A player may only challenge after they have completed a node and placed one of their tokens. The player can then immediately declare a challenge on any island (it does not have to be the one to which they have just added a token)

The challenger identifies a tribe that they control and a tribe that one of the other players control. Both of these tribes must be on the same island. The person with the smaller tribe (which could be the challenger), rolls one dice. The person with the larger tribe rolls from one to three dice. The number they roll is equal to the difference in the total number of tokens in the two tribes OR one dice (if the difference is less than one token) OR three dice (if the difference is greater than three tokens).

The winner of the challenge is the person who rolled the largest value on a single dice, or the challenger if the players' largest values were the same. The loser of the challenge takes back one of their own tokens from the losing tribe (the token for their own spirit that is nearest the top of the tribe). The winner adds one of the unplayed tokens for their spirit to the top of the losing tribe.

The effect of this is: in the losing tribe one of the tokens of the losing player is converted to a token of the winning player, while both tribes remain the same size and the challenging tribe is completely unchanged.

Yes. The tile may be part of a completed node.

Although tokens are generally placed on a node when it is formed, they are not associated with it in any way. If a tile that is part of a completed node is removed, any tokens on it remoin on the island.

The features on each tile consist of land or sea at each corner of the tile and an inward or outward jigsaw lobe on each side. The tile set is deeply symmetric in a mathematical sense.

All possible configurations of land and sea and in and out lobes on the side of a tile is equally likely. On the 136 tiles there are 8 different patterns on the side of a tile and each is equally common. So if you are matching on only one side (not forming a node) there will be exactly 68 edges in the set (on various tiles) which match that pattern. For symmetric patterns these will reside on 49 tiles, while for asymmetric patterns they will reside on 62 tiles. So regardless of the pattern you are trying to match, every second or third tile you check will match the pattern on at least one of its sides.

Most nodes are completed with a tile by matching the tile to the existing tiles on two of its sides. There will always be 8 tiles in the set that can accomplish this (at the start of the game). Of course as the tiles get used some patterns will be harder to fit.

For some people with good visual memory, it seems to be relatively easy to remember what the required tile looks like, although it does take some practice to get used to the patterns.

I personally find that a search strategy based on a mnemonic shape & pattern descriptor greatly facilitates the search process. For example, some of the tiles have a shape that resembles a horses head.

If the tile above is to be placed so that the top corner would form a node, then a sufficient description for such a tile would be: "a horse's head looking left with land on its head and its belly, and sea on its back". This completely specifies the space to be filled to complete the node, is easily remembered and can be used to evaluate each piece you look at.

For the sharp corner there are only four shapes of tile (I call them "horse looking left", "horse looking right," "broad spear head," and "sharp point"). For the blunt corner, I call the corresponding shapes "wave breaking left," "wave breaking right," "bracket," and "spade point". Of course you can give them whatever names you like.

By first looking for an appropriate shape then checking the land/sea pattern for matching shapes, we have found that this strategy can reduce the search time to well under 30 seconds to identify several tiles that will complete a node.

The tiles are designed with all possible patterns of land and sea on the tile-boundary for each shaped tile, with the constraint that the coastline must intersect the tile boundary in the middle of one of the inward or outward jigsaw lobes. What the land does inside the tile boundary does not affect this.

Rather than make these "peninsula" tiles, we chose to make tiles with two small pieces of land on each of the sharp corners like this:

This has the same boundary configuration as a 'peninsula tile' but has water separating the pieces of land from each other. This tends to produce a larger number of separate islands in the game than would occur with a peninsula configuration.

So called "analysis paralysis" can be a problem in many games. What a number of players have done when they encounter this problem is to use a sand timer to limit the time each player has in their turn. If the turn is not completed in the time then the player returns any tiles they are holding to the tile pool and the turn passes to the next player. Any challenges already underway should be completed.

I personally do not favour this solution. I would be inclined to either gently encourage the person to speed up their play by doing some of their analysis while they are waiting or failing that to find someone else to play the game with.

Such deep analysis is, in my view somewhat misguided, and misses the point of the game. For me the game is not really about this style of deep tactical analysis, and the relatively difficult nature of the tile matching task makes such an analysis (to more than one or two moves ahead) quite challenging. I find there is much greater benefit in negotiating appropriate alliances with other players for particular islands. Some tactical advantage can be gained in the early part of the game by considering how various moves affect the the resulting land configuration, and this is worth doing. However, this generally does not require a player to look at many options, nor to look too far ahead.

Deeper tactical play would mean trying to restrict the opponent's tile placement options and would require an analysis of both the resulting land configuration and the resulting interlock configurations. This greatly increases the required analysis and is what leads to long delays. However, it is quite difficult to gain an advantage from this sort of analysis in the early part of the game, as the opponent will usually have several available tiles for any given response. Towards the end of the game some advantage can be gained in this way, but then the number of tiles that need to be searched is much smaller.

Of course this problem does not arise in the no- turns (real time) game.

This was designed to balance the most valuable land nodes when playing with tribes. What we generally find is that it is very difficult to get a lake of more than two nodes. Either someone else grabs it first or they join it to open water and no-one scores for it.

In the game without tribes, there is an argument for reducing the value of the freshwater rights to match those of the land-node tokens. But again the lakes are very difficult to get. Making the freshwater rights worth so much changes the character of the game, but not in a way that unfairly benefits any individual player.

One of the most challenging things in making this game was cutting the jigsaw-like tiles. With a normal jigsaw, the pieces go together the way they are cut so they always line up perfectly. But if you take the same piece from two different copies of the same jigsaw, the alignment of the picture on the two tiles can vary by as much as ±3 mm. This is why jigsaw makers don't supply replacement pieces.

In Sunda to Sahul, the tiles go together in lots of different ways. So we have to have the shape of the tiles cut to an accuracy of better than ±0.2mm. As well as this shape accuracy, the cut must be in the correct place on the pictures as nearly as possible. We originally specified ±0.5 mm accuracy for the alignment of the cut to the picture, but during prototyping it became clear that this could not be achieved reliably in production. Our final specification was for alignment accuracy of ±1 mm. By and large the production run achieved ±0.5mm or better. But some of the sets do have a small number of tiles with alignment errors of as much as ±1mm. When two of these tiles are joined together so that the errors are in opposite directions, the misalignment between the patterns on the tiles can be as much as 2mm. If you have tiles with alignment worse than this, please contact us and we will arrange for a replacement set of tiles.

Firstly, all of the spirit names are fictitious. This is set 50,000 years ago so there did not seem to be much point drawing the names from modern languages. The names however are built on simple linguistic patterns to give them a consistent sound.

The use of the totems is something of a cheat and is certainly not based on any single existing culture, but many Australian and New Guinea cultures consider each individual to belong to one of several clans within a single community, and for each clan there is usually a totem (most often an animal spirit from the dreamtime in the aboriginal tradition).

In the game, the totems for each of the spirits are as follows

Marajun

Based on a New Guinea mud mask These are usually of simple design, are often decorated with various coloured ochres and could well date back to the period in question - although there is no direct evidence for this.

Akambu

An adze (a tool for making dugout canoes) based on the design of a New Guinea adze in the National Museum of Australia. The actual artifact is of course too modern for the 50,000 BCE date assumed in the game and it is not clear whether dugout canoes or the adze would have been in use at that time.

Sekuma

This is the shell of an Australian sea snail (just one I picked up on a beach so I am not sure of the species). In ancient Australia, many of the communities on the coast relied heavily on shellfish. Evidence of this can be seen today in the vast mounds (or middens) of shells, up to 20 metres high in some cases, that can be found all around the Australian coast.

Tumek

This is the tooth of a tiger shark, which is native to the Indonesian archipelago.

No. The islands that you get are not based on any real islands. They are simply a function of the tiling patterns and my artistic license in drawing their shapes to look as realistic as I could manage while satisfying the constraints required by the tiling. Strangely enough though if you look at a map of the modern island of Sulawesi, it bears an uncanny resemblance to some of the islands that are produced in the game. Stranger still, the large gulf at the bottom is called the "Gulf of Bone".