• "My challenge to you is as follows. Design a game which is appealing to play, which will go on to be a huge commercial success and yet illustrates through its systems the abject and total horror, the inhumanity, the alienation, the banality, the evil, and the hell-on-earth of a socio-political practise taken to extreme. The game must be named honestly. It must be easy to learn. It must be a game for all the family." As expected, this is great, but of course it is, because Martin is great. More to the point: it's shrewd and useful. (And: excellent nous from Cara to pick up this piece from someone who clearly could become a major games journalism talent. Please keep commissioning this "Martin Hollis")

Spelunky is a little clockwork world in which items and enemies behave in defined ways, but when mixed together cause a delicious feedback loops that you can, with experience, predict. My boy loves systemic games like this, games that are built on coherent systems that you can play in an open-ended way. Toy boxes like Minecraft and Plants Vs Zombies, Animal Crossing and (we play this together) Civilization – where he can tinker and learn cause and effect.

He spends hours playing them, or would if we let him. And these are the kind of games that, though they were much cruder back then, I liked when I was a boy too, especially Elite. Where anything seemed possible.

But the big games today, in which play comes fixed to immutable stories, aren’t like that. So I asked him: “Do you like games that tell stories that you follow as you play them, or do you like games that let you do what you want?”

“The second one.” My heart burst with pride.

Last Day Of School, from lovely chum Alex Wiltshire.

The big computer games I grew up with – the ones that made an impression – were Rogue, Prince of Persia, and countless flight sims (beginning with MS Flight Simulator 3.0 then 4.0, and then getting steadily less realistic through the MicroProse back catalogue). And there, really, is a lot of the things I like: deep systems, short repeated play sessions, complex things to master, coupled t worlds to do whatever you want in. I got to about level 14 of the dungeons of Yendor; I landed a Cessna 182 on a Nimitz class carrier without an arrestor wire (only just) and took it around the Michigan bay; I explored the deep 60-minute run/jump mechanics of Mechner’s early triumph.

My tendency to simplification as I grew up has a lot to recommend it – in particular, desigining for sofas or tiny bursts – but my heart swells too when I see the conversations Alex and his son have. Not just because of what they like – but because of how they like it, and, most importantly, how they talk about it together.

  • "It should be pointed out, however, that physics is not the only systemic toy upon which fun games can be built. Probability fields, such as those forged by the colours, numbers and suits in a deck of cards, and the stochastic patterns that emerge from mixing those cards up, are another well-known toy upon which many great games are built. In fact, there is a literal infinity of foundational systemic toys upon which meaningful games can be built, yet for the most part, the game industry focuses on building baseline game engines that simulate one single toy that is proven to only be marginally fun: physical reality."