Finding The Minimum Viable Game

Conventional wisdom dictates that it’s best to start with the most refined, precise possible example of your game before adding extra stuff. There’s a logic to this: Your extras depend on the core game and altering the core game will often mean discarding or at least mutilating them. If you design a really cool movement special ability before you’ve made sure your system for movement does the job then you’ll put yourself in a position of either deleting something that’s working great or having bad decisions at a lower level locked in.

To counter this you build the minimum playable version of your game, test and refine that until your core is great, then add things to the working core.

A lovely idea in principle, but how do you decide what the minimal viable version of your game is?

Let’s pretend you are the designer of Dominion (Unless you are Donald X in which case you don’t have to pretend) and have decided to work on the MVG before designing all of those fiddly special actions. What exactly is that?

Is it a game with only copper and estates? That’s enough to have a choice of what to buy every round and shows off the core tension between “I need points to win but points make it harder to buy stuff”

Or does it need to contain gold too? So you get that effect in which adding a lower value treasure is helpful early on but its presence will become detrimental later?

Or does it need at least one action card because they’re a fundamental part of the game and also running out action card piles is part of the game over condition.

Or does it need at least two action cards because the limit on how many can be played is part of the game and having one that allows +action and one that doesn’t is part of the core game?

The minimum viable game is a non-trivial problem. If you tried to refine the game from just having a copper pile and estate pile you’d either find you couldn’t make your supposed core fun to play no matter what or copper and estates would become complex in ways that mutated the game beyond all recognition.

On the other hand you could just as easily justify to yourself that you have to include everything for one reason or another which defeats the point of the exercise.

Personally I reckon it’d be somewhere around having the treasure and vp cards and 3-4 action cards that are intended as archetypical examples of common things players will do (Say Village, Smithy, Remodel, Militia and Woodcutter). But there’s no hard and fast way to say “Yeah, that’s the right amount of fancy actions for the game to be enjoyable so I can test and refine on that without getting bogged down in making others that might be made redundant as I change things”. It’s obvious that having none leaves you with something that’s not really much of a game, but for any particular list you could almost always justify removing some member from it.

So given those limitations how do you find the MVG for your own design?

Coming back to your games interesting decisions is never a bad move. If a game is a series of interesting choices, then the minimal viable game is the parts of the game sufficient to produce its most interesting choice.

Start by asking “At what moment will a player of this game go ‘I want to X and I want to Y, but if X then not Y but if Y then not X” Then once that’s identified look at what’s needed for that choice to work.

That’s not limited just to creating the choice, but also in making it have payoff. Sure the player has to have (at least) two options, but there has to be enough going on that it’s not trivially obvious which they should pick. There also needs to be a result that’ll let a player look back and see if they made that choice poorly or well.

Sometimes this requires things that are like the icing on the cake. It could be that the core of your game doesn’t work unless everyone’s got a special power or there are cards that create exceptions or whatever. That’s fine, your minimum game doesn’t have to be about “Destroy everything that looks like a sprinkle”, but it should manage “Use the minimum number of sprinkles”. If two different powers are enough to test that the core of your game is up to holding the weight of everything else you want to put on it then you don’t even need to design enough to get the game going at its highest player count.

Your minimum viable game might not be that minimal, in order to be viable, but it’s worth knowing what it is. Having a clear idea of what your core is, why it works, what’s critical to it and what’s you expanding on it will help with every other thing.

What does the minimal version of the last game you played look like?


I had an interesting experience working on Genesis the last couple of weeks. It involved spending money on things we turned out not to need and making changes then reversing them and did something to make me think of icons differently. I thought I’d share it in the hopes of preventing other people from making similar mistakes.

In this game the players are gods, each god has several purviews (Like mastery over water, death and animals or whatever) and each purview has a champion. Here’s Ripple, the champion of water, as her card appeared at the start of that time:

Obviously this is prototype time. The final thing won’t look like this and that’s not our art. It’s getting nowhere near a finished product. Just something to fill the space till we’ve paid an artist to draw something specially for the game.

I want to draw your attention to the icon at the bottom left. That’s a free icon from  which is a fantastic resource for any designer in the process of prototyping a game. The icons can even be freely used on a finished commercial product if you like, though generally I try not to.

Anyway do you see how its flecks are escaping the boundaries of the icon? That’s because it’s an icon that’s designed to go onto a square that I’ve put onto a circle. A lot of the games icons are like that, they annoy me every time I look at them.

It also looks bigger than the circle next to it. That’s an optical illusion, you can print the card and cut it in half and match up the circles to prove it to yourself if you like 😉 That they’re different styles makes them look different sizes which is all kinds of messy.

So I asked a designer to get an icon together for me that would fit the imagery of the rest of the card, looking to improve it with some more consistency. This is where I ended up:

Well not quite here – this time the left circle actually is smaller, but that’s my mistake. I don’t seem to have an image from after it was fixed but fixed it was and I got it in front of playtesters.

The feedback was immediate “The old version was better” “We actively liked it that the icon overspilled its bounds”. I figured maybe people liked what they were familiar with so tried playing it with some brand new people showing both sets of cards and saying “We’ve not settled on a direction”. It was unanimous – the coloured icons were preferred.

There was other feedback too and I continue to work on the game, changing the design in accordance with things people suggest and working on balance tweaks here and there. Here’s Ripple in the version I’m testing at the moment:

So what did I learn from all of this?

Well firstly that I need to be aware of my own perceptions. A lot of designers get games wrong for colourblind people because they’re not colourblind.

Conversely I am colourblind and pay little attention to it as a signifier. Moving all of the icons into the same palette significantly increased the time it took players to figure what did what. Only by seconds, but an increase in seconds for a task that’s done a lot of time in a game can be the difference between something which plays smoothly and something which feels juddery and uncertain.

Just because I recognise icons by shape first doesn’t mean other people aren’t seeing colour first and having them all be different colours makes it easier and faster for players to recognise them and make judgements.

Secondly I would never in a million years have asked a designer to draw me an icon using an instruction like “I need a 15mm circle with an icon inside it, but the icon should overflow the limits of the circle going up to 3mm over the line” but the proportion of players who thought that was a deliberate design choice and liked it is somewhat mindboggling. I’m sure someone drawing to make that an interesting style on purpose can do a better job than I did by just putting a square peg in a round hole – but I’ll get someone to knock together some icons like that deliberately drawn and see if we can’t improve the process while taking advantage of a happy accident.

I’m still not sure how to face the “optical illusion makes coloured icon next to the other icon look bigger” thing, especially if the icon overlapping its bonds is part of that, but I’m going to try a bunch of stuff.

When we test something we don’t always wind up testing what we intended to, but I think it’s good to be open to new ideas.