In this (rewind) edition of Designing Meeples (from March 2015), we interview famed tabletop designer, Matt Leacock shares some advice and tips on designing a good co-op game.

Please note this article was originally published on The Inquisitive Meeple on March 6, 2015. This is the whole article, though some of the layout has been changed and references to Thunderbirds Kickstarter has been erased. 

Matt Leacock is the designer of such cooperative play games as Pandemic and Pandemic: The Cure published by Z-Man GamesForbidden Island and Forbidden Desert published by Gamewright Games.

For those interested, we have in the past interviewed Matt for Pandemic: The Cure. The interview entitled “Matt Leacock Shares the Cure” can be found here.

Matt, you didn’t start out designing co-op games however, with the exception of a couple of games, your main focus in designing is co-ops. What draws you to designing them and what makes you good at designing them? 

Matt: I love designing cooperative games because they draw people together and get them talking and interacting more than many competitive games do. I also enjoy designing games that I can play with my family, and my wife and youngest daughter don’t really like games with heavy competition or negotiation.

I attribute a good part of my success to the hours I’ve spent simply watching people play the games — both to see what engages them and to understand what trips them up while playing.

I image that designing a co-op game would be different that designing a competitive game. What are some specific challenges you face when designing co-ops that you don’t face when designing competitive games?

Matt: When designing a cooperative game, you need to generate some sort of opposing force for the players to overcome. Finding a novel way to create a challenging opponent, given the limitations of board games (paper, cardboard, wooden bits, dice, and so on) can be difficult. With a competitive game, the players provide the opposing intelligence.

The evolution of the treatment center and event cards in Pandemic: The Cure

When you are designing a co-op game, do you worry about mitigating the “alpha player” problem at all or is that something you do not worry about since that is more reflective on who makes up each playgroup (personality styles)

Matt: First, it’s important to differentiate the players and give them something unique that they can contribute so the players don’t feel like they’re freely interchangeable. Then, you need to find ways to prevent one player from running the show. Realtime cooperative games (like Space Alert and Escape) do this by overwhelming the players with so much information that each player has to work independently. Other games (like Hanabi) accomplish this by limiting information or communication. I prefer open-ended games with a lot of discussion and creative problem-solving so I’ve been experimenting with other methods.

Pandemic: The Cure

For example, in Pandemic: The Cure, each player has direct control over their own dice which gives them more autonomy. Since it’s rude to reach over and roll someone else’s dice, players feel like they have much more independent control over what they’re doing on a turn. Another approach is to give players many different options with more subtle strategies. In Forbidden Desert , for example, it’s hard to point to any single optimal play on any given turn (since there are so many things you can do). This makes it easier to have multiple, independently valid strategies on the table to choose from on your turn.

How important is it to have varying difficulty levels when making a co-op game and what are some good tips in creating such variety?

Matt: I include multiple difficulty levels, so each game’s challenge can ramp up to meet the increasing skill of the players. If the games didn’t do this, the players would consider them “solved” at some point and move on. I try to make this a matter of changing a simple variable (such as the number of epidemic cards in Pandemic) to ease setup and to keep the game’s model simple for the players to understand. That is, I’d like the game to be challenging to play well, not challenging to comprehend. So, that said, my tip here would be to examine your game to determine what underlying variable could be tweaked to increase the challenge of the game without changing the way it’s played.

How do you ensure balance between giving the players the right amount of actions to take during their turn and feeling like they have at least accomplished something, while still making them feel like they need to do more on their next turn?

Matt: The first consideration for me is the audience’s ability to keep track of what they’ve accomplished on a given turn. Then I consider how much strategic depth the game should offer. To that end, knowing your audience is important. When I designed Forbidden Island, I reduced the number of actions per turn to three (down from the four you take in Pandemic) because it’s easier for kids to keep track of three things than four. I designed Forbidden Desert  with four actions per turn to help make it a deeper, more challenging sequel. Thunderbirds offers players many free things to do on their turn and is meant to appeal to kids, so it has three actions per turn. With four actions, it was too easy to lose track of how many you had spent.

If you offer too many actions per turn, you also run the risk of downtime due to analysis. Tikal, for example, had a tendency to drag out a bit since it offered ten action-points to spend each turn.

It’s a sloppy and imprecise process though—you try stuff out and see what works and iterate until it feels right for your intended audience.

I imagine you also have to consider things like making sure one individual player’s choice, while important, won’t ruin the whole game for the group. How important is it to achieve balance on individual actions with how it affects the overall group as whole?

Matt: I don’t think I’ve ever consciously modified a design to ensure that a given player’s actions wouldn’t upset the overall game. I’d rather err in the other direction, making each decision really important. That said, you want the game’s system to be robust enough so players can make small mistakes without the fear of immediate, crushing defeat. That way they can learn during a play session and not just during the game’s post mortem.

With your co-op games, you introduce different player abilities, via roles. What are some good general (standard) role actions to add to co-op games?

Matt: I’m not a big fan of over-generalizing play patterns as I think this leads to derivative games where designers take some ingredients from column A and mix them with ingredients from column B in order to design a new game. If you’re going to create individual differences between the players of the game, draw from what makes your game different first. One piece of advice though: if you are going to put in different role powers, go big. Don’t put in subtle tweaks, go as far as you can without breaking the game and iterate by halving or doubling the powers. It frustrates me when I play a game that purports to have different roles, but they have little to no impact on play.

Different roles certainly need a lot of testing not only to make sure they are not overpowered, but also to make sure they mix well with other roles. What is your process at finding role balance? Do you try out everything you can think of or do you generally start with toning them down (under powering them) and then making them stronger as you get more plays?

Matt: The best way to find balance is to play, observe, tweak, and iterate. I start by making the role abilities as strong as possible, then amp them up or tone them down as needed. Another tip: find a role that you want to serve as a benchmark. For example, in Pandemic, I can ask myself, “is this role really as good as the Medic?”

Would you say the crux of co-op game design is around getting the difficulty of the AI, or system you fight against, to the right difficulty?  Do you have any tips for achieving the balance?

Matt: Achieving balance, while time consuming, is a fairly known process. You play, make lots of notes, make changes, and iterate until you’ve filed off all the pointy bits. That’s an important part of co-op game development.

It’s much harder to come up with a novel system that players genuinely get excited about playing—one that’s both new and fun.

Pandemic: The Cure prototype
Finished product of Pandemic: The Cure

Do you have any suggestions about how to get the proper pacing of how the AI works in a co-op game?

Matt: One technique when designing the pacing and arc of a game is to try to imagine your game as a movie. Tokaido, for example, plays out like a zen-like road movie, Pandemic is more of a sci-fi/action thriller, and the Forbidden games play out like an adventure serial. Once you have that goal, you need to design the mechanisms that dole out experiences like rising tension, periods of rest and recovery, a climax, and denouement.

Try to give your players a bit of time to get their footing at the beginning of the game. If they’re immediately thrown into crisis while they’re still finding their bearings, their anxiety may shoot up, they’ll get overwhelmed, and may want to bail. That said, it’s important the players have something clear and immediate to accomplish on their first turn or they’ll have no idea where to start, worry that they’ll look unintelligent in front of their friends … and want to bail. Once you have established a few beats of the game though, you can really draw them in with bursts of crisis and relief. I’m new to these techniques though — I think it’d be fascinating to study the construction of screenplays to see how this is done across genres to see if there are approaches to storytelling that remain untapped by board games.

What are some tips in making sure that you have an effective AI, but at the same time ensuring it doesn’t become an effective hindrance to upkeep? 

Matt: Your best bet is to start small and try to find a simple core mechanism that generates challenging situations without a lot of inherent complexity. If you start with a complex system, it’s harder to balance, and will likely require more sub-systems, bookkeeping, and rules overhead for the players.

Forbidden Island prototype

With only an AI to complete with, how do you make sure the game can’t be “solved” easily?

Matt: So far, my approach has been to come up with AI that generates problems that are complex (that is, the decision tree is too deep to fully plot out) paired with a some degree of uncertainty around what will happen next. Whatever your approach though, you need to test it with your target audience to ensure it scratches their itch and holds up to repeated play.

Note that some games are solvable, but still make great kids’ games. For example, Susan McKinley Ross’s game, Hoot Owl Hoot, works well, yet it would be fairly easy to write an algorithm for optimal play. (The game works since most 4 year olds don’t yet have that skill.)

What is a good win/lost ration for a co-op game? How many times should players lose to keep the tension high and build excitement?

Matt: This really depends on the experience level of the players. If I throw novice players (that have never played a cooperative game) at the standard level of difficulty for a game, I expect them to lose 100% of the time the first time they play. If they try the intro game, I’d like them to feel like they had a chance of winning if they had only played better—so they play again. (And if they played exceptionally well during their first session and they win, great!)

Conversely, for expert players playing on the most difficult level, I want there to be a good chance that they will lose. If they do, I want them to be saying to each other, “we could have won if only we had accounted for X”—so they play again. There’s really no magic percentage here, it’s really about communicating expectations of the various difficulty levels and ensuring there’s enough room for novices to grow and expert players to be challenged continually.

Do co-op games require more playtesting than a competitive game, since varying difficulty levels and player counts can change the dynamic of the game?

Matt: The short answer is I don’t know. I haven’t designed enough competitive games to offer a meaningful comparison. All of my games (competitive or cooperative) see well over a hundred plays before the design is remotely considered “done.” I suspect the number of plays has more to do with the complexity of a given system than whether it’s cooperative or not.

Do you go through phases of playtesting to purposely see if you can break your co-op system? And what are some good suggestions about how to go about doing such testing, if it is important?

Matt: I stress-test each design during playtesting, encouraging playtesters to break the game through extreme play methods—taking one strategy and trying to play it as aggressively as possible, for example, to see if the system holds up. It’s also important to see how a variety of groups end up playing the game over multiple sessions. If a dominant strategy begins to emerge, I’ll go back in and knock it down a bit and shore up other paths to victory.

When testing co-op games, what are some good questions to ask your playtesters?

Matt: When I blind test, I ask that the participants video-record themselves. I prefer to include questions that the players can answer verbally, into the camera, at the end of the game as I can get much more nuanced feedback that way. Questions are usually simple and might include:

  •  Were there any points where you felt confused?
  •  Describe the game’s challenge. Did it feel too easy, too hard, or about right?
  •  Any comments on the length of the game?
  •  What was your favorite thing about the game?
  •  If you could change one thing about the game, what would it be?
  •  Any other comments?

I don’t ask if the players enjoyed their experience. It’s usually evident on the tape or offered voluntarily (if they enjoyed themselves).

Do you have any tips/suggestions about pitching a co-op game to a publishing company?

 Matt: Do your homework. Research publishers online to find out what kinds of games they’ve published to date and what they may be interested in publishing in the future. Find out what their submission policies are. If you have a prototype that you’d like to pitch, I recommend setting up an appointment with the company—in advance—at a conference or trade show so you can show them in person. Don’t send unsolicited prototypes. Don’t talk about how your game will be the “next Monopoly.” Refine your pitch, focusing on what makes your game different and memorable.

As we come to an end with our discussion about designing co-op games, do you have any general advice or tips we haven’t covered for aspiring co-op designers out there for make a solid co-op game/experience?

 Matt: We’re all just making this up as we go along. There are undoubtedly many new ideas for cooperative games that have never been tried. Experiment with crazy ideas when you’re designing.

As we finally wrap this up, do you have anything that you would like to add? 

Matt: No, I think we’re good.

Thank you, Matt, for taking the time out and agreeing to do this interview.


For those that like to read past Designing Meeples like this one, with Matt, you can do so by going to this section of the old The Inquisitive Meeple WordPress site. Get design advice from Jonathan Gilmour, David Short, Scott Almes, Kane Klenko, Daniel Solis, Gil Hova, and more.