Thread by @tronbevan

v dark spooky talk on cultist simulator coming in 10 MINS #GDC19

key point of talk: art is important, but food is also important. you can make experimental arty games and make money at the same time! #gdc19

what are experimental games? fallen london - a divisive, content-heavy f2p browser game with no discernible genre. is stupid and shouldn't work. and yet! made ~60k usd/mth 7yrs after launch #gdc19

sunless sea - world's slowest rpg with permadeath! also a stupid idea with no genre which divided people. AND YET! 4.5m usd yr1 against a 350k usd budget. #gdc19

cultist simulator - deliberately obscure occult solitaire with no tutorial (lol). no genre, divisive, and 2m usd rev yr1 against 190k usd budget + 2 bafta noms #gdc19

so doing stuff no one else is doing is a great advantage. but there's often v good reasons why no one else is doing it, and the key is working out what's actually a terrible idea and what's secretly a great idea just *wearing* a 'i am stupid' tshirt #gdc19

ak's successes all have in common: they're all distinctive and divisive, all reasonably good, all had super smol budgets and very smol timelines (sunless sea = 16mths, cultist = 11mths), and all used 'apophenic design' (more on that later) #gdc19

lean into your limitations as an indie! be distinctive: e.g. cultist's key art (thanks @ungapants) which catches the eye, evokes emotions and can be copied (because it's its own thing) for relatively cheap, allowing us to do a lot with only a little £ #gdc19

divisive comes at a cost: cultist tends to dip into mixed reviews on steam (😱) when it goes on sale, even though it's firmly at 77% overall; we get weird negative user reviews from people with 60hrs in the game... but divisive = ppl talking about it! yay! #gdc19

but going back to making a 'reasonably good' game... you gotta be a feedback sponge. feedback is vital. design your process for this: playable core loop prototype from day 1; open development; educate your community so they can give better feedback; roguelike is GREAT... #gdc19

...so that you don't burn out testers, who stop giving useful feedback on the opening moments/onboarding later down the line because they get too familiar with the game. #gdc19

now, production! 😎 spend as little as possible on your experiment, cos it might well fail. re: @GreyAlien's twitter poll about how long devs have been working on their game. and how little they're likely to end up being paid/hr as a result #gdc19

there's a good ANALOGY OF POTS here. if you make a bunch of cheap experimental games, one of them is likely to do well, and the others won't kill your studio. if you spend all your time and money making one single giant experimental game, i hope v much it's a hit #gdc19

now, wtf is 'apophenian design'. apophenia is the human tendency to see patterns in things, and to find meaning in disconnected things. harness this with game design: don't simulate, instead imply; make your mechanics visible; lower budgets & don't fill all the gaps #gdc19

use the power of the human mind to do your work for you! it's always better than what you could do, the player will feel more engaged, and the £ / cpu cost is a lot lower than the alternative #gdc19

text is a great tool here: it's cheap and relies on the human mind doing a bunch of the work anyway. players also *expect* to do a bunch of the work themselves. #gdc19

game mechanics can be great too. e.g. in cultist you only have one work verb: at one time, you can either do something fun and magical, or you can choose something sensible and lucrative. players think WORK OR MAGIC UNNGHHHH but the game doesn't have to explain that #gdc19

heretic operative by @CPromptGames is another great example of apophenic game design. take a look at their character stats, particularly 'corruption' vs 'tranquility'. says lots about the game world, the character & the situation, all in simple stats on a HUD #gdc19

the downside of leaving apophenic gaps is that players sometimes wander off in the wrong direction. so alexis's solution is theme-first design, so everything in the game - music, art, ui, narrative, music, the player's interest - all point in the same direction #gdc19

ak also recommends making your design visible in-game. players are smart and increasingly engaged with game design, and design can be beautiful like a skeleton watch showing its owner the intricacies of telling the time #gdc19

finally: be prepared to cut. in-game spaces are your friend, and remember, you want to make a cheap interesting experiment here. cutting helps focus on what's necessary, though leave a 'grieving period' for designers to let their babies go #gdc19

(the production vs design tension is a really useful tool, but can go south quickly, so be careful with human hearts) #gdc19

speaking of going south, here are some things we'd do differently. going without genre means you tend to balance yourself between other genres, lumping lots of expectations on the type of game it actually is. identify the faultline between those genres & explain them ASAP #gdc19

e.g. cultist's faultline was being a game of exploration vs RNG clicker progress. we could totally have explained that to our audience better, and refined its game design to make those two things work better with each other #gdc19

gdc asked ak to talk about cultist's ux, which was at the core of a lot of cultist's negative reviews. cultist's ux is a venn diagram of intentional obscurity, mechanical limitations & lack of polish (cos cheap, short, experimental) #gdc19

in experimental games, IT'S EVEN MORE IMPORTANT TO GET THE UX AND UI RIGHT. if people are expected to connect the dots themselves, you better make sure the dots are well-lit and easy to interact with. #gdc19

and that's it! if you have any questions, ask 'em here! #gdc19