Hnefatafl and Tablut: Sources and Corrections

No Viking rules for hnefatafl survive. The fullest rules for any game of its family are the Latin notes Carl Linnaeus made in his Lapland diary under 21 July 1732, on a Sami game called tablut, and in that Latin a king caught between two enemies is removed like any other piece: "etiam Rex", the king too. The English edition of 1811 said the king was "not liable to this misfortune", and the first modern reconstruction, in 1913, was built on that English. The popular 11 by 11 rules played today are modern designs of 2007 and 2012. This page gives the source for every slide, and says plainly what is still uncertain.

Every factual claim in our hnefatafl carousel and in its caption, with its source. Section numbers 01 to 05 match the number at the top of each slide; 07 to 10 are the later slides' places in the carousel.

Slide by slide

Cover. In 1732, two men could take your king. Today's 11x11 rules need four.

The first half of the cover is rule 9 of Linnaeus's notes on tablut, set out under 03 below. The second half, "Today's 11x11 rules need four", is the king-capture rule of the two popular 11 by 11 rule sets, Fetlar and Copenhagen, set out under 07.

Both halves are shorthand. In Linnaeus's rules the king needs four enemies around him on his castle square and three beside it; the two-man capture applies everywhere else. In the modern rules four is the usual case, but three are enough beside the throne, and on the board edge he cannot be captured at all, except in one endgame case in Fetlar. And 1732 is the year Linnaeus wrote the rules down, not the year the game began.

01. The Vikings left the pieces, not the rules

The grave is at Baldursheimur, by Lake Mývatn in the north of Iceland. The National Museum of Iceland holds the gaming set and its die as Þjms. 1, and its English display text reads: "It comprises 24 pieces, turned from marine mammal tooth. The die is cut from the leg bone of a cow or horse, and the hnefi or 'king' is a carved, whalebone, man like figure. When the game was found, half the pieces are said to have been red, and half white." The Icelandic record calls the burial karlmannskuml frá 10. öld, a man's grave of the tenth century.

The whalebone king, Þjms. 6, is a man who seems to squat on a base and heldur báðum höndum um langt, klofið hökuskegg, holds a long, forked beard in both hands. The man who registered him in 1863 first wondered whether he was a chessman or a small idol, then judged a hnefi from a hnefatafl set hið líklegasta, the most likely. The record also says that no board was found, taflborð fannst ekki; it has probably decayed in the ground.

The finder's report of 1860, quoted in the museum's record of an iron ring from the same grave, admits that young men dug the grave out carelessly. The gaming pieces lay on the dead man's right, with his spear, and about an ell beyond the foot of the grave lay a horse, í meðallagi stór, og úngur, of middling size and young.

The slide shortens the museum's last sentence, which in full reads: "The rules of the old game are not known with any certainty, other than that both sides took turns in attacking and defending the king." We know the shape of the game, one side defending a king against the other. We do not know its rules.

Sources: Þjóðminjasafn Íslands, Sarpur records 331609 (Þjms. 1, set and die), 331620 (Þjms. 6, the king) and 331618 (item 11, with Arngrímur Gíslason's report).

02. A botanist wrote them down

Linnaeus left Uppsala on 12 May 1732 and was back on 10 October, by his own journal. The Linnean Society of London, which holds the diary, notes that he "was just 25" when he set out, and names tablut among the Sami games it describes.

The rules are not a report of a game he watched. In the English edition they sit among his general notes on the Sami under the heading "July 21", an entry that opens with the diseases of reindeer; under "July 23" he "took leave of the alpine part of Lapland". So the slide says only that he noted the rules in his diary under 21 July.

The board is his own sketch, which the English editor, James Edward Smith, had cut in wood so exactly "that every stroke of the pen, even the most casual, is retained". It is a plain ruled grid of nine by nine, 81 squares, with no alternating colours. The centre square is the king's castle, "called konokis by the Laplanders". The Latin legend reads Sveci N:r 9 cum rege, the Swedes, nine in number with the king: a king and 8 Swedes. Rule 13 gives the other side: Muscovitae sine rege erint, suntque 16 in 4 phalangibus disponendi, the Muscovites have no king and are 16, set out in four groups, one in the middle of each edge. The king wins at the edge: rule 3 lets him leave along an open line, and rule 4 says Si Rex tali modo exit, est praelium finitum, if the king gets out this way, the battle is over.

Swedes and Muscovites are Linnaeus's Latin labels, which is why the slide adds "as he called them"; the only words he gives as Sami are konokis and the two calls in rule 6. Nor did he publish any of this. In the preface to his first volume, Smith calls the manuscript "such a journal as a man would write for his own use". It was first printed in English in 1811, and in its own Swedish and Latin in 1889 and 1913.

Sources: C. Linnaeus, Lachesis Lapponica, or a Tour in Lapland, ed. J.E. Smith (London, 1811), vol. I, preface and entry of 12 May 1732, and vol. II, entries of 21 and 23 July, pp. 55 to 58 with the woodcut; Linnean Society of London, "Linnaeus in Lapland" (2020); the Latin as cited under 03.

03. In his Latin, two men can take the king

Rule 9, as printed by Theodor Fries in the 1913 edition of the Iter Lapponicum: Si qvis hostem inter 2 sibi hostes collocare possit, est occisus et ejici debet, etiam Rex. If anyone can place an enemy between two of his own enemies, it is killed and must be removed, the king too. Ewald Ährling's earlier edition, in Carl von Linnés Ungdomsskrifter (1889), ends the rule item Rex, likewise the king.

Rules 10 and 11 make the exceptions. On his castle, the king can slip away while three sides are held, and si qvatuor hostes in 2, tum rex captus est, if four enemies hold the squares around it, the king is taken. Beside the castle, rule 11 needs three. Rule 9 is the general rule and includes the king; 10 and 11 add harder conditions around the castle only. Hence the slide: "Only on his castle square or beside it does it take four or three."

Both editions are printed transcriptions, by different editors, and they differ in small readings, etiam or item among them. The manuscript is in London, at the Linnean Society, and we have not seen it.

Sources: T.M. Fries (ed.), Iter Lapponicum, Skrifter af Carl von Linné V (Uppsala, 1913), pp. 154 to 156; E. Ährling (ed.), Carl von Linnés Ungdomsskrifter, second series (Stockholm, 1889), pp. 147 to 148; both read in the Internet Archive scans.

04. The English edition flipped it

Smith's Lachesis Lapponica, volume II, gives rule 9 as: "If any one man gets between two squares occupied by his enemies, he is killed and taken off, except the king, who is not liable to this misfortune." Nothing in either Latin edition stands behind that closing clause, and it says the opposite of etiam Rex. The English list also stops at rule 12, so rules 13 and 14, the kingless Muscovites and a castle that captures like an enemy piece, are missing. And it adds what the Latin lacks: "a checkered board" and "the sixteen embroidered squares", where the Latin has only locis characterisatis, marked squares.

The slide's "Its editor wrote that the translator knew little Latin" rests on the preface to volume I, where Smith writes that the manuscript was "Swedish, but so intermixed with Latin, even in half sentences, that the translator, not being much acquainted with this language, found it necessary to leave frequent blanks". Smith then rewrote the whole, "often obliged to supply the forms of whole sentences". The translator was Charles Troilius.

The chess historian is H.J.R. Murray. In A History of Chess (1913) he calls it "extremely probable" that tablut "is identical with the old hnefatafl", sets out the rules from Smith's English, renames the Muscovites "Russians", and writes: "The King cannot be taken". His footnote cites the 1811 edition, volume II, pages 55 to 58. J.C. Ashton (2010) traces how later game writers built on Murray. We stop there on purpose: the Latin of 1732, the reversal of 1811 and Murray's rule of 1913 are documented, but we cannot show that any modern rule author chose a four-sided capture because of Smith.

Sources: Lachesis Lapponica, ed. J.E. Smith (1811), vol. I, preface, and vol. II, pp. 55 to 58, on Project Gutenberg; H.J.R. Murray, A History of Chess (Oxford, 1913), pp. 443 to 446, in the Internet Archive scan; J.C. Ashton, "Linnaeus's Game of Tablut and its Relationship to the Ancient Viking Game Hnefatafl", The Heroic Age 13 (2010).

05. "I play hneftafl better"

The scene is a mannjafnaðr, a comparison of men, between two brothers who ruled Norway together in the early twelfth century, Sigurðr and Eysteinn Magnússon. In C.R. Unger's edition of Morkinskinna, page 186, Sigurðr says he is the stronger man and the better swimmer, maðr sterkari ok syndr betr. Eysteinn answers: Ek em maðr hagari ok tefli ek hneftafl betr. Era þat minna vert en afl þitt. I am the handier man, and I play hneftafl better. That is worth no less than your strength.

The spelling on the slide is the manuscript's, hneftafl. Unger dates the manuscript to the first half of the thirteenth century, about a hundred years after these kings. It is a saga scene, not a record of a conversation, and the brothers were twelfth-century kings, not Vikings.

Sources: C.R. Unger (ed.), Morkinskinna (Christiania, 1867), p. 186; our Old Norse is normalised from Unger's diplomatic text.

07. Two or four?

Both modern rule sets are published by their authors on Aage Nielsen's site. The Copenhagen rules "were outlined in 2012 by Aage Nielsen (Denmark), Adam Bartley (Norway) and Tim Millar (UK)"; the Fetlar rules "were outlined 2007 by the Fetlar Hnefatafl Panel, UK". Both are played on 11 by 11, and both say: "The king is captured when the attackers surround him on all four sides, except when he is next to the throne", where three are enough. Neither lets him be captured on the edge, except, in Fetlar, when he is the last defender left and completely surrounded. The date 1983, sometimes given for Fetlar, is not on its authors' page.

Eleven by eleven is not a Viking size either. Grids in the evidence run from 7 by 7 to 19 by 19, and only one Viking Age board with a known grid survives complete: the yew Ballinderry gaming board, with "forty-nine holes", 7 by 7, and interlace in the Scandinavian Borre style. The National Museum of Ireland thinks it was "most likely" made in Dublin. Fragments from the Gokstad ship burial and from Coppergate in York have been reconstructed as 13 by 13 and probably 15 by 15. The 19 by 19 is a drawing in a twelfth-century Irish gospel book, the Alea evangelii, a Christian allegory of the same family. The 11 by 11 rests on a Welsh drawing of 1587, known to us only as transcribed by F.R. Lewis in 1943.

Sources: aagenielsen.dk, Copenhagen and Fetlar rules; National Museum of Ireland, "The Ballinderry gaming board"; C.A. Morris, Wood and Woodworking in Anglo-Scandinavian and Medieval York, The Archaeology of York 17/13 (2000); Kulturhistorisk museum Oslo, C10406; Corpus Christi College Oxford MS 122, fol. 5v.

08. Your name, the way runes wrote it

Bruce Dickins, introducing the rune poems in Runic and Heroic Poems of the Old Teutonic Peoples (1915), writes that the Scandinavian runic alphabet "continually reduced the number of letters, until by the ninth century no more than sixteen were left". That sixteen-letter alphabet is the Younger Futhark.

It had no separate rune for G or D: the k-rune served for k and g, and the t-rune for t and d. Dickins shows it with the larger Jelling stone, raised by King Harald Bluetooth: Haraltr kunukR, King Harald, with T for d and K for g; aft Kurm faþur sin, after Gorm his father; and Denmark as Tanmaurk. Dotted runes for G and D were added later.

That is why our rune translator cuts SIGURD with a K and a T, and why the slide's card reads "sixteen runes, Younger Futhark". The tool spells the letters you type; a carver writing the Old Norse Sigurðr would have cut its ð with the þ-rune, as the Jelling carver did in faþur.

Sources: B. Dickins, Runic and Heroic Poems of the Old Teutonic Peoples (Cambridge, 1915), introduction, pp. 5 to 8, with the Jelling inscription in his transcription; National Museum of Denmark, "The Jelling Stone".

09. Glad and cheerful, win or lose

The line is Hávamál, stanza 15: Þagalt ok hugalt skyli þjóðans barn ok vígdjarft vera; glaðr ok reifr skyli gumna hverr, unz sinn bíðr bana. Silent and thoughtful should a king's child be, and bold in battle; glad and cheerful should every man be, until his death arrives. The translation is ours, and "win or lose" is our gloss for a slide about a game, not the poem's words. Hávamál survives in the Codex Regius, written in Iceland around 1270.

The blade is our Viking seax knife with a 12-inch forge-finished blade: 1075 high-carbon steel, dark from the forge along the spine and polished to the edge, on a rosewood handle. GLADR OK REIFR is the stanza's glaðr ok reifr in capitals, with D for ð. The words are optional: you choose them, and we laser-engrave them to order, letter for letter.

Sources: Hávamál, st. 15, Old Norse text on heimskringla.no.

10. The chain named for a king

"King's chain", or "king chain", is what jewellers call this dense braid of interlocking links. We know of no early source for the name and no Viking Age chain that goes by it. That is all the slide's "far older" rests on: the king on the game board goes back at least to the tenth-century grave at Baldursheimur. Our king's chain is 100 per cent stainless steel, built link by link by hand, in three link widths and three lengths.

Caption. Raicki, the riddle, and the fist

Raicki. Rule 6 in Fries's text: Ut Rex aditum apertum vidit, clamet Raicki, si duae viae apertae sunt tuicku. When the king sees an open way, let him call raicki; if two ways are open, tuicku. Ährling prints Raich' and tuicha, Smith raichi and tuichu, and Smith makes it a duty: whoever moves the king "must call out".

The riddle. In Hervarar saga ok Heiðreks, Gestumblindi asks: Hverjar eru þær brúðir, er sinn drottin vápnlausan vega; inar jarpari hlífa um alla daga, en inar fegri fara? Roughly: who are the brides who fight around their weaponless lord, the darker shielding him every day while the fairer go forth? King Heiðrekr answers: Þat er hnettafl; inar dekkri verja hnefann, en hvítar sækja. That is hnettafl; the darker ones defend the hnefi, and the white ones attack. This is manuscript R, of the late fourteenth or early fifteenth century. The Hauksbók version was compiled around 1300, the date the caption gives, but for this part it survives only in seventeenth-century copies; it spells the game hneftafl and has "the red ones" follow the king. No version gives numbers, so Murray's reading of the riddle as sixteen white men against a king and eight dark ones is not in the text. Another riddle in the saga, a creature with eight horns and no head, is sometimes quoted as the king piece; its answer is húnn, a thing that "runs as soon as it is thrown": a die.

The fist. Cleasby and Vigfússon's Icelandic-English Dictionary (1874) defines hnefi first as "the fist", and second as "the king in a kind of chess played by the ancients". Tafl is an early loan from Latin tabula. The game is the hnefi's board, not a "king's table". In Grettis saga, chapter 70, Þorbjörn öngull "tefldi hnettafl; þat var stórt halatafl", a big board whose pieces had tails, pegs underneath. The Baldursheimur pieces each have a hole underneath; the match is our inference, not the saga's.

Most texts say only tafl, which can mean any board game. In Völuspá the gods play tafl (stanza 8) and after Ragnarök find gullnar töflur í grasi, golden gaming pieces in the grass (stanza 61); in Orkneyinga saga Earl Rögnvaldr Kali boasts Tafl em ek örr at efla, I am quick to play at tafl. We do not count these as hnefatafl.

What the sources do not support. That we know the Viking rules: the museum says they are not known with any certainty, and Murray records that the scholar Willard Fiske, "after devoting several years to the recovery of the game, gave the problem up as insoluble". That the king stood for Odin: we found no text or museum record that says so, and the Baldursheimur registrar, who did wonder about an idol, named no god. The archaeologist Peter Pentz argues that beard-holding figures such as the Fløgstad figure were hnefatafl kings, tied to masculinity, royalty and fertility, not to Odin. That the game trained warriors: no Norse text says so. Of the Birka grave Bj 581, where a board and a bag of pieces lay with the dead, Neil Price and colleagues write that the full gaming set "suggests a potential command role". That is an interpretation of one grave, not a record of training.

Chess came later. Cleasby and Vigfússon find "no authentic record of chess" in Scandinavia before the twelfth century, and the National Museum of Iceland thinks hnefatafl fell out of use around 1200, as chess spread. The Lewis chessmen, carved around 1150 to 1200, belong to that later world, though the scholars who studied the hoard in 2009 suggested some pieces "may have been for hnefatafl, or both chess and hnefatafl".

Sources: Fries 1913, p. 155, and Ährling 1889, p. 148, rule 6; Smith 1811, vol. II, p. 57; Heiðreks saga, ed. Jón Helgason (Copenhagen, 1924), R and H texts, and C. Tolkien (trans.), The Saga of King Heidrek the Wise (1960); Cleasby and Vigfússon, An Icelandic-English Dictionary (Oxford, 1874), s.v. hnefi, tafl and skák-tafl; Grettis saga, ch. 70, Völuspá, st. 8 and 61, and Orkneyinga saga, on heimskringla.no; Murray 1913, p. 443; P. Pentz, "Understanding the Flygstad (Fløgstad) Figurine: Gaming Pieces, Kings, Gender and Fertility Rites", Medieval Archaeology 69 (2025), 169 to 199, abstract only; N. Price et al., "Viking warrior women? Reassessing Birka chamber grave Bj.581", Antiquity 93 (2019); D.H. Caldwell, M.A. Hall and C.M. Wilkinson, "The Lewis Hoard of Gaming Pieces", Medieval Archaeology 53 (2009), p. 198.

What we do not claim

• That any Viking rules survive, or that the rules you play are Viking rules, or that they come straight from Linnaeus. Tablut is 9 by 9; the popular 11 by 11 rules are Fetlar (2007) and Copenhagen (2012), modern designs.
• That the Copenhagen or Fetlar rules are a translation error. The documented chain is the Latin of 1732, Smith's English of 1811 and Murray's reconstruction of 1913, and it stops there.
• That tablut and hnefatafl are the same game. Murray thought it "extremely probable"; Linnaeus says nothing about where tablut came from.
• That Linnaeus watched a game on 21 July 1732, or published his notes, or that the Sami called the sides Swedes and Muscovites. "Embroidered" and "checkered" are Smith's words; the sketch has no alternating colours and gives no colours for the pieces.
• That Linnaeus's Latin is certain in every word. We have it from two printed transcriptions, 1889 and 1913; the manuscript is in London.
• That the sagas are eyewitness history. They were written down in the thirteenth and fourteenth centuries, and they spell the game hneftafl or hnettafl. Nor is every tafl hnefatafl: the word covers board games in general.
• That the Baldursheimur pieces are whalebone or walrus ivory. The museum says marine mammal tooth, with a whalebone king.
• That the Vikings played on 11 by 11 boards. Board sizes vary from 7 by 7 to 19 by 19, and only the 7 by 7 Ballinderry board is complete with a known grid.
• That the Ockelbo runestone, which shows two figures at a board, still exists as carved. The original stone, Gs 19, was destroyed in the church fire of 1904; what stands at Ockelbo is a copy raised in 1932, and the scene is known from nineteenth-century drawings.
• That the game was played with dice, or without them. Some sets were buried with a die, and no one knows why; tablut has none.
• That the king represents Odin, that Vikings used the game to train warriors, or that they brought it to Ireland and Britain, where related games had names of their own.
• That the Vikings played chess, or that the king's chain is a Viking chain.

Corrections

None yet. If you find an error, tell us, with a citation.

Frequently asked questions

What were the original rules of hnefatafl?

Nobody knows. No Viking Age source sets out the rules, and the National Museum of Iceland says they are not known with any certainty, beyond one side defending a king while the other attacks. The fullest rules for any game of the family are Carl Linnaeus's Latin notes on the Sami game tablut, written in his diary in 1732.

Can two pieces capture the king in hnefatafl?

In Linnaeus's tablut, yes. His Latin rule 9 says a piece caught between two enemies is removed, the king too, except on or beside his castle square, where four or three enemies are needed. The English edition of 1811 reversed this, and the two popular modern rule sets, Copenhagen and Fetlar, need four attackers around the king.

What does hnefatafl mean?

Hnefi means fist in Old Norse, and in the game it was the name of the king piece. Tafl, borrowed early from Latin tabula, means a board game. So hnefatafl is the board game of the hnefi. The medieval manuscripts spell it hneftafl or hnettafl, and it does not mean king's table.

Did the Vikings play hnefatafl on an 11 by 11 board?

There is no evidence that they did. The only complete Viking Age board with a known grid, from Ballinderry in Ireland, has 7 by 7 holes, and other finds are fragments reconstructed at 13 by 13 or 15 by 15. The 11 by 11 size goes back to a Welsh drawing of 1587, known only from a later transcription, and to the modern Fetlar and Copenhagen rules.