I'm incredibly excited: the latest Kickstarter update says that Spirit Island has been printed and is on its way towards our shores! I'd been holding off on more designer diaries until the production issues were hammered out - now that that the games are in transit I feel like it's not too much of a tease to start these up again.
From the very beginning, the island has been divided into lands, where each land is "a distinct place". Invaders in one land aren't in another one; a Power used on a land affects only things which are there; range is measured in "# of lands distant"; etc.
But the earliest boards didn't look at all like the current ones.
The very first version was a hand-scrawled, non-modular island. While I had the notion that every land would eventually be a particular type of wilderness, the Invaders ignored terrain, expanding along arrows from one land to another.
This didn't work very well in solo experimentation - the arrows seemed super-busy and hard to parse. (I've since seen it done better, I think in Legends of Andor?) But that experimentation also led to the Explore-Build-Ravage system the Invaders use (more on that in another post), which in turn made it clear that terrain was important for more than just Spirit abilities.
The second board looked like this:
It's still a single island, but with a hex layout rather than organic hand-drawn lands. Messing around with this validated the basic Invader AI concepts, but the game scaling was a bit problematic.
At this point, I back-burnered the project for a while - I was daunted by its sheer scale, the prospect of how much stuff I'd need to create and playtest to do it justice as a strongly thematic game - none of my prior designs had involved so many asymmetric positions, loads of special powers, or much in the way of worldbuilding.
I pulled it back into focus in early 2012 after a conversation with Ted Vessenes, who was really enthusiastic about this Spirit Island idea. Enthusiasm from other people has always been a strong motivator for me, so I dusted off my notes, did a huge pile of brainstorming on "things that spirits might do", and came up with new maps:
(B) is the initial hex-based board draft from early 2012. I decided the best way to scale the game for more/fewer players was to scale the size of the island; plus, modular boards would allow for game-to-game variability. I went for a weird shape so the island wouldn't look like a square or a simple hexagon (few islands do), though with regular hex borders between boards so they could (in theory) interlock in lots of interesting ways.
(C) is the first board any playtesters ever used, at Intercon in March of 2012. I made the lands organic (though the boards themselves remain hex-based); it seemed more thematic and made it clearer that each area of contiguous terrain was one discrete location (rather than "3 hexes of forest"). There's very little coastline (which at this point had no mechanical function), 10 lands (of 5 different terrain types), and a "Spirit homeland" - that funny circle-flare land near the top. (The initial working title was Genius Loci, and each spirit was the Spirit Of A Place - the volcano, the river-source, the ancient copse - and that special land was The Place. Each Spirit would have its own board.)
Those tests proved incredibly valuable, both for lessons learned and for enthusiasm. One map-related lesson was that playtesters could not keep different types of rainforest conceptually distinct in their head - at least, not well enough to use them as different terrain types, so terrains would have to be a little more iconically distinctive and a little less realistic. On the enthusiasm front, people were not merely wanting to play again but dragging friends over and saying "come try this out!". That's unusual for a game in such a rough state, even with friends, and was a strong signal to keep working on this game.
(D) post-dates (C) by about a week, updating the board shape - the combinations with hexes just weren't working out as interestingly as I'd hoped; there were too many "not quite right" fits with insufficient contact between parts of the board. I fell back on my knowledge of tessellations and experimented with a bunch of triangle/square/hexagon variations, eventually settling on a rhombus with organically distorted edges that (since they tessellated) would interlock neatly. The (much expanded) ocean edge is deliberately different so you can't place it as an interior edge by accident. The tessellation also let me define a set of fixed boundary-points for lands, so that land borders on one board would never be touching/too close to land borders on an adjacent board, eliminating the "...are those lands touching by 1mm?" problem.
This revision also introduced setup iconography: #1 and #2 in white boxes for starting Invader territories, brown discs for Dahan, red discs for Presence, stars for Sacred Sites (which were a separate type of piece back then). I used #1 and #2 rather than any iconography because I was still experimenting heavily with what an appropriate starting complement of Invaders was.
I waffled back and forth over whether to craft the board tessellations so that you could flip the boards over and use Side B (offset by half a board-width) with Side A for more variety - I eventually decided against it because it made the edges look too regular / non-natural. As things turned out, I'm doubly glad I did this, because the thematic maps (on the reverse side) really shouldn't be mixed play-wise or graphically with the balanced maps on the front.
(E) is from a month or two later. It drops from 5 terrains to 4, and with that from 10 lands to 8. I came to this decision reluctantly, as 10 lands allowed for more nuance in the positional maneuverings - carving out defended areas that the Invaders couldn't Explore into. But 5 terrains was just too high-variance, and it took too long on average before a given terrain got revisited (if it did at all). But the core dynamic was still there... and as it turned out, the thematic side let me play around with 9-10 lands per board again, so all's well that ends well.
This is the last revision with the Spirit Land - the next iteration dropped it, and with it a host of special-case rules and possibilities for stalled games:
(F) is from later that month. Between dropping the Spirit Land and going from 10 lands to 8, piece overflow happened a lot less often.
(At this point, I had my first child, and my design pace slowed waaaaay down.)
(G) is from 2013, several iterations later, which were mostly about updating iconography. All lands now have numbers, both so they can be referenced by setup instructions / Adversary effects / events, and also so that lands can be unambiguously referenced. Invader pieces are shown explicitly rather than implied. It's also tilted, as that allowed for fitting a slightly larger map on an 8.5"x11" piece of paper. Note the exterior tic-lines around the 3 land-based edges; those are the possible points where lands can have boundaries.
There were many iterations after that, but little change in their general appearance - the further exploration was more about board topology, and what sorts of boards work well or poorly for interesting and balanced gameplay. For instance, the board shown in (G) got cut because land #4 is adjacent to every other land on the board (and at least one off of it). This was a small problem because you could block off the Invaders from lands 6 and 7 by keeping just that one land clear, and a bigger problem because Spirits could set up a Sacred Site in it and have range-1-from-Sacred-Site access to more than an entire board. Other trials found that having variance in the number of coastal lands made a noticeable difference to gameplay, particularly after the "Invaders Explore from Ocean" rule was introduced.
Finally, we have the (almost) production versions. For the balanced sides of the boards, the graphic designer decided to stick with a textured look (albeit with professionally commissioned textures rather than my off-the-shelf slapdashery) and a stitched-border motif to make it feel like a map the Dahan might use:
(The black lines on the Wetlands are a texture artifact of that particular proofing file; they aren't in the later proofs I looked at.)
The rulebook gives standard ways to combine the balanced boards for 2p/3p/4p play, but if you want to (and you don't mind a bit of change in game difficulty) you can mix it up and fit them together however you like. (Rules of thumb: the more edges connecting between boards, the harder the game. The more coastal cul-de-sacs with few land connections, the harder the game. The more inland cul-de-sacs, the easier the game.)
For the thematic side of the boards, Greater than Games commissioned a large artwork map of the entire island, with a more graphical / satellite's-eye-view style:
The terrain types are still distinct, but not so blazingly different, and there's some shading into each other (though not so much that you can't tell which terrain type each land is). There's a bit of variation in what a given terrain looks like - see the Sands at top left vs. the one at bottom right - and some ornamentation that doesn't affect gameplay, like the river through land #1 and the picture of A Spread of Rampant Green in land #5.
...and that brings us up to now! The boards are one of the things I'm most looking forward to about playing with a final-printed version of the game - paper boards curl, and slide over each other so easily; having boards that actually bump up against one another will be lovely. :-)
Mostly for design diaries and retrospectives, perhaps branching out into posts on more general design thoughts.
- [+] Dice rolls