Battlefield Atlas
Fall of Byzantium · 1071 First 3 stages free

Manzikert, 26 August 1071

On 26 August 1071, near the frontier fortress of Manzikert (Malazgirt, north of Lake Van in eastern Turkey), the Seljuk sultan Alp Arslan defeated and captured the Byzantine emperor Romanos IV Diogenes. Romanos had divided his army, sending a large part toward Khliat (Ahlat), and retook Manzikert; after two days of skirmishing he advanced across the plain against a Seljuk army that fell back before him. When he turned back to his camp at dusk, the rearguard under his rival Andronikos Doukas withdrew, the wings broke, and the emperor was surrounded and taken. Losses were not enormous, but the defeat and the civil wars that followed opened Anatolia to Turkish settlement. Where exactly the battle was fought is still being investigated: the finds so far lie at work sites between Gülkoru and Malazgirt within a survey area of about 150 square kilometers, no single field has been confirmed, and the medieval name of Alp Arslan's camp (Rahwa) is a generic word for 'plain'. This map follows that strip of finds, and all unit positions on the plain are reconstructions.

Play the first 3 stages free7 stages on modern satellite imagery and terrain, with 3D elevation

The first 3 of its 7 stages are free to play. The rest opens with Battlefield Explorer ($4.99 a month or $39 a year), which unlocks every battle in the atlas and every one added later. Lifetime access and war packs are one-time options. Free battles include Thermopylae, Plataea, Actium, Viking raid on Lindisfarne, Hastings, Hattin, Agincourt, Vítkov Hill, Kutná Hora, Fall of Constantinople, First St Albans, Lützen, Siege of Buda, Culloden, Fort William Henry, Yorktown, Lake Erie, Waterloo, The Alamo, Monterrey, Antietam, Little Bighorn, Manila Bay, Belleau Wood, Saint-Mihiel, Omaha Beach, Inchon, Ia Drang, Goose Green, 73 Easting, Wanat, Baghdad Thunder Runs, Manbij, Battle of Raqqa, Donetsk Airport, Ilovaisk, Debaltseve, Battle of Kyiv, Battle of Hostomel Airport, Siege of Mariupol, Bilohorivka (Siverskyi Donets crossing), Kharkiv counteroffensive, Kherson counteroffensive, Bakhmut, Vuhledar, Mala Tokmachka and Robotyne, Avdiivka, Kursk incursion.

Alp Arslan and the captive Romanos IV
Alp Arslan and the captive Romanos IV — Bibliothèque nationale de France · Public domain

The battle, stage by stage

  1. About 20–22 August 1071

    The emperor comes east

    In the spring of 1071 Emperor Romanos IV Diogenes leads a large army — Byzantine regulars, provincial troops and hired Normans, Franks, Uzes and Pechenegs — into Armenia to retake the frontier fortresses the Seljuks had taken. Modern estimates of its size range from about 40,000 to 60,000–70,000; medieval claims of 200,000 or more are not credible. From Theodosiopolis (Erzurum) it marches by Hınıs toward Manzikert, which Sultan Alp Arslan's men had seized. Believing, on reports that may have been wrong, that the sultan has gone south toward Syria and Baghdad, Romanos divides his army: a large part — by some accounts the larger and better part — goes south toward Khliat (Ahlat) on Lake Van under Joseph Tarchaneiotes, with Roussel de Bailleul's Frankish mercenaries. The routes shown are schematic.

  2. 23 August 1071

    Manzikert retaken

    On 23 August Manzikert's Seljuk garrison, faced with the emperor's army, gives up the fortress with little or no fight, and Romanos garrisons the black-basalt walls and pitches his camp outside the town. Alp Arslan, who had been campaigning in Syria, has in fact turned back and is close at hand; sources disagree on the route he took. Tarchaneiotes' detachment is some 60 km away by road at Khliat and, in the end, plays no part in the battle. Where the camp stood is not known; it is drawn just south-east of the town, toward the plain where battle finds have been made.

  3. 24–25 August 1071

    Skirmishes on the plain

    Turkish horse archers begin harrying Byzantine foragers. Romanos sends out the Armenian general Nikephoros Basilakes with cavalry; he pursues too far toward the Seljuk camp and is routed and captured (probably on 24 August; the date is inferred), and Nikephoros Bryennios, sent with the left wing to retrieve the situation, only narrowly fights his way back, wounded. On the 25th a Seljuk embassy offers terms and Romanos refuses them; he also learns that the Khliat detachment is not coming back. Part of the Uz mercenaries, kin of the Seljuk Turks, go over to the enemy on the 25th, the day before the battle; the Pechenegs and the other Uzes stay. Kadri Perk (1946) put Alp Arslan's camp near Harabereş on the Süphan foothills south of Gülkoru, where it is drawn; his reading has been significantly challenged by fieldwork since 2020.

  4. 26 August 1071 (Friday), afternoon

    Advance into empty ground

    Romanos draws up in front of the camp: himself in the center with the guard and tagmata, Bryennios on the left, Theodore Alyates on the right, and a rearguard of the eastern magnates' retinues and levies under Andronikos Doukas, nephew of the former emperor and no friend of Romanos. Alp Arslan, said in Muslim tradition to have put on a white robe like a burial shroud, has a mostly mounted army, commonly estimated at about 20,000–30,000 (there are no reliable figures; one modern summary gives the Byzantines about twice the Seljuk number), in a loose crescent a few kilometers off. The Byzantine line advances in good order, but the Seljuks will not stand: the crescent falls back while horse archers ride in on the flanks, shoot and wheel away. Hour after hour the army moves farther from its camp, and the rearguard, by one modern reading, falls far behind. Arrowheads and spearheads have been found at work sites between Gülkoru and Malazgirt within a survey area of about 150 square kilometers, but no single field has been confirmed, and Rahwa, the medieval name for where Alp Arslan camped, is a generic word for 'plain'. The lines drawn are interpretive.

  5. 26 August 1071, dusk

    The standards turn back

    Late in the day the Byzantines reach the Seljuk camp and find it abandoned. With evening coming on, an empty plain ahead, and a camp left lightly guarded behind, Romanos orders the imperial standards turned back toward camp. Turning a long line around in front of a mounted enemy is a dangerous maneuver. Parts of the army, out of sight of the center, take the turn of the standards for a defeat; the Seljuk troops held back in the hills begin to close in.

  6. 26 August 1071, nightfall

    Andronikos Doukas withdraws

    As the Turks attack the turning army, the rearguard under Andronikos Doukas, instead of coming up, marches back to the camp and away. Michael Attaleiates, who was in the camp, says many reported that Doukas spread word the emperor had fallen, and calls it treachery; Nikephoros Bryennios the Younger, writing later, blames the rearguard's failure without saying Doukas lied, and some modern historians (Theotokis) argue distance and confusion, not betrayal, explain it. The right wing breaks almost at once.

  7. 26–27 August 1071

    The emperor taken

    Bryennios' left wing tries to reach the center and is driven back. Romanos and his center are surrounded; the emperor fights on after his horse is killed and is wounded in the hand, then taken. Brought before Alp Arslan the next day, he is at first not believed to be the emperor. The sultan, by the accounts, set his foot on his neck as a token of submission, then treated him well and released him after about a week on terms of ransom, tribute and the surrender of several cities. What became of the Khliat detachment is unclear. Attaleiates says Tarchaneiotes led it back through Mesopotamia to Roman territory; Ibn al-Adim says it withdrew intact after the defeat; other Muslim writers claim Alp Arslan destroyed it. The route shown, west toward Muş (Theotokis's suggestion), is conjectural. Byzantine dead are put by modern summaries at perhaps 2,000–8,000; many units survived, but the eastern army's professional core was broken and the empire fell into civil war.

Forces

The Byzantine army that set out was probably the larger, perhaps about twice the Seljuk force, but Romanos fought with only part of it after sending a large detachment to Khliat. Every figure is an estimate; medieval numbers on both sides are greatly inflated.

Seljuk Turks (Sultan Alp Arslan)

Present (modern estimate)about 20,000–30,000

Commonly estimated at about 20,000–30,000, mostly horsemen; there are no reliable figures (Britannica). Medieval Armenian and Muslim writers give far larger numbers. Encyclopaedia Britannica, Battle of Manzikert; Mark Cartwright, World History Encyclopedia

Byzantine Empire (Romanos IV Diogenes)

Whole expedition (modern estimate)about 40,000–70,000

The World History Encyclopedia gives modern estimates of 60,000–70,000, about double the Seljuks, against medieval claims of 300,000; other modern estimates are about 40,000. Mark Cartwright, World History Encyclopedia

With the emperor on 26 Augustabout half the army or less

Romanos detached nearly half his army toward Khliat, by Theotokis's reading the larger and better part, and part of the Uz deserted before the battle. Medievalists.net, Betrayal or Mistake? (Theotokis); Mark Cartwright, World History Encyclopedia

Casualties

No reliable casualty figures exist for either side. Byzantine dead were probably a few thousand, and the real losses were the capture of the emperor, the destruction of part of the professional core and the civil war that followed. Certainty: poor.

Seljuk Turks (Sultan Alp Arslan)

Killed and woundedunknown

No source gives a usable figure for Seljuk losses; Britannica says there are no reliable figures for the battle. Encyclopaedia Britannica, Battle of Manzikert

Byzantine Empire (Romanos IV Diogenes)

Killedunknown; probably a few thousand (2,000–8,000 (modern summaries))

Britannica says there are no reliable figures; the World History Encyclopedia says the battle was not a catastrophe in casualties. The 2,000–8,000 range appears in modern summaries whose original basis could not be checked. Encyclopaedia Britannica, Battle of Manzikert; Mark Cartwright, World History Encyclopedia

Capturedthe emperor and an unknown number of others

Romanos IV was wounded and taken prisoner, and released after about a week on terms; Nikephoros Basilakes had been captured in the skirmishing beforehand. Britannica notes heavy losses among the professional, elite troops. Encyclopaedia Britannica, Battle of Manzikert; Mark Cartwright, World History Encyclopedia

Units lost before or outside the battlelarge but uncounted

The Khliat detachment, by some accounts the larger part of the army, took no part; part of the Uz mercenaries deserted on 25 August; the rearguard under Andronikos Doukas left the field intact (Theotokis, via Medievalists.net). Medievalists.net, Betrayal or Mistake? (Theotokis)

Then and now: how the ground has changed

The plain south-east of Malazgirt is still open, bare pasture and farmland at about 1,600 m, so the ground looks broadly as it did, but the town has spread beyond its walls and the exact field of 1071 has not been identified.

Weapons

Seljuk Turks (Sultan Alp Arslan)

Turkish composite bow

Turkish composite bow

Recurved composite bow · in use for many centuries

At Manzikert: Turkish horse archers shot and wheeled away all through the afternoon of 26 August, wearing down the Byzantine wings without standing to fight.

Small, very strong recurved bows built up from horn, wood and sinew. Mounted archers could shoot them at the gallop and wheel away before a charge reached them, the tactic that wore down Crusader columns on the march.

Build
horn, wood and sinew, glued and bound
Use
shot from horseback at the gallop
Range
effective to about 100–150 yd; farther for volleys
Photo: Daderot (a later Ottoman example, Higgins Armory Museum) · Public domain
Straight Islamic sword (sayf)

Straight Islamic sword (sayf)

Straight, mostly double-edged sword · in use for centuries; curved sabers spread later

At Manzikert: The straight sword was the sidearm of the sultan's household horsemen (ghulams) who closed in at dusk.

Saladin's horsemen fought with straight swords much like the Franks', alongside lance, mace and bow. The curved saber popularly linked with the Crusades became widespread in the Middle East mainly after the 13th century. This sword is a 13th-century Mamluk example.

Blade
straight, about 80–90 cm
Use
cavalry and infantry sidearm
Note
the curved scimitar of legend became common only after this period
Photo: Dosseman, Istanbul Military Museum (13th-century Mamluk sword) · CC BY-SA 4.0

Byzantine Empire (Romanos IV Diogenes)

Kontarion (Byzantine cavalry spear)

Kontarion (Byzantine cavalry spear)

Spear · Middle Byzantine, 10th–11th centuries

At Manzikert: The heavy spear of the Byzantine cavalry of the tagmata and the eastern themes, who made up the center and wings.

The long spear of Middle Byzantine cavalry. Romanos IV's center and wings were built around such horsemen, who could break any enemy that stood to receive a charge; at Manzikert the Seljuk horse archers simply would not stand.

Type
long cavalry spear (lengths varied)
Use
Byzantine heavy and medium cavalry
Sidearm
sword
Photo: Madrid Skylitzes · public domain
Byzantine lamellar armor

Byzantine lamellar armor

Body armor · Middle Byzantine, 10th–11th centuries

At Manzikert: Lamellar and mail protected the better-equipped Byzantine horsemen and the imperial guard.

Armor of small plates laced in rows, worn by the better-equipped Byzantine soldiers, often over or under mail and a padded coat. It protected well against arrows but was heavy and hot on a long August advance.

Build
small iron or hardened-leather plates laced together
Worn with
mail, padded coats, helmets
Wearers
heavy cavalry and guard troops
Photo: MedievalGreek · CC BY-SA 4.0 (reconstruction)

Paintings and photographs

Alp Arslan and the captive Romanos IV
Alp Arslan and the captive Romanos IV Alp Arslan humbling the captive Romanos IV, from a 15th-century French manuscript of Boccaccio (BnF Français 232). Bibliothèque nationale de France · Public domain · source
Battle of Malazgirt (medieval miniature)
Battle of Malazgirt (medieval miniature) The battle as imagined by a later medieval French illuminator. Public domain · source
Coin of Romanos IV with his co-emperors
Coin of Romanos IV with his co-emperors Gold coin of Romanos IV; coins bearing his image are among the survey finds on the plain. Wikimedia Commons · CC BY-SA 2.5 · source

Free for everyone every year on its anniversary: August 26.

Accuracy score: 50/100 (limited)

How the accuracy score works

Changelog

Corrections and updates to this map since 9 November 2026, newest first. All changes

Sources and method

Manzikert, 26 August 1071. No U.S. Army staff ride handbook covers this battle. Fixed points: Malazgirt Castle and town (coordinates from Wikipedia, used only for position), Gülkoru village, and Ahlat. Battlefield location: the archaeological survey led by Prof. Adnan Çevik (Muğla Sıtkı Koçman University) with Muş Alparslan University and the Ahlat Museum, reported in HeritageDaily (2024), Daily Sabah (finds between Gülkoru and Malazgirt), Daily Sabah (Afşin), Daily Sabah (2,000 finds), Hürriyet Daily News, TRT Haber and Sabah (1 August 2026); the review of Kadri Perk's 1946 field study and the Rahwa/Zahwa tradition in a DergiPark article. Narrative: Encyclopaedia Britannica, 'Battle of Manzikert'; George Theotokis, as summarized in 'Betrayal or Mistake? Rethinking the Defeat at Manzikert' (Medievalists.net, 2026); Mark Cartwright, 'Battle of Manzikert', World History Encyclopedia; Foundation of the Hellenic World. Contemporary and near-contemporary sources, known here through these works: Michael Attaleiates (in the camp; pro-Romanos), Nikephoros Bryennios the Younger, Michael Psellos, Skylitzes Continuatus and Zonaras, Matthew of Edessa, and the Muslim historians Sibt ibn al-Jawzi, Ibn al-Adim, Ibn al-Athir and al-Bundari. Orders of battle, dates of the skirmishes and casualty ranges were cross-checked against the Wikipedia article. Strengths are disputed and given as ranges; Byzantine and Muslim figures for both armies are greatly inflated. Camp sites, the lines on the plain and all routes are conjectural. Images from Wikimedia Commons.

Where the positions come from

Further reading