I’ve been toying with this game design for a few weeks. The primary thematic element involves running children across minefields. Those that survive the trip get to escape to presumably live good lives. Those that don’t, well, don’t. The players effectively take the part of family heads determining which and how many of their own children to send out when for potential slaughter or safe passage (they have lots of kids). Key elements are that the game is semi-cooperative, explicitly zero-sum, inferential deduction of likely landmine locations, setting up other player’s kids to clear the path by stepping on and blowing up the landmines for your kids to then follow safely, deliberately sacrificing your own early kids in order to determine and secure a path for later stock etc. The goal (scoring) is simple maximal genetic survival, primarily of direct family then extended and then remote family.

A minor goal of the game would be to provide an opportunity for discussion and education on topics ranging from landmines to exploitation of children, sacrificial costing, personal versus family/group survival, lose-lose decision structures, the meaning of personal sacrifice, relativism etc.

Basic premise:

• Hex map
• All hexes contain tokens. Tokens are either mines or dummy mines. Tokens are face down.
• One military player against the rest. This player plays the side of the military. (In a large player count game there may be two such players?)
• Players have two types of tokens (direct family and indirect family)
• Players start at locations along one edge of the hex map.
• Two types of locations: locations exclusive to each player and locations which start out with a mix of tokens (lids) for different players.
• Initial allocation of tokens to locations is done round-robin during setup.
• After setup players round-robin give each other relationship markers for in-laws and cousins (one of each)
• Each location has a token pool and a queue
• Turn order is variable. Details TBD.
• On their turn a player may do one of 3 things:
  • Move one of their kids already on the minefield N steps across the minefield, or move one of their kids at the front of a location queue N steps into the minefield. (N/2 if injured kid)
  • Acquire more kids
  • Move a token in a location (their’s or other’s) from the pool to the queue
• Each action may be chosen a limited number of times. Once two actions are exhausted all actions reset their use counters.
• When two actions exhaust their use counters military player gets to (secretly) add more mines to the minefield
• Initial mines are placed secretly.
• Mines come in two types. Simple kill mines and scatter mines (kills and scatters half-kill mines in surrounding hexes. The player’s stock of mines is public.
• Player also has dummy mines.
• At the end of the turn the player must feed their kids. Food resources are proportional to family death rate: the more die in your family the more food received from charity organisations
• When a kid moves on the field the military player states for each hex entered whether a mine there explodes. If a mine explodes the military player must reveal the mine, deliver the appropriate damage, and distribute new half-kill mines as applicable.
• Military player receives new stock of mines proportional to number of kids killed
• Game ends when all of one family has been eliminated or made it across the minefield.
• Players score points for members of their (extended) family that made it safely across, less points for cousins and less again for in-laws (the other players). Player with the most points wins.
• Kids behind the minefield do not score.

New rules.

New player aid.

Brand new game player introduction (with pictures of the game in play).

Most of the many changes are due to the generous help and commentary of Scott Russell.

Additional playtests have put the game time solidly around 150 minutes with new players, a little less with experienced players.

New rules.

New player aid.

In a fit of masochistic self-pity I proffered the rules for criticism on the Spielfrieks mailing list. Happily they weren’t gentle.

• The most common complaint was resolution of turn order, specifically the handling of values of shells versus fish and whether or not multiple players could bid on the same route. It is likely that “bid” is the wrong word for this section but I’ve yet to come up with a better one. Rewritten

• Clarified that bidding on a route also explored it

• Clarified that deliveries could use other player’s routes

• Clarified that only directly connected routes produce during delivery

• Clarified that discarded fish, shells and kula are returned to the supply

• Clarified that kula may be re-given

• Re-ordered prestige generation rules for clarity

• About Explorers and Proas section removed, content rewritten and folded into the Explore and Delivery sections.

• Duplicate content removed from Delivery (specification of stopping at first island with a matching market)

• Further annotated and signalled Advanced Game portions

• Other small edits

• Matching changes made to player aid

New rules.

New player aid.

Slight edits to the map:

• Added a red section to the prestige track to call out the game ending condition of 33 or more prestige

• Removed the previous player aids on the map as they’ve been supplanted by the actual player aids

– Trimmed the size of the background image to be narrower as we had difficulty fitting the game on the narrow tables in this weekend’s sessions

As I’ve had difficulty finding a good map of the thousand islands region of the St Lawrence River, I’m going to toss the Earthsea map and just use the same base map for three players by just varying the setup:

• Back down to 5 colours of markets instead of 6. This should change the rate of market stacking on development from 1/3 to 2/5 (ie slightly more frequent). The result will be a marginally more resource and thus prestige rich game, but also one in which delivery opportunities and intersections matter more.

• One less market per development stack. This should accentuate the slight starvation patterns encouraged by the reduction in colours. It should also encourage a more diffuse network building pattern. Happily the math works out too.

Both changes should lead to a tighter, leaner and more aggressive game, as suited for 3 players.

There were several more playtests this weekend. All played smoothly and I’m well pleased. I’ll attempt to remember to get the pictures off my camera and post them. Calling out two sessions in particular:

Session 1:

4 players, 3 new, playtime of 165 minutes. Game ended on 33 prestige but would have continued no more than 2 more turns before one market ran out. All four players had more than 30 prestige. Less than a dozen points spanned from first to last, making this the closest game to date. The eliminated player had the second highest score before elimination.

The new player aid was frequently used in the rules teaching and was regularly referred to during the first quarter of play (up until just after focus moved from resources to kula). We started out with a 3 player game but then added a fourth (Eddie) just after the rules explanation. Jason did a notable job of teaching Eddie the rules while I ate lunch. While I gave him a copy of the rules, he taught the game entirely from and with frequent reference to the player aid. Jason called out the player aid as both useful both in comprehending the game and in adding a comfort level for players attempting to digest the game’s complexity (‘’I can ignore the details, concentrate on the big patterns, and look up the details on the player aid later when I need them’‘).~~

Session 2:

4 players, 2 new, playtime of 130 minutes. Game ended on 33 prestige but would have continued no more than another turn before one market ran out. Two players had more than 30 prestige. More than 30 points spanned from first to last. One of the players from the first session above beat me by 1.5 points – he had second-most VPs but just enough more prestige than me to take the win. The eliminated player had the lowest score before elimination and appeared overwhelmed for much of the game. Again the player aid was frequently referred to during rules teaching and up until shortly after the kula race started.

While players were (almost) always able to answer their questions from the player aid, there seemed to be some uncertainty on where to look on the player aid for the answer to a particular question. My current summary is that the player aid has the right factual content, but may need a presentation/sectioning/titling adjustment.~~

New rules.

New player aid.

Added endgame condition of 33 prestige.

Slight formatting and language changes.

We’ve done four more 4 player playtests without much event. However the last couple 3 player playtests have been more interesting:

First session

After just over 15 minutes of rules explanation we played a 3 player game in a little under 150 minutes. The game worked well and generally as predicted and was well received.. Oddly almost 3/4 of market development stacked immediately. Unlikely. As a result scores were unusually high in the early game, and deliveries were scarce and aggressively fought over (turn order) Exchanging VPs for resources remained popular far later in the game than is usual and as a result the game never got into the mad struggle to convert resources into prestige that is desired. Yet…it worked. Perhaps not ideally in my mind, but it functioned and was more than interesting.

As a result of that playtest I added an island and a few routes to the Earthsea map to try and balance out the (highly thematic) advantages of c These are the same changes mentioned in the last blog entry.

Second session

This time the market distribution was equally unusual with almost no stacking and an extremely even distribution of market colours across the board and within development stacks. The result was a 210 minute game before three colours finally all ran out on the same turn. Aggressive play (as the leader) on my part could have shortened the game but I would have had to sacrifice positional advantage to do so. Prestige levels were absurdly high due to the continuously rich delivery field. Whereas games will usually end with the highest prestige in the high 20s to low 30s, I ended the game with 67 prestige with another player also above 60 and the trailer in the 50s. Not good. The game was clearly determined long before it ended (I ran away by almost a factor of 2). This suggests that an alternate game ending condition which would trigger at the early determination point is indicated. Functionally the game worked. Both players found it enjoyable with the second place player happily bemused and pondering at game end (a desired response that also occurred in the previous 3 player session).

With the current game end definitions play length is effectively a function of the normality of the market distribution. If the random market distribution tends towards clumping and stacking, then the game will run shorter. The more the random market distribution runs toward an (equally unlikely) even (and unstacked) distribution, the longer the game will play. The range is roughly from 135 minutes - 240 minutes assuming “typical” players. The goal is 150 minutes with new players. An obvious temptation is to end the game when a player achieves a prestige of 35 or more. That’s just far enough into the final multiplier bracket that a close competitor is likely to follow, but not so far that the game will exhaust. The result should be that only unusually even distributions invoke the end condition and that the new condition should curb the game appropriately.

Roke remains too strong. Much as I Like the idea of a specifically 3 player map, it is not clearly needed or clear that the default Polynesian map wouldn’t function equally well for 3 players. The problem is that Roke effectively sits in the centre of the map, doesn’t have the problems with draining too quickly like Havnor and is trivially connected to all the major lobes of the map. The real problem is that the map is roughly circular and asymmetric with Roke in the lynchpin position. I expect that I’ll abandon the Earthsea map rather than fix it. Much as I like Earthsea (I adored the books as a kid) the long term licensing fees and problems surrounding a licensed product are too large to bother with. I’m tempted to do a Science Fiction map set in the worlds of Perry Rhodan (an absurdly popular SF pulp series in Germany that I also like), or, more likely, in the more than 1,800 islands along the St Lawrence River between Kingston and Brockville. At least there the amerindians have something of a gift economy tradition to fit that oh-so-essential theme and the map is naturally long/narrow.

New rules – no big changes, just tweaks on wording.

Polynesian map rescalled and trimmed for easier play, but no graph changes.

Earthsea map added an island and a couple links to stretch the endgame by 1-2 turns.

The new map images are a little easier to play on due to better node spacing.

New rules.

I’ve moved kahuna to the Advanced Game and for now I’ve decided to go with the increased kula prices rather than shifting the multiplier ranges. According to the spreadsheet it works out well enough. Oh, and I upped the market colour count up from 5 to 6. We’ve played several games this way and it makes market delivery patterns slightly tighter earlier in the mid-game and brings in the end-game a turn or two faster.

The only rules change required for the smaller Earthsea map is a reduction in the market distribution to match the smaller set of nodes (25) and distribution stack markets (80) (as versus 30 nodes and 95 distribution stack markets in Polynesia).

No other changes seem to be necessary or called for in the game other than the above. It is feeling very close to done. Now to test the Earthsea map and re-verify the Kahuna-less game.

Much more work and thought has been going into this project than I’ve had time to document. Some notes:

• Most important: Scoring is (good_press-bad_press)^2+money. Ties go to highest good press. Player with the most bad press is auto-eliminated

• Event cards come with money (total defined on card). Other values on events are good and bad press. Good press slightly outnumbers bad press. This guarantees a winner outside of ties.

• Draft an event higher up the stack results in money from the bank going on lower cards

• Select an event and resolve it against your network, or play money and resolve it against another player’s network

• There are three phases per turn: networking, event, press resolution.

• In networking each player plays two cards, or plays one card and pays money to move one network card

• In press resolution rounds only one step of the graph is resolved, not the full graph.

• After all event cards are processed, networking and press resolution alternate until graph is empty

Another set of playtests on the Polynesia map went extremely well.

Despite my efforts the game persists in trying to last about 3 hours with all new players. 2.5 hours is definitely in reach for an all-new player group, but it is a fight.

The only aspect that I’m still fiddling with is the Kahuna. I’ve been convinced for a while that they are not strictly necessary to the game, but I like them and so left them in, seeing them as a possible expansion mechanism. I’ve played and without them. One of today’s games however revealed their downside with efficiency-minded players: They’re resource production magnifiers and when they’re used efficiently resource production rates get very high (I was producing almost 30 fish plus more than a dozen shell every turn) and the lower value kula tokens are thus simply not as interesting. That’s a problem. Fortunately the obvious resolution is simple:

1) Remove kahuna from the base rules and make them an optional advanced game

2) Rescale the prestige multipliers to every 15 rather than every 10 when using kahuna

or:

2) Increase the kula token costs to 5/11/7

That’s simple enough (I’ll probably drop the second rule change for the third after I work over the spreadsheet some more). Everything else is working well. Hopefully we’ll get in another game on Monday and again on Tuesday if Corrupt Benifecence doesn’t distract with its own playtesting.

Recent playtests have also started to clearly reveal that the game really has three phases. It always did, but it took a bit before I noticed it so clearly:

1) The first portion of the game is dominated by infrastructure building and is very zero sum competitive.

2) The mid and most of the late game is dominated by sustaining the income rates required for ideal gift-giving (kula).

3) The end-game, which is relatively short, is mostly defined by limiting and constraining the delivery opportunities of key players through route claiming and which specific markets are available for delivery when.

The only other change I’m looking to make is:

• 3 prestige may be discard to allow both deliveries in a turn to use one more proa than the player has

Not a biggie but it makes prestige even more fungible. It is a small change that, again, should occur in less than half of games, but it will make for more interesting end games.

I should remember to get pictures of the game in play posted…