actually a brilliant example of four-dimensional chess. In particular, you can easily find arguments from both pro-Ukrainian and pro-Russian commentators claiming that Bakhmut was used as a trap to draw in the other side’s manpower and material for destruction, while buying time to accumulate fighting power.

Pro-Ukrainian sources are adamant that a huge amount of Russian combat power was destroyed in Bakhmut, while the AFU received western armor and training to build out a mechanized package to go back on the offensive. Pro-Russian writers similarly seem convinced that the AFU burned a huge amount of manpower, while the Russian army preserved its strength by letting Wagner do most of the fighting. 

Clearly, they cannot both be correct. 

In this article, I would like to take a holistic survey of the Battle of Bakhmut and adjudicate the evidence. Which army was really destroyed in this “strategically insignificant” city? Which army was being profligately wasteful of its manpower? And most importantly - why did this middling city become the site of the largest battle of the century? Homicide was committed, but nobody can agree on who murdered whom. So, let us conduct an autopsy. 

The Road to the Death Pit

The Battle of Bakhmut lasted for so long that it can be easy to forget how the front ended up there, and how Bakhmut fits into the operations in the summer of 2022. Russian operations in the summer were focused on the reduction of the Ukrainian salient around Lysychansk and Severodonetsk, and came to a climactic head when Russian forces broke open the heavily defended Ukrainian stronghold of Popasna, encircled a pocket of Ukrainian forces around Zolote, and approached the Bakhmut-Lysychansk highway. The actual fall of the Lysychansk-Severodonetsk urban agglomeration came relatively quickly, with Russian forces threatening to encircle the entire bag and forcing a Ukrainian withdrawal.

Just for reference, here is what the frontline in the central Donbas looked like on May 1, 2022, courtesy of MilitaryLand:

In this context, Bakhmut already threatened to become a major battleground. It lay at a literal crossroad directly in the center of the Ukrainian salient. As Ukrainian positions in Lysychansk, Popasna, and Svitlodarsk were broken open, the axes of Russian advance converged on Bakhmut. 

Ukrainian forces greatly needed to stabilize the front, and establish a stable blocking position, and there was really nowhere else to do this except at Bakhmut. Between Lysychansk and Bakhmut there are no sufficiently robust urban areas to anchor the defense, and there was absolutely no question of failing to adequately defend Bakhmut, for a few reasons that we can enumerate:

1. Bakhmut is in the central position in this sector of front, and its loss would both threaten Siversk with envelopment and allow Russian forces to bypass the well-fortified and strongly held defenses at Toretsk. 

2. The Russian strategic objective at Slavyansk-Kramatorsk cannot be successfully defended if the Russian army controls both the heights to the east (in the Bakhmut area) and Izyum. 

Bakhmut itself, meanwhile, was a defensible urban area with dominating heights in its rear, multiple routes of supply, good linkages to other sectors of front, and a peripheral belt of smaller urban areas protecting its flanks. 

This set up a rather obvious operational decision for Ukrainian forces. The choice was, all things considered, to either commit reserves to stabilize the front at Bakhmut (a strong and operational vital defensive anchor) or risk letting Russia bypass and sweep away an entire belt of defenses in places like Siversk and Toretsk. Asked to choose between a reasonably good option and an extremely bad option, there was no great controversy in the deciding. 

Falling back from the loss of their eastern defensive belt, Ukraine needed to stabilize the front somewhere, and the only suitable place was Bakhmut - so this is where Ukrainian reserves were sent in force, and the AFU chose to fight. Operational logic, indifferent to those things that normally recommend cities to us as “important”, decreed that the Styx should flow through Bakhmut.

Russia came to meet this challenge - bringing as their spearhead a mercenary group, staffed by convicts, wielding shovels, run by a bald caterer. What could go wrong? 

Operational Progression

Because the general impression of Bakhmut is characterized by urban combat, it can be easy to forget that most of the battle took place outside the city itself, in the exurbs and fields around the urban center. The approach to Bakhmut is cluttered with a ring of smaller villages (places like Klynove, Pokrovs’ke, and Zaitseve) from which the AFU was able to fight a tenacious defense with the support of artillery in the city itself. 

While Russian forces nominally reached the approach to Bakhmut late in June (even before Lysychansk was captured) and the city came under the extremes of shelling range, they did not immediately begin a concerted push to reach it. On August 1, the first assaults on the outer belt of villages began, and the Russian Ministry of Defense stated in its briefings that “battles for Bakhmut” had begun. This date is the most logical starting date for historiographical purposes, so we may firmly say that the Battle of Bakhmut was fought from August 1, 2022 to May 20, 2023 - a total of 293 days.

The first two months of the battle saw the Russian capture of most of the settlements east of the T0513 highway south of the city and the T1302 highway to the north, stripping Bakhmut and Soledar of most of their eastern buffer zones and pushing the line of contact right up to the edge of the urban areas proper.

Phase 1: The Outer Belt

At this point, the frontlines largely froze up for the remainder of the year, before Wagner set the stage for further advances with the capture of the small village of Yakovlivka, to the north of Soledar. This success can be construed as the first domino in a chain of events which led to Ukrainian defeat in Bakhmut. 

Soledar itself serves a unique and critical role in the operational geography of Bakhmut. Laid out in a relatively long and thin strip, Soledar and its suburbs form a continuous urban shield stretching from the T0513 highway (which runs north to Siversk) all the way to the T0504 road (running east to Popasna). This makes Soledar a natural satellite stronghold which defends Bakhmut across nearly ninety degrees of approach. Soledar is also liberally gifted with industrial build, including the salt mine for which it is named, which makes it a relatively friendly place to wage a static defense, full of deep places and strong walls. 

Wagner’s capture of Yakovlivka on December 16, however, marked the first sign that the defense of Soledar was in trouble. Yakovlikva sits on an elevated position to the northeast of Soledar, and its capture gave Wagner a powerful position atop Soledar’s flank. The Ukrainians recognized this, and Soledar was powerfully reinforced in response to the loss of Yakovlivka and the anticipated oncoming assault. The capture of Bakhmutske on December 27 (a suburb of Soledar directly on its southern approach) set the stage for a successful assault. 

The attack on Soledar ended up being relatively fast and extremely violent, characterized by intense levels of Russian artillery support. The assault began almost immediately after the loss of Bakhmutske on December 27th, and by January 10th Ukraine’s cohesive defense had been shattered. Ukrainian leadership, of course, denied losing the town and wove a story about glorious counterattacks, but even the Institute for the Study of War (a propaganda arm of the US State Department) later admitted that Russia had captured Soledar by January 11th. 

The loss of Soledar, in combination with the early January capture of Klischiivka to the south, put Wagner in a position to begin a partial envelopment of Bakhmut.

Phase 2: Clearing the Flanks

It was at this point that the discussion shifted towards a potential Russian encirclement of Bakhmut. To be sure, the Russian wings did expand rapidly around the city, placing it in a firebag, but there was never a concerted effort to take the city into a proper encirclement. The Russian advance subsided on the approach to Ivanivske in the south, and over the vital M03 highway in the north. 

A genuine encirclement was probably never in the cards, mainly because of the complication of Chasiv Yar - a strongly held rear area stronghold. To fully encircle Bakhmut, Russian forces would have been forced to choose between two difficult options: either blockade the road from Chasiv Yar to Bakhmut, or flare the envelopment wide enough to take Chasiv Yar into the pocket as well. Either option would have greatly complicated the operation, and so Bakhmut was never genuinely encircled. 

What the Russians did succeed in doing, however, was establish dominant position on the flanks which accrued three significant advantages. First, they were able to direct fire on Bakhmut’s remaining supply lines. Secondly, they were able to pummel Bakhmut itself with intense artillery fire from a variety of axes. Third - and perhaps most importantly - they were able to assault the Bakhmut urban center itself from three different directions. This, in the end, greatly hastened the fall of the city. By April, it was clear that the focus had shifted from expanding the envelopment on the flanks to assaulting Bakhmut itself, and it was reported that Russian regular units had taken custody of the flanks so that Wagner could clear the city.

Phase 3: Flaring the Wings

Fighting throughout April and early May at last shifted to the struggle in the urban center. AFU units in the city ultimately proved incapable of stopping Wagner’s advance, largely due to tight Russian fires coordination and the cramped confines of the Ukrainian defense - with Wagner advancing into the city from three axes, the firing grids for Russian artillery became very narrow, and the AFU’s static defense - while bravely contested - was slowly ground down. 

By early May, it was clear that the city would fall soon, with the AFU desperately holding on to the western edge of the city. Attention soon shifted, however, to a Ukrainian counterattack on the flanks. 

This became a rather classic instance of events on the ground being outrun by the narrative. There had been rumors of an impending Ukrainian counterattack circulating for quite some time, advanced by both Ukrainian and Russian sources. Ukrainian channels were predicated on the idea that General Oleksandr Syrskyi (commander of AFU ground forces) had hatched a scheme to draw the Russians into Bakhmut before launching a counterattack on the wings. This idea was seemingly corroborated frantic warnings from Wagner head Yevgeny Prigozhin that the Ukrainians had massed enormous forces in the rear areas behind Bakhmut which would be unleashed to counter-encircle the city. 

In any case, the spring months came and went without any astonishing AFU counterattack, and all manner of material shortages and weather delays were blamed. Then, on May 15, all hell seemed to be breaking loose. The AFU finally attacked, and Prigozhin screeched that the situation on the flanks was approaching the worst-case scenario. 

In fact, what happened was rather anticlimactic. The AFU did bring a hefty grouping of units to play, including several of their best and most veteran formations. These included units from:

• The 56th Brigade

• The 57th Mechanized Brigade

• The 67th Mechanized Brigade

• The 92nd Mechanized Brigade

• The 3rd Assault Brigade (Azov)

• The 80th Air Assault Brigade

• The 5th Assault Brigade

This sizeable strike package attacked a handful of mediocre Russian Motor Rifle brigades, achieved a bit of initial success, and culminated with heavy losses. Despite Prigozhin’s assertion that the Russian regulars abandoned their posts and left the Russian wings undefended, we later learned that these forces - including mobilized motor rifle units - doggedly defended their positions and only withdrew under orders from above. These withdrawals (distances of a few hundred meters at most) brought the Russian defensive line to strongly held positions along a series of canals and reservoirs, which the AFU was unable to push through. 

Now, this is not to say that Russia did not suffer losses defending against a tenacious Ukrainian attack. The 4th Motor Rifle Brigade, which was largely responsible for the successful defense outside Klishchiivka, was badly chewed up, its commander was killed, and it had to be promptly rotated out. However, the offensive potential of the Ukrainian assault package was exhausted, and there have been no follow on attempts in the past two weeks.

The Final Act: Ukraine’s Counterattack

In the end, the vaunted Syrskyi plan looked rather lame. The counterattack did successfully unblock a few key roads out of Bakhmut, but it did nothing to prevent Wagner from finalizing the capture of the city, it burned through the combat power of several premier brigades, and on May 20th the last Ukrainian positions in the city were liquidated. 

So. This was a strange battle. An agonizingly slow creep around the flanks of the city, a materializing threat of encirclement, and a sudden concentration of Wagner’s combat energy in the city itself - all taking place under threats of an enormous counteroffensive by the AFU, which turned out to be ineffective and ephemeral. 

It’s not obvious, then, how this battle suited the operational logic of either army, nor that anyone would come away fully satisfied. Ukraine obviously lost the battle in nominal terms, but the Russian advance seemed so slow and Bakhmut so strategically random (at least superficially) that Wagner’s success can be portrayed as a pyrrhic victory. To fully adjudicate the Battle of Bakhmut, we need to contemplate relative losses and expenditure of combat power. 

The Butcher’s Bill

Estimating combat losses in Ukraine is a difficult task, largely because “official” casualty estimates are often patently absurd. This leaves us with a need to fumble for reasonable figures using proxies and ancillary information. One such important source of knowledge is deployment data - we can get a general sense of the burn rate by the scale and frequency of unit allocation. In this particular case, however, we find that unit deployments are somewhat difficult to work with. Let’s parse through this. 

First and foremost, we need to grapple with the incontrovertible fact that a huge share of the Ukrainian military was deployed at Bakhmut at one point or another. The Telegram Channel Grey Zone compiled a list of all the Ukrainian units that were positively identified (usually by social media posts or AFU updates) as being deployed in Bakhmut throughout the nine month battle (that is, they were not there all at once):

This is an absolutely enormous commitment (37 brigades, 2 regiments, and 18 separate battalions (plus irregular formations like the Georgian Legion) which indicates obviously severe losses (for what it’s worth, the pro-Ukrainian MilitaryLand Deployment Map admits a similarly titanic Ukrainian deployment in Bakhmut). However, this does not really get us close to accurately assessing losses, largely because Ukraine’s Order of Battle (ORBAT) is a bit confused. Ukraine frequently parcels out units below the brigade level (for example, their artillery brigades never deploy as such) and they have a bad habit of unit cannibalization. 

Doing some extremely rough back of the envelop math, minimal scratching off of just the 37 brigades could easily have pushed Ukraine past 25,000 casualties, but there are all manner of shaky assumptions here. First, this assumes that Ukraine withdraws its brigades when they reach combat ineffective loss levels (15% would be a placeholder number here), which isn’t necessarily true - there is precedent for the AFU leaving troops in place to die, especially from lower quality units like Territorial Defense. In fact, an Australian volunteer (interview linked later on) claimed that the 24th mechanized brigade suffered 80% casualties in Bakhmut, so it’s possible that a great many of these brigades were chewed up beyond task ineffectiveness levels (that is, they were not correctly rotated out) but were instead destroyed entirely. A recent article in the New Yorker, for example, interviewed survivors of a battalion that was almost entirely wiped out. In another instance, a retired Marine Colonel said that units at the frontline routinely suffer 70% casualties. 

We can say a few things for certain. First, that Ukraine had an extremely high burn rate which forced it to commit nearly a third of its total ORBAT. Secondly, we know that at least some of these formations were left at the front until they were destroyed. Finally, we can definitively say that pro-Ukrainian accounts are incorrect (or maybe lying) when they say that the defense at Bakhmut was conducted to buy time for Ukraine to build up strength in the rear. We know this first and foremost because Bakhmut insatiably sucked in additional units, and secondly because this burn included a large number of Ukraine’s premier and veteran forces, including fully a dozen assault, airborne, and armored brigades. 

There’s another problem with the ORBAT approach to casualties, however, and this concerns Wagner. You see, one of our objectives here is to try to get a sense of the comparative rates of loss, and ORBAT simply isn’t a good way to do this in the particular case of Bakhmut. This is because the battle was mostly fought from the Russian side by the Wagner Group, which is a huge formation with an opaque internal structure. 

Whereas on the Ukrainian side we can enumerate a long list of formations that fought at Bakhmut, on the Russian side we just put the 50,000 strong Wagner Group. Wagner of course has internal sub-formations and rotations, but these are not visible to those of us on the outside, and so we cannot get a sense of Wagner’s internal ORBAT or force commitment. We understand generally that Wagner has a structure of assault detachments (probably a battalion equivalent), platoons, and squads, but we do not have a sense of where these units are deployed in real time or how quickly they are rotated or burned through. Sadly, when Prigozhin went in front of cameras he brought maps without unit dispositions on them, leaving ORBAT nerds squinting in vain trying to extract useful information. So, lacking good insight into Wagner’s deployments, we are unable to make an adequate comparison to the bloated Ukrainian ORBAT in Bakhmut. 

There are other ways that we can get at the casualties, however. The Russian dissident (that is, anti-Putin) organization Mediazona tracks Russian losses by tabulating obituaries, death announcements on social media, and official announcements. For the entire period of the Battle of Bakhmut (August 1 - May 20), they counted 6,184 total deaths among PMC personnel, inmates, and airborne forces (these three categories accounting for most of the Russian force in Bakhmut). 

Meanwhile, Prigozhin claimed that Wagner had suffered 20,000 KIA in Bakhmut while inflicting 50,000 KIA on the Ukrainians. Concerning the first number - the context of this claim was an interview in which he was lambasting the Russian Ministry of Defense (as is his habit), and he has an incentive to overstate Wagner’s losses (since he is trying to play up Wagner’s sacrifice for the Russian people). 

So, here is where we are at with Wagner losses. We have a “floor”, or absolute minimum of a little over 6,000 KIA (these being positively identified by name) with a significant upward margin of error , and something like a ceiling of 20,000. The number that I have been working with is approximately 17,000 total Wagner KIA in the Bakhmut operation (with a min-max range of 14,000 and 20,000, respectively).

However, something we need to consider is the composition of these forces. Among the positively identified KIA, convicts outnumber professional PMC operators by about 2.6 to 1 (that is, Wagner’s dead would be about 73% convicts). According to the Pentagon, however (taken with a large grain of salt), nearly 90% of Wagner’s losses are convicts. Taking a conservative 75/25 split and rounding the numbers to make them pretty, my estimate is that Wagner lost about 13,000 convicts and 4,000 professional operators.Adding in VDV losses and motorized rifle units fighting on the flanks, and total Russian KIA in Bakhmut are likely on the order of 20-22,000.

So, what about Ukrainian losses? The major outstanding question remains: who is on the right end of the loss ratios? 

Ukrainian commentators consistently ask us to believe that Russian losses were far worse due to their use of “human wave” attacks. There are several reasons why this can be dismissed. 

First, we have to acknowledge that after nine months of combat we have not yet seen a single video showing one of these purported human waves (that is, Wagner convicts attacking in a massed formation). Keeping in mind that Ukraine loves to share footage of embarrassing Russian mistakes, that they have no qualms about sharing gory war porn, and that this is a war being fought with thousands of eyes in the sky in the form of reconnaissance drones, it must strike us as curious that not one of these alleged human wave attacks has yet been caught on camera. When videos are shared purporting to show human waves, they invariably show small groups of 6-8 infantry (we call this a squad, not a human wave). 

However, the absence of evidence is not evidence of absence. That being said, the “human wave” narrative has been contradicted on multiple occasions. Just for starters, General Syrskyi himself contradicted the human wave narrative and said that Wagner’s methodology is to push small assault groups forward under intense artillery cover. Witnesses from the front concur. An Australian Army Veteran volunteering in Ukraine gave a very interesting interview in which he downplayed Wagner casualties and instead emphasized that “Ukraine is taking way too many casualties” - he later adds that the 24th brigade suffered 80% casualties in Bakhmut. He also notes that Wagner favors infiltration groups and small units - the veritable opposite of massed human waves.

I found this article from the Wall Street Journal to be nicely emblematic of the human wave issue. It contains the obligatory claim of human wave tactics - “The enemy pays no attention to huge losses of its personnel and continues the active assault. The approaches to our positions are simply littered with the bodies of the adversary’s dead soldiers.” This description, however, comes from the bureaucratic apparatus at the Ministry of Defense. What about people on the ground? A Ukrainian officer at the front says: “So far, the exchange rate of trading our lives for theirs favors the Russians. If this goes on like this, we could run out.” 

Ultimately, it’s difficult to believe that the kill ratio favors Ukraine for the simple reason that the Russians have enjoyed a tremendous advantage in firepower. Ukrainian soldiers speak freely about Russia’s enormous superiority in artillery, and at one point it was suggested that the AFU was outgunned by ten to one. The New Yorker’sinterview subjects claimed that their battalion’s mortar section had a ration of a mere five shells per day!

The enormous Russian advantage in artillery and standoff weaponry suggests the a-priori assumption that the AFU would be taking horrific casualties, and indeed that’s what we hear from myriad sources at the front. Then, of course, there was the shocking February claim by a former US Marine in Bakhmut that the life expectancy at the front line was a mere four hours. 

All of this is really ancillary to the larger point. The enormous inventory of AFU units that were churned through Bakhmut included something on the order of 160,000 total personnel. Taking loss rates of between 25 and 30% (roughly on par with Wagner’s burn rate), it’s clear that Ukraine’s losses were extreme. I believe total irretrievable losses for Ukraine in Bakhmut were approximately 45,000, with some +/- 7,000 margin of error. 

So, my current working estimates for losses in the Battle of Bakhmut are some 45,000 for Ukraine, 17,0000 for Wagner, and 5,000 for other Russian forces. 

But perhaps even this misses the point. 

Ukraine was losing its army, Russia was losing its prison population. 

Adjudicating the Battle of Bakhmut is relatively easy when one looks at what units were brought to the table. Bakhmut burned through an enormous portion of the AFU’s inventory, including many of its veteran assault brigades, while virtually none of Russia’s conventional forces were damaged (with the notable exception of the Motor Rifle brigades that defeated the Ukrainian counterattack). Even the Pentagon has admitted that the vast majority of Russian casualties in Ukraine were convicts. 

Now, this is all rather cynical - nobody can deny it. But from the unsentimental calculus of strategic logic, Russia churned through its single most disposable military asset, leaving its regular ORBAT not only completely intact, but actually larger than it was last year. 

Meanwhile, Ukraine was left with virtually no indigenous offensive power - the only way it can conduct offensive operations is with a mechanized package built from scratch by NATO. For all Ukraine’s bluster, the force commitment at Bakhmut left it unable to undertake any proactive operations all through the winter and spring, its multi-brigade counterattack at Bakhmut lamely fizzled out, and it left its supporters grasping at straws about an immanent counteroffensive to encircle Wagner by a reserve army that doesn’t exist. It was even reduced to sending small flying columns into Belgorod Oblast to launch terror raids, only to have them blown up - discovering that the Russian border is in fact crawling with forces of the very much intact Russian army. 

I think that ultimately, neither army anticipated that Bakhmut would become the focal point of such high-intensity combat, but the arrival of Ukrainian reserves in force created a unique situation. Russia was beginning a process of major force generation (with mobilization beginning in September), and the gridlocked, slow-moving, Verdun-like environs of Bakhmut offered a good place for Wagner to bear the combat load while much of the regular Russian forces underwent expansion and refitting. 

Ukraine, meanwhile, fell into the sunk cost fallacy and began to believe its own propaganda about “Fortress Bakhmut”, and allowed brigade after brigade to be sucked in, turning the city and its environs into a killing zone. 

Now that Bakhmut is lost (or as Zelensky put it, exists “only in our hearts”), Ukraine faces an operational impasse. Bakhmut was after all a very good place to fight a static defense. If the AFU could not hold it, or even produce a favorable loss exchange, can a strategy of holding static fortified belts really be deemed viable? Meanwhile, the failure of the Syrskyi plan and the defeat of a multi-brigade counterattack by Russian motor rifle brigades casts serious doubt on Ukraine’s ability to advance on strongly held Russian positions. 

Ultimately, both Ukraine and Russia traded for time in Bakhmut, but whereas Russia put up a PMC which primarily lost convicts, Ukraine bought time by chewing up a significant amount of its combat power. They bought time - but time to do what? Can Ukraine do anything that will be worth the lives it spent in Bakhmut, or was it all just blood for the blood god?