Different Approaches to Micro-game Design

This past year (and longer, really), I’ve been exercising my design muscles by making really tiny games. I talked about why designing microgames is a good design exercise a while ago. The first one was Nantucket, which ended up being a few cards, and I discussed the process behind that a in the same article. Nantucket really started with the mechanics of New Bedford, and I adapted them to the smaller simpler format. This year, I had BoxScore as a stretch goal promo with the Bottom of the 9th Clubhouse expansion. And now you can get another game Rocky Road Dice Cream as a deluxe pledge for Rocky Road a la Mode by Joshua J. Mills from Green Couch Games. As I’ve continued working and developing my skills, I’ve observed that the approach I took with Nantucket is just one of several different ways to adapt a game. So today, I thought I would talk about those three approaches with 3 other games: Espresso, BoxScore, and Dice Cream.

Espresso – Bottom up

EspressoDiceTinyThe design for Espresso started with a concept for a business card game. A roll-and-move racing game except there is a role selection aspect to determine when everyone rolls and when you moves (i.e. everyone does it, but the player whose turn it is gets a bonus). The initial version was a race around the outside of a business card, with the rules in the middle. The actions were rolling, moving, swapping dice with another player, and simply increasing your die. This was too predictable. I needed another power, and went with flipping the die to the other side.

But I also realized I didn’t even need a track because players could simply keep score. The card could have the roles on it instead, turning into something more like worker placement. Moving turned into scoring. And increasing your die turned into simply taking an extra point. The number of points fluctuated between 1 and 2, but it turns out anything more than 1 is an automatic choice. The big breakthrough was varying player order, based on the action. Weaker actions would have first picks in the next round. Stronger actions (like scoring) went later. Bonuses for the active player varied a little, but I eventually settled on most of them simply increasing your die, rather than extra points. That helps players with lower rolls, without helping people with 6s, providing a little rubber banding.

But the game
was still a little too predictable. once you got a 1 or a 6, you would just swap or flip until it scores, taking away half of the “roll and move”. Every round was evaluated the same. The solution was to force the dice to change, by adding a mechanic where the dice decrease by 1 at the end of every round. You never start a round with a 6, so there is always a chance to improve your position. (And this makes the Flip action more beneficial to players with lower rolls, while mitigating the punishment to higher rolls.)

At this point, it was still almost a completely abstract game. I was considering adding special dice with powers, (and making it expandable to any die with a logo). I started thinking about Angry Dice, which does something similar, but then it struck me that this die decrease was like the blends degrading in Viva Java: The Dice Game (and, in turn, like Viva Java). And there were coffee bean dice! The coffee process matched up well with the actions. Roll to get beans, roast them to improve them, blend with other players, research (which is an sort of separate path) and finally brewing/serving the coffee. Those mappings led me to change the blend action, so that you and another player gain 1 to your dice instead of losing it. It made the action more desirable, and has great Positive Player Interaction. In fact, all the actions except research can potentially help someone else, which is something I always strive for in a game. And figuring out who to help and when is really the heart of the game.

But this is a case where the original mechanics came first, and I was able to map it onto an existing theme. And while I’m now working on translating it to a new theme of Sushi, where the fish slowly loses freshness, I can do that because I found a theme that really made the game I already had work.

Box Score – Top Down

BoxScoreTinyBottom of the 9th [Bot9]is a great game. The first version of the game was apparently small enough to be played in a car cup holder, and an early prototype board fit on a single card, so you could take it with you. But the extra cards, board, and tokens all make the game a little too big to fit in your pocket. I knew it was a game I could make small again. I looked at the two key elements of Bot9, which were the dice rolling and the stare-down phase. Take away the stare down and it’s all random. Take away the dice and the stare-down isn’t complex enough.

So I combined them into a single operation. You simply simultaneously choose a die number and reveal it, then evaluate a table to determine the outcome. A key to making something small like this work is perturbation. When you’re repeating a very simple action, you need to find a way to change the state of the game from round to round, or the decisions become repetitive. So I decided you couldn’t show the same number twice in a row.

I didn’t know what to do about tracking at bats. I considered trying to have players fill up the grid, so you simply accumulate bases, outs and strikes. That would self-balance but has problems and needs either a pen or too many tokens. I next considered requiring players to use 5 of the 6 options before re-using them. But that still requires a lot of tokens. So I decided to add a simple space to track balls and strikes, and simply stuck with not being able to choose the same number twice in a row.

Finding the right distribution for the results table was a small challenge. It seemed obvious that, like Bot9, if the batter matches the pitcher, it’s good. But I also tried to figure out what the die represents. I could have gone with where you actually hit the ball (so farther is better, and not down baselines. But that’s really hard to create a good strategy for. And so I returned to Bot9 and used pitch speed. That means that the batter should probably be a little faster than the pitcher, but too fast or too slow is bad. I put together a grid with strikes, outs, and hits. Balls ended up being too much. I assigned values to each, with positive being good for the batter. Outs were -1/3, strikes are -1/9, singles are 1/4, doubles are 1/2, and a home run is 1. (It’s not exactly right because you play to first run or 3 outs, which hides some of the details, but with testing, it seems close enough).

Then I started arranging things so the patterns made sense. Hitting the ball harder is better for more runs, but risks an out more than a soft swing. 6,1 and 1,6 are great for the pitcher, so you need a home run at 1,1 to even things out. Fill in with strikes, which can actually be good for the batter by giving them an extra chance. To make the final balance work out to zero (i.e. balanced between batter and pitcher), I turned one corner into the Double Play. Thematically, this all works out because strikeouts are hard. Your mid pitches and swings are safe but slow.

In Box Score, I took the mechanisms (and components) that came in Bot9 and tried to represent them as closely as possible while stripping as much away as possible. You have the dice rolling to give a wide range of outcomes, and the feel of the dice game. But it’s completely non-random to capture the actual feel of out-thinking the opponent. BoxScore needs only the components from Bot9, in addition to the card. And as a bonus, because there’s no actual rolling, you can play without a table, perhaps while in line at the ball game.

Dice Cream – Hybrid

DiceCreamTinyDice Cream started in Salt Lake City. Joshua J Mills had a game he had been working on for a while called Rocky Road a la Mode, and asked me to take a look at it at SaltCon. [I’ll refer to the base Rocky Road a la Mode as Rocky Road, and my game as Dice Cream, even though they’re both under the Rocky Road Brand.] I really loved the game, just like the other designs he’s shown me. I had already shown him a few single-card games I had been working on, and Josh said he wanted me to see if I could make a single-card version of Rocky Road. I think I may have scared him a bit because my face went blank, and I made some strange facial expressions and hand movements while silently mumbling to myself for a few minutes. I sketched out something on a scrap paper nearby. But I looked at it and realized it wasn’t going to do what I wanted it to. Perhaps Josh remembers it differently. By the end of the flight home a few days later, I had the rudimentary workings of the game.

Part of the game came from Rocky Road. Mechanically, it had to include moving around a time-track, and an engine-building aspect. Thematically, it kept collecting and serving ice cream, and I knew it would re-use moving forward for each resource collected, and when serving. But I didn’t know how to control how much to serve, or track how much you had. There weren’t many components besides a few stacking tokens and the cards. I needed something more. So I had to approach the design from another angle.

I realized I had just enough room on the card sides for a supply track running 1-6, using half inch tokens. Rocky Road is set collection, so clearly, you needed to collect the right sets of flavors, moving tokens up and down the track to count your current supply. [I had abstracted away the theme for the time being, except to the extent that there were definitely 3 types of resource “flavors” like Rocky Road.] I was loosely inspired here by the way each player tracks multiple resources in Scott Almes’ Harbour, as it’s a really efficient way of handling resources without many components.

Dice are a really handy way of both tracking information and adding randomness to a game without a lot of components. But I didn’t want to add too many dice, so I used 3, matching the 3 types of treat in Rocky Road. Roll the dice to see how many to deliver. But instead of just serving a mix of just a couple, you had to match the number on the die. And you could choose how many different ones to serve at once. When you serve, re-roll the die to change the demand. And if you score all 3, you get a bonus point. As a change from Rocky Road the more you score, the farther you move, which then gives the other player more turns to catch up. Also, the addition of dice lead to it being called Dice Cream. A good pun is like 50% of the initial design. And of course, if it’s ice cream, the 3 flavors are chocolate, vanilla, and strawberry to make up Neapolitan.

If you’re only using one side of something in a microgame, there’s room to add something. That’s how I got an engine building element. You would be able to flip your resource token to gain a permanent bonus. But it has some challenges. If you simply reduce the resource cost by 1, you can potentially keep scoring the same type over and over, which is no fun. And it’s too complicated and powerful if it’s every time you collect that resource. Instead, you gain the resource each time you become the last player in line. [Since you can take multiple actions in a row, the concept of a “turn” is a bit loose here.] There’s a trade-off of scoring. You can either score, or gain the bonus resource, not both, which diverges from Rocky Road.

The game was a bit static because high numbers wouldn’t ever get delivered. Josh was just starting to add some extra abilities to Rocky Road when I played it. I didn’t have much room, but I added fixed spaces on the board that forced you to re-roll one of the flavor dice. It was initially a specific flavor, which was still too static while you drove around the road, so I changed it to re-rolling any one die. This gave more action, but you almost always re-rolled the highest die. After a few plays at Unpub, I wanted something that would speed up the game, too, since it could take a while if you keep re-rolling high, Finally, I changed it so that when you pass the space, you must decrease a die by 1. This potentially helps both players, which fits my design philosophy. The final addition was a comment from Josh Mills. He wanted me to make the dice rolls more fun, and give players something to hope for. So in collaboration with him, I made it a bonus resource when you roll the same as an un-rolled die. This also makes smaller deliveries more powerful, balancing the game a little better.

It took me several iterations to work out the setup. The initial versions used half of the back of the card for setup. If your setup is more complicated to explain than the rest of the game, you’re probably doing something wrong.  I finally decided to let players choose what resources to start with and that eliminated about 75% of the setup. You don’t need to overthink the easy stuff, especially when making a microgame. Sometimes the easiest method is best simply because it’s easiest.

With Dice Cream I wasn’t just taking an adaptation, but it wasn’t entirely new, either. There are some elements that carried over, new mechanics inspired by the original, and new mechanics developed to help the gameplay. And the final result is something that still captures the feel of the original game without feeling like a re-hash or a lazy copy.

Concluding Thoughts

A microgame shouldn’t just be a physically small copy of a big game. It shouldn’t be a limited version that cuts out content. It should stand on its own in terms of gameplay. Where microgames can lean on larger games is the context. Would BoxScore make as much sense isolated from the Bottom of the 9th world? Would Dice Cream? Maybe. But the real strength is as a supplement, reinforcing the larger game while also leaning on it to enhance the experience.

If you’re interested in BoxScore, look for me, Dice Hate Me Games, and Greater than Games at conventions this fall after Bottom of the 9th Clubhouse arrives. If you’re interested in Dice Cream, head over to the Rocky Road a la Mode Kickstarter and select the Deluxe pledge level.


  1. #1 by John Kemp on June 24, 2016 - 5:02 pm

    Great article! Thanks!

  1. Nanodaptions – The Indie Game Report (TIGR)
  2. Nanodaptions – The Inquisitive Meeple
  3. Milkman Mills Shares the Scoop on Rocky Road a la Mode – The Inquisitive Meeple
  4. Nanodaptions – The Inquisitive Meeple

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: