Magnificent spreadsheets

There was a bit of a comparative discussion on Twitter among a few game designers about the use of spreadsheets. For my part, I find them useful in maintaining balance in a game's economy, in the relative values of different components of the game. In "Magnificent Marvels," Keith and I recognized the need to be sure that the different components with widely varied point values would need appropriately balanced building costs, and we put together a spreadsheet to try to manage that.

The core resource for player decisions is actions, which can be used to obtain money, more workers (by spending money), raw materials (by committing workers), and refined materials (by converting raw materials). So I put together the original spreadsheet of marvel components with a calculation on how many actions it would take to build each component, and then a ratio of points per action. The metric is not a perfect reflection of the cost-benefit value of each component, but it served as a helpful starting point. Very quickly, we got the game up and running based on our initial spreadsheet calculations.

The Achilles heel of our design method came in our modifications to the game flow based on playtest results. One of the things we decided we needed to accelerate was refining materials, so we made the converters immediately available and more efficient in their conversions. What we didn't realize at the time was that we'd distorted the relationship between cost and point value for certain components that required several of the same refined material to build. Whereas they had originally taken perhaps six actions to assemble before (and therefore provided a high point value), the accelerating refining and other changes meant that really only three actions or so were now necessary to build the same component. It wasn't until one of our playtesters specifically pointed out the ease with which he'd built that particular component that we realized our spreadsheet hadn't kept up with our rules structure.

So the lesson learned was that it is important as the game responds to playtesting results that the models used to manage the economy of the game respond along with them. We'd fallen into a classic trap of thinking all our original assumptions were still valid. Fortunately, playtesting serves to shake out those kinds of imbalances early on. That's what it's for, after all.