FACT-CHECK: Kostiantynivka, 31, 21, then 24 attacks—three true figures that no one ever reads together

On the evening of August 19, the second report, as of 10:00 p.m.

Since the start of the day: 189 clashes across the entire front, and 21 attacks repelled along the Kostiantynivka axis, around Ivanopillia, Kostiantynivka, Illinivka, Oleksandro-Shultyne, and Novoselivka.

Twenty-one, then—ten fewer than this morning, and a total that seems to contradict the surge reported just a few hours earlier.

Has the front calmed down?

It’s impossible to say: this report covers only fourteen hours, and fighting is still ongoing as it is published.

On neighboring fronts along the same fortified belt, the same snapshot shows 14 assaults halted near Sloviansk and 2 near Kramatorsk—numbers that are just as provisional, just as accurate, and just as poorly dated afterward.

A report that is, by its very nature, incomplete

The 10:00 p.m. report is a snapshot taken before the end of the film.

The night—the favorite time for infiltration groups—is not included.

Comparing it to the previous day’s total is like comparing halftime to an entire game; the general staff knows this, makes it clear, yet the hasty follow-ups overlook it.

Twenty-one attacks at 10:00 p.m. is not a lull: it’s a snapshot taken before nightfall.

Thursday, August 20, 8:00 a.m.

The report finally provides the full tally for the 19th: 234 clashes on the front lines and 24 attacks along the Kostiantynivka axis—around Kostiantynivka, Illinivka, and Ivanopillia, toward Novodmytrivka, Pavlivka, Oleksandro-Shultyne, and Novoselivka.

Twenty-four.

Three more than the 10:00 p.m. update: the night brought its own share.

Meanwhile, Pokrovsk has climbed back up to 26, regaining its status as the most-attacked axis.

The neighboring fortified belt is following the same pattern: 19 assault operations along the Sloviansk axis for the entire day, 4 along the Kramatorsk axis—compared to 16 and 1 in the previous morning’s report.

The pressure isn’t shifting; it’s spreading out.

The complete series, for the record

Let’s line up the mornings, since they are the ones that matter: 27 attacks on this axis in the August 18 report, 31 in the August 19 report, and 24 in the August 20 report.

A peak, not a decline—and a peak sufficient to turn a medium-sized city in the Donbas, for the span of a single morning, into the nerve center of the entire war.

Kostiantynivka bore the title of “primary front line” for just one morning—and that morning was a reflection of the day before.

31 was not a mistake, 21 was not a retraction, 24 was not a correction.

Let’s take a step back and look at this calmly.

This is where readers get lost—and where we lose them.

31: twenty-four hours ending at 8:00 a.m. on the 19th, so essentially covering the day of the 18th.

21: fourteen hours of the 19th, frozen at 10:00 p.m., with ongoing combat not counted.

24: the entire day of the 19th, frozen on the morning of the 20th.

Three time frames, three partial truths, zero contradictions.

The real problem lies elsewhere

The problem isn’t the consistency of the general staff, which publishes its “blackout” hours in black and white.

The flaw lies in the news cycle: news dispatches that list the three figures under the same date without specifying the time window, and news aggregators that pile them up.

Imagine the news editor—a generic, self-acknowledged figure—who has to choose a headline at 8:40 a.m.: he has three official figures before him, sixty characters of space, and no room to explain the time windows.

He’ll go with the biggest one.

Everyone goes with the highest number.

And the largest figure that morning was 31—an exact number, misdated.

At the end of the chain, a reader acting in good faith concludes—and we can’t really blame him for it—that “the Ukrainian figures contradict each other.”

That conclusion is false—and it plays right into the hands of the opposing side’s propaganda.

What a “clash” amounts to

Add a layer of uncertainty that no one mentions: the unit of measurement itself.

A “clash” or an “attack,” in these reports, is an event reported by units in contact: a repelled platoon-level assault counts as one, just as a push by two men emerging from a basement counts as one.

The actual intensity—troop strength, duration, support—is nowhere to be found.

Blind to the gravity of the situation, honest about the numbers: that is exactly what this tally is worth—no more, no less.

Three honest reports, poorly transcribed, create a lie that no one wrote.

As the clock ticks on, the Institute for the Study of War is examining the same city using other tools: geolocated imagery, sources from both sides, and control maps.

Its August 19 assessment does not count the assaults.

It describes a modus operandi: infiltration missions by small groups in and around Kostiantynivka to bypass Ukrainian positions and cut off supply lines.

Footage dated August 18

Geolocated footage published on August 18 shows, according to the institute, Ukrainian forces shelling Russian positions southwest of Kostiantynivka and northeast of Osykove—traces of these very infiltrations.

In other words: the enemy is not just on the outskirts—it is already, in scattered groups, within the very fabric of the city, between the positions and behind them.

Two men in a basement do not constitute a conquest.

A hundred cellars, however, do.

And Inflated Claims

The same report notes that Russian sources continue to “exaggerate” their claims of advances along this axis—the institute’s wording is harsher: they are inflating them.

Therein lies the full documentary paradox: Ukrainian reports that are true but misinterpreted, Russian claims that are exaggerated but widely disseminated.

In between, a city.

And a Washington-based institute that, unable to count the assaults, counts the meters—which is another way, neither better nor worse, of failing to see the whole picture.

Kostiantynivka is not under direct assault: it is being infiltrated, which is a slower and more serious process.

We must remember what this city stands for, and why every attack there carries more weight than elsewhere.

Losing this strategic corridor would not simply mean losing a line on a map: it would, according to the very logic Moscow espouses, open the way to what remains of the free Donbas.

Kostiantynivka is the southernmost of the four fortress cities in the Donetsk Oblast that Moscow has vowed to capture in order to seize the rest of the region.

Vladimir Putin said it himself, as quoted in late July: this city is “the key to the liberation of the entire territory of the Donetsk People’s Republic”—his words, not ours.

Estimates circulating in late July

Before his dismissal on July 21, the then-commander-in-chief, Oleksandr Syrskii, publicly estimated the number of Russian infiltrators in the city at around 145, claiming they were suffering losses of about 25 men per day—Ukrainian claims, unverifiable, as reported by Al Jazeera.

145 men.

In mid-July, the ISW estimated the Russian presence at about 37% of the city, primarily in the form of infiltrators; its subsequent assessments indicated that it had spread to most of the city.

37%, then more.

A Ukrainian soldier quoted in late July by Hromadske went further: in his view, the city was already effectively lost—a judgment based on the situation on the ground, not an established fact, which August reports have not yet confirmed.

The Bombs That Change Everything

And beyond the assault counts, there are the KABs: guided aerial bombs weighing 500 to 3,000 kilograms, dropped at a rate sometimes exceeding 300 per day across the entire front.

On August 19 alone, there were 330; on August 20, the number reached 330.

A Ukrainian commander engaged in the city, referred to in late July by the call sign “Kurt,” put it bluntly: his men held most of the city, but the guided bombs “wipe out positions”—the only thing, in his view, that was still bringing the Russians success.

A repelled assault counts.

A position wiped off the map by a one-metric-ton bomb isn’t counted in the assault tally—and yet it’s that very bomb that shifts the front line.

Assaults fill the tally; guided bombs, on the other hand, clear out the positions.

The method consists of three steps.

First: check the bulletin’s cut-off time—8:00 a.m. or 10:00 p.m.—before even reading the number, because that time is the document’s true signature.

Next: match the number to its time slot, even if it means correcting the date on the statement.

Finally: never add two statements together, and never conclude that there is a contradiction between two different time frames.

The time scale is also misleading

The same caution applies to long-term trends: according to ISW estimates reported in late July, Russia was gaining 16.6 square kilometers per day in the first half of 2025, barely 1 kilometer per day in June 2026—and it was losing ground, about 1.6 square kilometers per day, in July.

A rising assault count can therefore coexist with an army that is retreating.

Robert Brovdi, commander of Ukraine’s unmanned systems forces, quantified the resurgence of Russian offensive efforts in July: 12% more offensive personnel than in June, 18% more than in May—again, according to Ukrainian assessments.

Who Benefits from the Confusion

Every time a Western reader believes they’ve caught Kyiv red-handed in accounting inconsistencies, mistrust sets in—and this mistrust is free ammunition for Moscow, which no longer even needs to lie on this front: it simply needs to let our own misinterpretations do the work.

Rigor is not a luxury reserved for specialists.

It is a line of defense.

Misreading a true report is like manufacturing ammunition for the adversary yourself.

Let’s summarize what these thirty-six hours have shown—and nothing else.

On the morning of August 19, the Kostiantynivka sector was, with 31 attacks in twenty-four hours, the most heavily attacked sector on the front, according to the Ukrainian General Staff.

By evening, 21 attacks had been repelled there since dawn.

Then the total for the entire day on the 19th remained at 24; early the next morning, Pokrovsk regained the top spot with 26.

None of these figures contradict one another; none can be verified by a third party; all describe, window by window, a city that the enemy is gradually eroding through infiltration while guided bombs crush its defenders.

The city holds.

The counters keep turning.

They will keep ticking even when the city can no longer hold—or when it has been cleared; that is the nature of counters: they outlive everything they count—positions, villages, men.

Three sheets stacked on top of one another, three different numbers—and beneath them, a single city bleeding.

Tomorrow morning, at 8:00 a.m., a fourth sheet will cover the other three, and it will all start over.

And you—the next time two official figures seem to contradict each other—will you first look for the time at which each one closed its ledger?

Signed, Maxime Marquette, columnist