Which Came First?

... the theme or the mechanic?

Eurogames are often accused of having paper-thin veneer themes. They tend to have fairly abstract mechanics with a basis in some form of mathematics with a theme on top. Examples of some game themes are:

• Puerto Rico - Colonising the New World,
• Tikal - Exploring the jungle ruins at Tikal,
• Carcassonne - Developing the land arond the French city of Carcassonne,
• Caylus - Building a castle for the King.

I'm not really interested in why games are themed - I'm guessing it's to draw the players in and also as a marketing ploy. What I am interested in is the development of the game. Did the designer start with a theme and then choose mechanics to fit? Did they start with some abstract mechanics and add a theme on top? Did the publisher add the theme for marketing purposes? If either of the latter two options, did the designer redesign the game at all to better fit the theme?

I've designed one game and started designing three others. I've tried both the first two methods. My first game Border Reivers started with the mechanics, the other three all started with a theme. I started developing Border Reivers after a painfully long game of Mighty Empires with Tim and Dunk (and that wasn't even fighting the battles with miniatures). I really liked the style of game (medieval empire-building), but after 36 hours I was losing with Tim and Dunk clearly ahead of me (Tim winning by a nose) and the game wasn't going anywhere. We decided to give up after one more turn, and then something totally random happened almost wiping out Tim ( a dragon attack or somesuch). It was ridiculous, after such a long game for the playing field to completely change like that. So I thought I'd make a similar type of game that was less random and much shorter. I was aiming for under an hour, but Border Reivers can stretch to a bit longer than that with three players - like many three player games when the two losing players gang up on the winning player. Once then rules settled down after a couple of years, I started to look for a theme and chose the Border Reivers as they were a local, historical phenomenon and they fit the mechanics fairly well (especially the raiding card). I'll be honest, it's not the strongest themed game.

My next two attempts were theme driven. A Samurai duelling card game and a mission-based game inspired by the Firefly sci-fi TV series (it's great - I highly recommend it). The mechanics of both of these games changed over time as I tried to get the game to mirror the theme. Eventually I shelved both of them.

My latest game Codename: Dollyo is also theme driven, inspired by a trilogy of books that I love. It's a game of politics and intrigue in a setting similar to feudal Japan or Korea. Over the last year or two it's mechanics have completely changed on a couple of occasions as I discover new ideas. Each time I try to pick mechanics that give a fun gaming experience while remaining true to my subject matter.

Which works best? Judging by my experience, designing the mechanics first and then giving it a thin coat of theme means you end up with a finished game. Starting with a theme and trying to pick and chose mechanics to fit the theme ends up with lots of half-finshed games. I wonder which method other games designers use? It's a question I'll have to ask any I bump into at Essen this year.