The DOOM Economy

December 3, 2016

I think they key thing about the DOOM economy in Colossus of Atlantis, is that I have absolutely no idea what will happen when the game is actually played. This is equal parts exciting and terrifying.


Made in China as a flower pot holder. Can you spot the Alien influence?

Player actions in the game will increase the DOOM score. If the total DOOM score from all player actions hits a secret and predetermined point, the game ends with Atlantis sinking beneath the waves. The team with the lowest DOOM score wins a moral victory. Up to five players can assure personal survival through the cataclysm if “The Ark” Great Wonder has been built. A couple of the other wonders can influence the Atlantis DOOM score, halting its increase for a turn, a one off reduction in score, or allowing House scores to be reduced through sacrifice (which does not change whether Atlantis sinks, but can boost your chance of a Moral victory).

House and Atlantis DOOM scores are public information.

DOOM is a collective action problem inspired by the Prisoner’s Dilemma. Actions that increase DOOM have the potential to benefit the player whose action has triggered the DOOM increase. If your House nobly eschews the use of actions that increase DOOM, you may well save Atlantis, but your rivals who do increase DOOM may outperform you in the game.

DOOM tokens are also a negative feedback loop. They are a way of giving a boost to a player who has not done well earlier in the game, by giving them an option to catch up in effective game actions. Negative feedback acts to stabilize and prolong the game (See chapter 18 games as Cybernetic Systems in Rules of Play by Salen and Zimmerman MIT Press 2004).

Getting into the gritty detail, DOOM is increased:

  • by the number rolled on a DOOM die (a d13)
  • by the cost of each Sorcery card purchased by a player
  • by the value of any Governance cards that DOOM is used to power up for City options
  • by the value of any Governance cards used to activate Sorcery cards
  • for roleplaying reasons (if Atlantis falls into anarchy, or the players are misbehaving)

Each map table will probably have at least one DOOM die rolled on it each game turn. So if we have five map tables and ten game turns, then that will make the Atlantis DOOM score increase by roughly 380. But a player with a DOOM die involved in multiple conflicts might roll the DOOM die four times in one turn. The DOOM ray technology upgrade also allows more DOOM dice to be rolled. So perhaps the upper range of DOOM from normal combat activities is closer to 1,000.

There are 18 sorcery cards. The cost to buy a complete set is 171 DOOM tokens. So if we have seven factions and all the cards are purchased, that is 1,197 DOOM. That is a large investment of DOOM tokens, when most teams will get perhaps 5-10 tokens a turn between all their players.

The cost of Governance cards used to power up cities and fuel sorcery card use is hard predict. The cards have a value range of 1-10, with an average of around 6.5. Higher value cards have usually (but not always) stronger effects. The limit here is going to be the number of DOOM tokens that can be used as fuel.

You get one DOOM token per conflict you lose, and each time you leave a Council meeting without any rewards. Player choice can see rewards distributed evenly, or hoarded by a few. So social inequality increases the chances of Atlantis sinking – I am okay with that as a design feature. The number of conflicts each turn is variable. Again player choice can lead to conflict free map tables. Or there could be a lot of eris on a table, and a player could get anywhere from 1-4 DOOM tokens.

Finger in air time, I expect most players to use DOOM tokens at least some of the time. The temptation to use DOOM will always be there. Maybe it will be 1,000 points of DOOM over the course of the game.

So, if I pick a secret DOOM threshold of 3,000, then its quite likely that Atlantis will sink. If the threshold is 2,000 it will probably sink quite early. If the DOOM threshold is 4,000 or more, then the chance of Atlantis sinking goes down.

So Sorcery cards are equivalent to nuclear weapons. You want some to act as a deterrent force against other players, but you may not actually want to use them in play. Given human emotions, once one House starts using Sorcery, the other factions may respond in kind. Another possibility is the kind of player who likes smashing sandcastles other players build, deliberately maximizing their DOOM generation (social mechanisms in the game might be able to deal with that – its up to the players to spot it happening and do something about it). A House that feels they are losing badly after a series of betrayals may also feel justified in dragging Atlantis down with them.

In my next post on The Colossus of Atlantis I will address aspects of emergent play in the game.


Conflict in Colossus of Atlantis

November 10, 2016

Let us start with two video clips. The first is the Talos scene from Jason and the Argonauts. It neatly demonstrates a Colossus versus Trireme battle. The second is from Oliver Stone’s Alexander movie, the main battle scene (not the best quality sadly). Which I think is a much better take on Hoplite warfare than in the movies Troy or 300.

Design goals

My design goals for conflict between the forces that players control in The Colossus of Atlantis are for conflict to be:

  • Simple
  • Fun
  • have scope for mastery
  • fit with the theme
  • Quick

I have tried to keep conflict relatively simple by:

  • making conflict automatic if two or more players send units to the same region
    • the only exception to this is if a player declares they are a coward (by saying “Phobos!”), and withdraws their forces back to their home city
    • cowardice has a potential cost – you can be fined by the courts, and if you don’t pay, your home city can be attacked (see below for an image of what that looks like)
  • keeping the maths to simple addition (no subtraction, multiplication or long division)
  • restricting the types of forces and the number of possible attacks to four

The maths is simple – add your strength ratings together for all your controlled units. Roll your dice, and add the two together. Highest player wins, both players lose in a tie.

I have two main ways in which conflict is fun. First, you get to roll dice, and you have a chance that each roll will explode (allowing you to keep rolling until the die stops exploding, and adding all the numbers together). Every now and then the Chaos dice will allow a player to pull off a lucky win against the odds. So there will always be some tension in a dice throw. Secondly, when you win you always get rewarded. You get some money, some victory points (VP), and some other useful game resources.


Course of Empire by Thomas Cole [Public Domain], via Wikimedia Commons

I believe there is scope for mastery of the game to be demonstrated. This arises from observation of the emergent play – if you can see the choices other players are making in how they allocate forces to regions, and react to them with skill, then you should win more conflicts (and earn more rewards). A cautious player may concentrate their forces on one or two areas. A bold player may split them between four areas, hoping to get lucky and find a region no one else is contesting for a cheap victory. Understanding how the dice can work out will also probably help with system mastery.

Exploding Dice

Because the Chaos die generates numbers from 1-4, and explodes on a four, its actually impossible to get a roll of four on the dice. Four +1-4 gets 5-8, and of a course a second four explodes again. You will roll from 1-3 75% of the time, and then 25% of the time you will roll 5+:

  • 75% of the time you will roll from 1-3
  • 18.75% of the time you will roll from 5-7
  • ~4.5% of the time you will roll 9-11
  • ~2% of the time you will roll 13+

One implication here, is that if you have a one bonus die advantage over your enemy, but they have four more combat strength, then you will have roughly a 25% chance of winning.

The DOOM die generates numbers from 1-13, but only explodes on a 13. So it exploding is a rare event – you might not see it happen in the game.

  • 92.3% of the time you will see a roll from 1-12
  • 7% of the time you will see a roll from 14-25
  • less than 1% of the time you will see a roll of 27+

Intuitively you would think that the d13 has a better chance of giving a victory against long odds, but its good to see it confirmed in the math.


I believe that a megagame conflict system trying to model Ancient Greek hoplite combat for a megagame should keep it simple, following a cultural style of combat where both parties agreed to fight at a particular place, the battle was fought quickly, and the defeated party conceded defeat. I thought about incorporating rules for skirmishers, cavalry, chariots and fortifications, but playtesting showed that this made it too complicated.

I added Triremes after the second playtest, as I redrew the maps, splitting them into one set of six land regions, and a second set of coastal regions, plus one gateway to the hollow earth. This reflects the importance of naval warfare in ancient Greece, and also fit with a change to how trading worked in the game – opening up possibilities for blockades and smuggling.


This relates to simple. A few things have been done to keep play fast. First, all players in a region are in conflict with each other. Diplomacy over forming temporary coalitions just takes up too much time in the fifteen minute game turn. With 13 potential conflicts to resolve, battles need to be done in under a minute. Second, players can resolve more than one conflict at a time. If two players are resolving a conflict in region one, three different players with a conflict in region 5 can be working through that. That is another reason why there will be lot of dice for each game table.

I also removed an early playtest option for spending DOOM tokens to adjust combat outcomes. This created an auction bidding mechanic, which just took too much time. Now you need the right Colossus upgrade (the DOOM Ray) and you spend the token before you roll the dice (a DOOM die).

Part of being quick is in constraining the choices the players can make. Most humans can only effectively evaluate three to five options. More than that and they start taking mental short cuts. So while I have 13 colony regions for players to squabble over, and they each offer slightly different rewards, there are only three types: land, coastal, and hollow earth. The fourth option, attack on home cities is constrained by law and politics. I also only allow four attacks maximum – and no splitting up of unit types. All your hoplites go to one region, and all your triremes go to one region (and it could be the same region).

Right, back to editing the game document!