The history of hnefatafl

Where it came from

The tafl family is a group of hunt games: one small force at the centre against one large force around it. Boards and playing pieces that look like tafl equipment turn up in northern Europe from roughly the fourth century onwards, in graves, in bogs, and in the waterlogged layers of trading towns. They come from Scandinavia, from the Baltic, from Ireland, from Scotland and from the Northern Isles.

The commonest explanation for where the idea came from is Roman. The Roman army played a two-player board game called ludus latrunculorum, the game of little mercenaries, and the fragments we have of it describe a game with orthogonal movement and capture by trapping a piece between two of the enemy — which is, in outline, the tafl capture. The Roman game was played along the empire’s northern frontier, and northern peoples had every opportunity to see it.

That is a plausible line of descent, and it is worth saying clearly that it is nothing more. Nobody has found a transitional board, or a text describing the borrowing, or any statement by an ancient or medieval writer that one game came out of the other. What we have is two games that resemble each other and two peoples who were in contact. The resemblance may be inheritance, or it may be that custodial capture is simply an obvious idea that people arrive at more than once. Treat the Roman ancestry as a hypothesis with reasonable support, not as a fact.

One thing the evidence does settle: tafl is not an ancestor of chess, a relative of chess, or a version of chess. Chess arrived in Europe from India by way of Persia and the Islamic world, with two matched armies of differently moving pieces. Tafl has one kind of soldier, unequal sides and a fugitive king. The two games share the idea of a marked board, and nothing else.

Where it was played, and when it stopped

Tafl travelled with the people who played it. Boards have been recovered across Scandinavia, in Viking-age Ireland, in Scotland and the islands, and as far as Greenland. The Welsh played a version called tawlbwrdd, the Irish a small game called brandubh, and the Sami of northern Scandinavia a nine by nine game called tablut, which they were still playing in the eighteenth century.

Board sizes vary, and the variation is real rather than an artefact of poor preservation. Seven by seven, nine by nine, eleven by eleven and thirteen by thirteen boards have all been found. The families of names vary too. Whether these were one game with local sizes or several related games is not something the finds can answer.

Tafl also appears in literature, though never with its rules. Norse sagas and poems mention the game and its pieces in passing, the way a modern novel might mention a hand of cards without explaining the deal. Irish texts name brandubh and describe a board of seven times seven with five men at its heart, which is a description of the layout and not of the play. These references prove the game mattered. They do not tell us how a turn worked.

From about the twelfth century the references thin out, and chess takes over the space that tafl had occupied. In most of the region the game was gone by the late Middle Ages. In two places it hung on: Wales, where a copyist was still able to draw a board in 1587, and Lapland, where it was a living game two centuries later still.

What the sources actually say

The honest summary is short. No complete rule set for any tafl game survives from the period when the game was widely played. Every modern code is built from three thin kinds of evidence.

• Objects. Boards, gaming pieces and the odd king-piece from graves and bog deposits. These give us board sizes and, sometimes, the number of pieces. They cannot give us a single rule.
• Passing mentions. Sagas, poems and glossaries that name the game or a piece without describing the play, plus riddles and figures of speech that only make sense if the reader already knew how it worked.
• Two late witnesses. A Welsh drawing from 1587 and a Latin traveller’s notebook from 1732. These are the only accounts that try to say what the layout and the play were, and both of them are incomplete.

The Welsh witness is Robert ap Ifan, who in 1587 copied out a figure of a tawlbwrdd board: a king at the centre of a ring of twelve men, with forty-eight attackers arranged around them. His figure and his numbers do not sit comfortably on any board that has actually been found, which is why modern tawlbwrdd is a reduction of what he drew rather than a copy of it.

The other witness is Carl Linnaeus, who watched Sami players during his journey through Lapland in 1732 and wrote down what he saw. His is the only account we have from somebody who was actually looking at a game in progress. Even so it was written in Latin, at speed, by a botanist rather than a games writer, and it leaves several questions open — most famously the exact conditions for taking the king. Almost every modern tablut disagreement traces back to a gap in those notes.

The great modern gathering-up of all of this is H. J. R. Murray’s A History of Board-Games Other Than Chess, published in 1952, which pulled the scattered evidence for tafl into one place and gave later players something to argue with. Murray was working from the same thin sources as everybody else, and his reconstruction has been picked apart since, but it is where the modern conversation starts.

The modern game

Tafl came back in the second half of the twentieth century, first as a curiosity in books about old games, then as something people actually played. Playing exposed a problem that reading had not: reconstructed tafl is easy to make unbalanced. Change the escape squares from the corners to the edge, or let the king be taken by two pieces instead of four, and the whole game tips.

The response was to write codes rather than to recover them. On the Shetland island of Fetlar, a panel drew up a set of eleven by eleven rules for an annual tournament, deliberately plain and easy to teach. A group of players in Copenhagen went further and added three extra endings — the shieldwall, the edge fort and encirclement — specifically to close off the stalling and blocking strategies that spoiled tournament games. The Copenhagen code is now what most clubs and most software mean by hnefatafl.

It is important to be clear about what those groups did. They did not discover lost rules. They wrote good ones, tested them against real play, and published them. A modern hnefatafl game is a playable descendant of a medieval game, in the way that a restored building is a descendant of the one that stood there: the footprint is genuine, and much of what you see above it is careful modern work.

That is also why this site keeps six codes rather than one. Copenhagen and Fetlar are the modern tournament games. Tawlbwrdd, tablut and brandubh sit closer to a particular historical source, with all of that source’s gaps. And Ard Ri is included as a curiosity: no medieval text describes it, and the version that circulates in books and software looks like a twentieth-century invention made by analogy with the bigger games. Every one of them is labelled here for what it is.

Sources

Everything on this page is written from scratch. These are the works it leans on, and what each one settles.

• Carl Linnaeus, Lapland journal, 1732

  The only surviving account by somebody who watched tafl being played. It gives us tablut: the board, the two armies and the general shape of the play. It is also incomplete, so this site states its own settlement of the points Linnaeus left open, above all the capture of the king.

• Robert ap Ifan, tawlbwrdd figure, 1587

  A Welsh copyist’s drawing of a tawlbwrdd board, with a king ringed by twelve men and a large force of attackers around them. It is the source for the Welsh game, and because his figure cannot be read literally onto a real board, this site follows the usual modern reduction to eleven by eleven with twenty-four attackers.

• H. J. R. Murray, A History of Board-Games Other Than Chess, 1952

  The standard scholarly survey, which collected the archaeological and literary evidence for tafl in one place. It settles nothing about play on its own, but it is the reason the modern debate exists and the common reference point for every reconstruction since.

• The Fetlar Hnefatafl Panel rules

  The eleven by eleven code written for the annual tournament on Fetlar, in Shetland. It settles the plain version of the modern game on this site: corner escape, a king who needs all four sides covered, and none of the extra win conditions.

• The Copenhagen hnefatafl code

  The tournament rule set agreed by the Copenhagen hnefatafl group and adopted by most modern clubs. It settles this site’s default game, including the three extras that the Fetlar code leaves out: the shieldwall capture, the edge fort win and the win by encirclement.