In depth | Defending Vuhledar — how one open dirt road stopped an armoured offensive

2026-10-13 · 2023-01-24 – 2023-02-23

Overview

Part 1: Defending Vuhledar (January–February 2023) told the story of the battle; this piece looks at the small decisions inside it. From 24 January 2023 Russian naval infantry attacked Vuhledar for a month without gaining a foothold [2][12]. The core of the defence was an engagement area (EA) on a dirt road south-east of the town: road left clear, fields either side mined, anti-tank teams in the tree lines, artillery laid on the road [1]. Below, terrain and weather are re-read through data, and the scenario's phases p1–p8 are taken in turn.

A five-storey apartment block with broken, partly fire-blackened windows behind a snowy yard — Photo 1. Vuhledar, 14 February 2023. Blocks like this gave the defenders their view over the fields — Photo: National Police of Ukraine (npu.gov.ua), CC BY 4.0, via Wikimedia Commons

Strategic and operational context

Russia's main effort that winter was Bakhmut; Vuhledar sat where the eastern and southern fronts met [8]. Its height put Pavlivka, the T-0509 road and the rail line supplying Donetsk and Mariupol within Ukrainian artillery range [2]. Russia had announced on 14 November 2022 that it controlled the town; it did not [2].

The attack belonged to the Eastern Group of Forces. On 23 February the UK Ministry of Defence (MoD) judged that its commander was under heavy pressure to deliver but probably lacked a force able to break through [5]. DFRLab linked the timing to the first anniversary of the invasion on 24 February [8]. The Hudson Institute noted that the 155th and 40th Naval Infantry Brigades were trained for coastal and amphibious tasks, not combined-arms manoeuvre [9].

The 72nd Separate Mechanised Brigade held the town [10][14]. Its artillery had remote anti-armour mine (RAAM) rounds, first listed in US aid in September 2022 [7]; about 6,000 were reportedly supplied by late 2022 [12].

Terrain and weather in detail (OAKOC)

Sketch map of the town, Pavlivka, Mykilske and the mined engagement area on the dirt road between them, with unit symbols and numbered Russian axes — Figure 1. The eastern engagement area (sketch; unit positions estimated; to February 2023 only). Places from OSM [17], elevations from Open-Meteo [21] — Own work, after sources 1, 2, 4, 6–8, 10

Observation. The town centre stands at about 190 m, Pavlivka at 150 m and Mykilske at 160 m [21]. On treeless steppe 30–40 m opens kilometres of view, and five-storey blocks add more (Photo 1). Live drone feeds reached the battalion command post [1].

Avenues of approach. Pavlivka lies about 3.4 km south-south-west of the centre and Mykilske 5.4 km south-east [17]; RFE/RL's "less than 2 km" reads as edge to edge [2]. The southern approach climbs out of the Kashlahach valley; the south-eastern one crosses 4–5 km of open fields on a few dirt roads. The north and west were bypasses, both tried in February [2][4].

Key terrain. The high-rise blocks, and the point where the dirt road enters the south-eastern edge.

Obstacles. The decisive obstacles were minefields, and the weather changed their meaning. ERA5 gives lows of minus 6 to minus 11°C for 6–10 February [16]. Frozen ground lets tracked vehicles leave the road, so the fields themselves had to be closed. Mining the fields while leaving the road open fitted that condition.

Cover and concealment. Windbreaks hid the anti-tank teams [1] and were the attacker's only cover; Russian infantry later used them (p6). On 8 February daylight ran from 07:49 to 17:42 [16].

Organisation and command relationships

Neither side published company-level dispositions; "est." marks estimates.

Side	Echelon	Unit / element	Role and equipment	Source
Ukraine	Brigade	72nd Separate Mechanised Brigade	Held the town	[10][14]
Ukraine	Battalion	1st Mechanised Battalion, 72nd Bde	Officers described the dirt-road ambush	[1]
Ukraine	Company	Forward, edge, reserve companies	Unpublished (est.: one forward, one or two on the edge, one reserve)	scenario
Ukraine	Teams	Javelin and Stugna-P teams	Hidden in tree lines beside the road	[1]
Ukraine	Tanks	T-64 crews	Camouflaged about 8 km behind the line	[1]
Ukraine	Fires	M777, Caesar, HIMARS	Registered on the road	[1]
Ukraine	Fires	RAAM rounds	Seeded over and behind advancing units	[7][9]
Ukraine	Special	Special Operations Forces	Claimed tank ambush, 3 Feb	[10]
Russia	Group	Eastern Group of Forces	Responsible for the operation	[5]
Russia	Brigade	155th Naval Infantry Bde (Primorye)	Led the assaults	[2][6]
Russia	Brigade	40th Naval Infantry Bde (Kamchatka)	Main assault force	[2][9]
Russia	Battalion	Alga (Tatarstan), Atal (Chuvashia), 3rd Army Corps	Alga at most c. 500 men (estimate)	[2][8]
Russia	Other	Elements of 14th Spetsnaz Brigade	Infantry assaults, mid-February	[8]
Russia	Equipment	T-80, BTR-82, BMP-3	Column of about 15 vehicles on video	[3][8]

The telling difference is where the authority to fire sat: the Ukrainian battalion command post watched the drone feed and timed the fire by radio [1], while on the Russian side inexperienced replacements made officers' micromanagement worse [6].

Day-by-day account

24–26 January: the night assault

Naval infantry in BMP-3s attacked at minus 5–6°C and took ground they could not hold under fire [3][16]. On 28 January the Moscow-installed head of occupied Donetsk claimed the town was encircled; a Ukrainian spokesman spoke of "fierce combat" [11].

27 January – 2 February: shelling and rework

The UK MoD called these probing attacks with no substantive gains [3]. While the town was shelled, the defenders mined the fields and hid anti-tank teams [1].

3 February: the ambush claim

Ukraine's Special Operations Forces and the 72nd Brigade said they stopped a T-80 heading for a firing position and captured two crewmen [10].

6–8 February: the armoured column

The night of 8 February fell to minus 9.7°C [16]. On frozen ground a column of about 15 tanks and armoured personnel carriers came down the eastern dirt road. When firing began, drivers turning round ran onto the mined verges, wrecks blocked the rest and the artillery struck the halted column [1]. No source gives the date or hour; the scenario uses 0700 on 8 February.

9–10 February: counting

UK MoD imagery from 9 February showed Russian vehicle losses concentrated around Vuhledar [6]. On 10 February the MoD assessed that at least 30 mostly intact armoured vehicles had been abandoned after one failed assault, and that Russian units had probably made tactical gains at the western edge [4]. ISW called the tactics on video "highly dysfunctional" [2].

11–23 February: infantry, the northern flank and the anniversary

Russia switched to infantry assaults under artillery cover, with occasional Lancet strikes [8]. The Rybar channel reported a failed flanking attack from the north [2]. On 23 February the UK MoD judged a breakthrough unlikely [5]. The town was still Ukrainian at the month's end.

Decision points in depth

The headings follow the training scenario; p4x and p6x are what-if phases (Figure 2).

Daily temperatures and snow days from 24 January to 23 February 2023 with phase bands and decision points p1 to p8 — Figure 2. Decision points and weather. Temperatures: ERA5 reanalysis for the Vuhledar grid cell [16]; decision times follow the training scenario — Own work

p1 The forward line gives way (25 Jan, 0600)

Knew: part of the forward line had fallen; the enemy was still moving in the open; the approaches were registered; a tank thrust was rumoured. Options: counterattack, fall back and strike from the heights, reinforce what was left, or recommend withdrawal. Decision: a single report says the defenders pulled back into the town; the Russians could not hold what they took [3]. Otherwise: an open-field counterattack would have spent the only reserve before the main assault.

p2 Reshaping under the guns (28 Jan, 1400)

Knew: mining the fields was possible now; closing the road too would take two more days. Options: close the roads, build an engagement area on an open road, or shelter and wait. Decision: the road was left clear and the fields mined [1]. Otherwise: on frozen ground a closed road would have pushed the enemy elsewhere and spread the defenders' fire.

p3 Pressure from two sides (5 Feb, 0900)

Knew: a local official (C3) spoke of encirclement; drones saw build-up mainly south-east; the reserve was one company. Options: hold with the reserve central, raid behind the lines, or shell assembly areas first. Decision: the town was held; the raid of 3 February was a special forces task [10]. Otherwise: rationed RAAM and shells spent early would have thinned the engagement area on 8 February.

p4 Column on the dirt road (8 Feb, 0700)

Knew: about 15 vehicles in single file, lead short of the EA (A1); crews newly mobilised. Options: fire at long range, wait for the whole column, meet it with tanks, or mine the road ahead. Decision: fire was ordered by radio once the column was inside; turning vehicles hit the mines and blocked the rest [1]. Otherwise: long-range fire would have scattered the column across frozen fields; mining the road ahead would have left it a way back. Waiting trapped front and rear, at the risk p4x explores.

p4x Armour at the eastern edge (8 Feb, 0800, what-if)

Situation: lead vehicles reach the eastern blocks while follow-on vehicles hesitate. Recommended: hold the buildings, engage the lead from upper floors and leave the follow-on vehicles to the artillery. In reality the column never reached the town [1].

p5 Wrecks in the fields (9 Feb, 1000)

Knew: wrecks had torn gaps in the verges; enemy engineers were at work at night; RAAM was rationed. Options: re-seed with remote mines, re-lay by hand at night, rotate troops, or shell the wrecks. Decision: the UK MoD credited RAAM, sometimes fired over and behind advancing units, as a factor in the losses [7]. One round lays a minefield in minutes (Photo 2). Otherwise: hand-laying in the open risks the engineers; open gaps invite the next attack.

Exterior and cutaway of a 155 mm shell with disc-shaped mines stacked inside — Photo 2. An M741 RAAM round, exterior and cutaway. The shell releases its mines in flight. Not taken at Vuhledar — Photo: Vslv, CC BY-SA 4.0, via Wikimedia Commons

p6 Infantry instead of armour (13 Feb, 0500)

Knew: small infantry groups along the tree lines at dawn and dusk; a rumour (F6) of tanks behind smoke. Options: drones and artillery while holding the buildings, ambush patrols, giving up outer blocks, or thinning the line. Decision: the defenders held; analysts saw no Russian advantage beyond occasional Lancet strikes [8]. Otherwise: RUSI describes expendable Russian infantry sent to expose firing positions for others to strike [13]; patrols would have walked into that.

p6x Enemy in the outer blocks (14 Feb, 0400, what-if)

Situation: two or three squads in an outer block, more gathering. Recommended: isolate the block with fire and remote mines, then clear it by day with a small group, in line with FM 3-90's use of local counterattacks to restore the defence [15]. In reality the line did not change meaningfully in mid-February [5][8].

p7 The northern flank (17 Feb, 1600)

Knew: signs of small groups in the northern tree line (B3); possibly a diversion (C3). Options: a small reserve and one anti-tank team north, a whole company, nothing, or pre-emptive fire. Decision: the Russian flanking attempt failed against fire from the heights and the minefields [2]. Otherwise: a whole company would have thinned the south; doing nothing would have left no time.

p8 Before the anniversary (22 Feb, 0600)

Knew: the enemy wanted a result but a breakthrough was judged unlikely; refusal reports were weak (C4); remote mines were nearly gone. Options: counterattack, hold and seal breaches, shorten the line, or raid. Decision: the defenders held; the UK MoD saw no Russian breakthrough force [5]. Otherwise: a counterattack across the fields would have reversed the roles of the past four weeks.

A camouflage-netted infantry fighting vehicle with Ukrainian markings carrying two soldiers across dry grass — Photo 3. A BMP-3 in Ukrainian service, July 2024, said by its unit to have been captured near Vuhledar — Photo: unknown author (Ukrainian MoD / 56th Brigade post), CC BY 4.0, via Wikimedia Commons

Doctrine re-read

Engagement area development (FM 3-90, ch. 8, paras 8-51, 8-52). The seven steps run from identifying avenues of approach and the likely enemy course of action, through deciding where to destroy the enemy and positioning weapons, to planning obstacles and fires and rehearsing [15]. Vuhledar's defenders did each of the first six [1]; only rehearsal is missing from the record.

Obstacle effects (ch. 17, paras 17-18 to 17-22, table 17-1). The effects are block, disrupt, fix and turn, produced by obstacles combined with fires, not obstacles alone [15]. Mines in the fields turned vehicles onto the road; wrecks in the verges fixed those behind; RAAM behind the column came close to disrupt [7].

The reserve (ch. 9, para 9-10). An area defence should succeed without its reserve, which lower echelons use for local counterattacks [15]. Holding it at p1, p3 and p7 fits.

Breaching (ch. 18, para 18-23). Reducing a defended obstacle needs suppress, obscure, secure, reduce and assault (SOSRA) [15]. The columns on video came in single file without smoke or engineers [1][2]. RUSI found the loss of a naval infantry brigade on such a thrust made Russia wary of armoured breakthroughs [13]. Russian manuals were not consulted; RUSI's observations stand in for them.

Animated map showing the six phases in turn; the still frame shows 23 February — Figure 3. Manoeuvre by phase (16-second loop; reduced motion shows only the final situation). Positions sketched and partly estimated — Own work

Source criticism

Vehicle losses. Ukraine's General Staff claimed at least 130 armoured vehicles including 36 tanks, which the NYT could not verify [1][2]. Video and aerial counts gave 20–40 [2]; the UK MoD, at least 30 abandoned after one assault [4]. The MoD figure is an imagery-based floor for one action; the Ukrainian figure is a three-week claim by a party to the fight. They measure different things.

Personnel. A 155th Brigade survivor said eight men of his company survived; a marines' open letter cited 300 lost, wounded or missing in four days [2]. The then UK Defence Secretary cited reports of 1,000 dead on 14–15 February [12]. Russian internal testimony runs against interest but is anonymous; the minister's figure, resting on unnamed reports, is the weakest.

Dates. The renewed offensive is dated 24 January [12] or "the last week of January" [2]. No source dates the column; the 9 February imagery [6] sets the latest date.

Causes. ISW blamed poorly trained mobilised troops [2]; Russian bloggers blamed fast Ukrainian reserves, which Hudson thought secondary [9]; RUSI pointed to how armour was employed [13]. Green replacements invite micromanagement [6], which repeats the same attack down the same road.

Control claims. Russia's November 2022 and 28 January claims were false [2][11]. The UK MoD's "tactical gains" at the western edge [4] and the NYT's "routed" [1] describe a local advance and a failed assault; both hold.

Lessons

An obstacle gathers before it stops. Leaving a road open can be part of the obstacle plan, above all when frozen ground opens the fields.
Whoever holds the information sets the moment to fire.
Obstacles are consumables. Plan how to re-seal them and ration the rounds that do it.
Spend the only reserve last. Meet flank threats and lost ground first with observation and fire.
Do not move troops on weak reports. An F6 rumour is a reason to watch harder, not to redeploy.

Unresolved questions

The date and hour of the armoured assault, and how many columns came.
The Ukrainian battalion's company layout and fire-control details.
Which Russian brigade attacked on which axis.
How many RAAM rounds were fired and what share of losses mines caused.
The 3 February ambush and the "Cossack" refusal are Ukrainian claims; Ukrainian losses were not disclosed [1].

The 72nd Brigade withdrew on orders in October 2024 as its flanks gave way [14].

Sources

Andrew E. Kramer, "In an Epic Battle of Tanks, Russia Was Routed, Repeating Earlier Mistakes", The New York Times (via Boston Globe), 2023-03-01. https://www.bostonglobe.com/2023/03/01/world/an-epic-battle-tanks-russia-was-routed-repeating-earlier-mistakes (accessed 2026-10-13)
Mike Eckel, "What Happened in Vuhledar? A Battle Points to Major Russian Military Problems", RFE/RL, 2023-02-17. https://www.rferl.org/a/ukraine-russia-battle-vuhledar/32276547.html (accessed 2026-10-13)
The War Zone, "Ukraine Situation Report: Heavy Fighting As Battle For Vuhledar Rages", 2023-01-28. https://www.twz.com/ukraine-situation-report-heavy-fighting-as-battle-for-vuhledar-rages (accessed 2026-10-13)
Ukrinform, report of the UK MoD intelligence update of 10 February: at least 30 armoured vehicles abandoned, 2023-02-10. https://ukrinform.net/rubric-ato/3668281-russian-troops-likely-abandoned-at-least-30-armored-vehicles-after-failed-assault-on-vuhledar-uk-intelligence.html (accessed 2026-10-13)
Euromaidan Press, report of the UK MoD intelligence update of 23 February: pressure to deliver at Vuhledar, 2023-02-23. https://euromaidanpress.com/2023/02/23/russian-command-under-pressure-to-deliver-success-at-vuhledar-but-this-is-unlikely-british-intelligence/ (accessed 2026-10-13)
Novaya Gazeta Europe, report of a UK MoD intelligence update: 155th Naval Infantry Brigade losses and replacements, 2023-02-26. https://novayagazeta.eu/articles/2023/02/26/uk-intel-russian-elite-infantry-apparently-suffered-extremely-high-casualties-at-vuhledar-en-news (accessed 2026-10-13)
Novaya Gazeta Europe, report of a UK MoD intelligence update: RAAM at Vuhledar, 2023-03-16. https://novayagazeta.eu/en/articles/2023/03/16/uk-intel-russia-suffered-heavy-losses-at-vuhledar-due-to-ukraines-use-of-raam-anti-tank-mines-en-news (accessed 2026-10-13)
Atlantic Council DFRLab, "Russian War Report: Failed Russian push on Vuhledar results in losses as Moscow increases troops", 2023-02-17. https://www.atlanticcouncil.org/blogs/new-atlanticist/russian-war-report-failed-russian-push-on-vuhledar-results-in-losses-as-moscow-increases-troops/ (accessed 2026-10-13)
Hudson Institute, "Ukraine Military Situation Report, February 15, 2023", 2023-02-15. https://hudson.org/defense-strategy/ukraine-military-situation-report-february-15-2023 (accessed 2026-10-13)
Ukrainska Pravda, Special Operations Forces and 72nd Brigade ambush behind enemy lines (Ukrainian claim), 2023-02-03. https://www.pravda.com.ua/eng/news/2023/02/03/7387764/ (accessed 2026-10-13)
France 24, live coverage of the fighting for Vuhledar, 2023-01-28. https://www.france24.com/en/europe/20230128-live-fierce-battle-raging-in-ukraine-for-control-of-vuhledar-near-donetsk (accessed 2026-10-13)
Nicholas Slayton, "How Ukraine is using US mines to decimate Russian tanks", Task & Purpose, 2023-02-25. https://taskandpurpose.com/news/ukraine-vuhledar-mines-decimate-russian-tanks/ (accessed 2026-10-13)
Jack Watling and Nick Reynolds, Meatgrinder: Russian Tactics in the Second Year of Its Invasion of Ukraine, RUSI Special Report, 2023-05-19. https://static.rusi.org/403-SR-Russian-Tactics-web-final.pdf (accessed 2026-10-13)
Sergii Kostezh, "Vuhledar: Why Ukrainian Troops Lost the Mining City", Kyiv Post, 2024-10-12. https://www.kyivpost.com/analysis/40288 (accessed 2026-10-13)
Headquarters, Department of the Army, FM 3-90 Tactics, 2023-05-01 (ch. 8 paras 8-51, 8-52; ch. 9 paras 9-10, 9-14; ch. 17 paras 17-18 to 17-22 and table 17-1; ch. 18 para 18-23). https://armypubs.army.mil/epubs/DR_pubs/DR_a/ARN38160-FM_3-90-000-WEB-1.pdf (accessed 2026-10-13)
Open-Meteo, Historical Weather API (ERA5 reanalysis), Vuhledar grid cell, daily temperature, precipitation, snowfall, sunrise and sunset, 24 Jan–23 Feb 2023. https://archive-api.open-meteo.com/v1/archive?latitude=47.781&longitude=37.246&start_date=2023-01-24&end_date=2023-02-23&daily=temperature_2m_max,temperature_2m_min,precipitation_sum,snowfall_sum,sunrise,sunset&timezone=Europe%2FKyiv (accessed 2026-10-13)
OpenStreetMap Nominatim, place search for Vuhledar, Pavlivka and Mykilske. https://nominatim.openstreetmap.org/search?q=Vuhledar&format=json&limit=3&countrycodes=ua (accessed 2026-10-13)
Wikimedia Commons, File:Vuhledar (2023-02-14) 02.jpg (National Police of Ukraine, CC BY 4.0). https://commons.wikimedia.org/wiki/File:Vuhledar_(2023-02-14)_02.jpg (accessed 2026-10-13)
Wikimedia Commons, File:RAAMS.jpg (Vslv, CC BY-SA 4.0). https://commons.wikimedia.org/wiki/File:RAAMS.jpg (accessed 2026-10-13)
Wikimedia Commons, File:UA BMP-3 Vuhledar.jpg (CC BY 4.0). https://commons.wikimedia.org/wiki/File:UA_BMP-3_Vuhledar.jpg (accessed 2026-10-13)
Open-Meteo, Elevation API, elevations for Vuhledar, Pavlivka and Mykilske. https://api.open-meteo.com/v1/elevation?latitude=47.781,47.7555,47.7503,47.80&longitude=37.246,37.2214,37.3026,37.20 (accessed 2026-10-13)