Love’s Labour's Lost

NOTE: This essay was originally posted as the fourth Designer Diary on the Molly House Board Game Geek page. This essay refers to an older version of the game. 

In the last entry I concluded by mentioning how a pushing on one subsystem led to the complete reworking of the design. That subsystem was the relationship system. Today, I’m going to write a bit about that system’s role in the design and how its deconstruction gave us the tools we needed to finish the game.

At some point in the design of Molly House, we made it a priority to feature a kind of dating game. I am consistently disappointed that more games don’t explore interpersonal relationships, and was very supportive of the idea. Dating games of all sorts present really interesting design problems and, in a certain light, are not so different from the kinds of alignment and portfolio problems that I’ve explored elsewhere. In contrast to the loyalty system of Pamir or John Company’s various share holding systems, a dating game is anchored by an intimate, personal scale. It seemed like fruitful territory to explore.

In the original cut of Molly House that Drew and I played, the players could form direct relationships with one another through special festivities. This was later expanded to include lots of different kinds of relationships. Eventually, these relationships were dramatically simplified and pretty soon found themselves reduced to an exercise in area-majority dynamics. Essentially, whoever had the biggest stack of cubes on the Molly wins.

Strictly speaking, this worked just fine. And, there was still something compelling about the system, even in this abstract form. At the same time, it didn’t seem right. The area majority mechanics were too simple and it flattened the personality differences between the mollies.

The system also relied heavily on a longstanding problem area within the design: resource generation. Early on we had essentially three resources: items, cash, and actions themselves. Players could develop relationships with mollies by spending some combination of those resources. But, these resources didn’t really allow for players to develop unique player positions. This created a very boring gameplay loop where players would take actions to generate money, then actions to turn that money into items, then spend those items at festivities to develop relationships with the mollies that happened to be in attendance.

We spent a lot of time trying to make this loop interesting. At one point you could freely spend time with mollies, allowing you to convert actions to relationship cubes. Ugh. I feel gross just typing that sentence. The more we worked on relationships the more we realized that the game’s core economy simply wasn’t robust enough to support a relationship system.

And, the essential problem was that the player positions were so flexible anyone could contribute to anything. If a high value molly came into the game, players would rush to “bid” on it. There was no friction, no tension to the romance. As an aside, it’s worth pointing out that this dynamic had emerged out of all of the efforts to keep the design simple. When you cut complexity, something is almost always lost, and in this case we were “spending” the game’s complexity budget on things like an elaborate constable raid system, festivities, and other elements. Relationships were therefore becoming a bit of a side-show.

We fixed this by doing two things. First, we added spatial friction. The game got a randomized map which folks had to move around. This meant that if a molly showed up in Covent Garden and you were at Lincoln’s Inn, you might not be able to get there in time. Second, we decided to create a new universal currency using suited tokens.

Suited tokens gave us so many new options. We imagined there would be one or more bags of tokens which would reflect four classes of desires following arch types loosely inspired by traditional playing cards and the tarot. Now, instead of just gaining cash and items, players could gain tokens which they could then spend in a variety of ways, including courting the non-player mollies.

At first we decided to link each non-player molly to one suit. This would naturally create differentiation of player position as the random draws of the tokens would lead players to invest in different mollies. However, we soon realized that this was removing an awful lot of strategy from the game. You tended to invest in whatever mollies solely based on what resources you had. Though spatial concerns mattered, they were not enough to overcome this pressure. By introducing some texture into the game, we had essentially sapped the game of some of its strategy.

Because the game felt close to its complexity budget, I felt it was important to come up with a solution that didn’t add too many rules to the core design. In such situations, my first instinct was to go to a spatial element. Here I’ll admit I was inspired by the share-pricing dynamics of the game Shark, which used a simple set of spatial behaviors and valuations to determine share price. The result is a chaotic game that feels far more like a dynamic market than something like Acquire, but doesn’t require a stack of event cards like something like Bear Raid or any number of more modern business games.

The basic idea here is that every non-player molly would have a personality board. This would be a network with each node being associated with one of the game’s suits. Certain spaces on the edge of the web would be entry points. If you played a matching token you could put your relationship cube in that space. From there though you had to build your relationship following the adjacency presented by the network. Then, at the end of each week, relationships were scored based on which player had the most cubes on each molly.

This had a ton of interesting dynamics. By making the placement rules slightly more complex than the standard area majority system, it allowed us to keep scoring simple. It also gave us the ability to quickly make tons of different network maps (“personality types”), which gave the mollies all a different feel. They didn’t quite behave like living, breathing people but they were one step further away from a safety deposit box.

Still, there were problems. We had to spend a lot of time dialing in the right level of dynamism in the system. Should cubes stick around from one week to another or be cleared? Could players attack each other’s cubes? How many suits should be on each Molly? How should this system tie into festivities? There is a proper diary entry that could be written about the process of answering each of these questions. And, in normal circumstances I’d be happy to write those entries. But, the whole relationship system ended up being dropped because of that last question.

As we worked on the relationship system, we were continually presented with the problem of what role that system had in the broader game. Early in the design, we had two victory points: communal joy and personal joy. Gradually communal joy morphed into something like house reputation. The idea was that festivities would offer you esteem (reputation and prestige within the community) and cruising and developing relationships with mollies would give you personal joy. But the more we pushed at the difference between these things the wobblier they seemed. Slowly but surely, just one victory currency emerged. This was primarily because of how the actions chained into one another. For instance, by doing well in a festivity players got the resources they needed to develop relationships and score personal joy. We tried various ways to delink this connection, but within the game’s larger mechanical/thematic framework it felt very wrong.

Eventually this boiled down to a pretty simple loop. Players would work to perform festivities. These festivities would give players experience tokens (which had the suits) and they would then invest these tokens in the various mollies which would score joy at the end of each week in area majority fashion.

We spent a lot of time trying to get the balance right between these concerns. And we did eventually get to a place where things were balanced. But, this balance didn’t bring the game into focus. We started to get the feeling that there was something wrong at a more fundamental level. While being central to the design, festivities were not half as responsive or expressive as the relationship system. But, for all of its expressive range, the relationship system still felt primarily guided by random draws and didn’t have the surprise or texture that it needed. The whole thing felt very flat.

Then, in basically two moments (separated by a couple months) the design was turned on its head. It started by realizing that the various “experience tokens” should be linked to the constable system. Originally the constables would mill about the board and look for the players who had the most cash or items or whatnot. But this never quite felt robust enough to support the various player positions and to make folks truly worry. Months after we introduced the experience tokens, we realized that the constables should be suited according to those tokens. It was so obvious! If the constable of cups was triggered, then players with those experience tokens would be in trouble. But just counting experience tokens didn’t provide the range needed—generally players were only holding 0-2 of a particular type, which meant there would be a lot of unsatisfactory ties. We could change this if we gave every token a guilt value of say 1 to 9.

Of course, this would work better if experience tokens weren’t tokens at all. They should obviously be cards. Then, instead of using dice to determine when the constable checks happened, we could just mix a few constable cards into the deck and have their checks triggered when they were drawn. Instead of rolling “heat dice” each turn as you moved around, you would just draw cards into your hand.

This had the follow on effect of linking the suited resources to risk, which meant that players were now free to develop relationships with mollies outside of the context of festivities. This was a big problem, because it meant that folks could just ignore the molly house. After some months of tinkering, we realized we could get around this problem by using a card-flow system. Instead of using modifiers or tasks to judge festivities, we created a system where players would contribute cards to a goal (e.g. “let’s get 20+ value of cups on the table). If the festivity was successful, those contributing cards would then graduate to become experiences and therefore be used in following turns to place bonds on the different mollies. Eventually we realized that these communal goals could mirror the logic of poker hands. One type of party might involve generating a six card run. Another might require a 5 card flush. Festivities were, for the first time, playful in their own right.

This system held out for several months. When we visited Jo at UKGE earlier this summer, it was largely in place. Then, later that summer, we expanded our playtesting efforts to include new testers. In these first tests a critical realization was made. The design was simply too busy. Folks loved the new festivity system but didn’t see why they were being forced to spend so much of their time playing a goofy network area majority game. They wanted more parties!

Initially I reacted very strongly against this impulse. We had worked so hard building out the relationship system! The nodes and links of each personality were so neat! I didn't want to abandon its mechanical and even aesthetic potential. And yet, with a few days' reflection, I could see they were totally right. The center of Molly House should, of course, be the molly house. So, one long Saturday we stripped the relationship game fully out of the design.

What then, should be done with the mollies? Almost in an instant, the answer came. They were the parties. The previous festivities had been triggered by any number of odd systems—festivity cards, task cards, end-of-turn goals, and whatnot. But, if we just made the mollies hold the festivity goals, then the relationship system could still exist in the game in an implicit way. Here were the mollies, milling about on the board. Players could invite them into their hand, bring them to the molly house, and play them to help throw their parties. They could run the parties, establishing the goals, or could attend the festivities like any other card. They were now at the very heart of the game. Molly House had found its mechanical footing.

I'm hoping to publish the last of this series early next week before the campaign launches. In that entry I'll take everyone through the current design from top to bottom, so you can see where this all ended up.