Sponsored By

The recipe for Candy Crush Saga: luck, skill and puzzles

Why is Candy Crush Saga so successful? I've think I've found 12 recipe ingredients - including the way the skill/luck mix changes across levels and episodes, and the use of variable perceptual super-rewards. The ability to buy unhappiness is also key.

Heather Stark, Blogger

October 3, 2013

13 Min Read

King.com is pulling in one billion daily gameplays for its f2p games, according to Reuters. Candy Crush Saga, its top performer, is estimated to be bringing in $840k per day in the US iOS market alone, according to ThinkGaming.com. Since King.com dropped in-game advertising earlier this year and has only recently started to offer Candy Crush branded socks, it's fair to guess pretty much all this revenue comes from in-app-purchases.

Level 1 of Candy Crush Saga

Happy times for King.com (the trading name of the game's developer), MidasPlayer.com (the UK registered company that trades as King.com), and MidasPlayer.com's Maltese-registered parent (Midasplayer International Holding Company). An IPO is thought to be brewing. There are, of course, detractors. But compared to the backlash against Zynga's monetisation tactics, the nature of the complaints is somewhat different. Civilian comments often centre around the complaint that the pay offers sprinkled throughout the game make it too easy to rack up spend (e.g. this LA times review). Rather intriguingly, some people dislike the game, but pay for it anyway, like the gizmodo commentator who called it "a simple and blatantly unoriginal time-waster" and in the same breath said they had spent money on it - but only to to get from one episode to the next.

Under-rated but essential: luck

Luck has been hugely important to Candy Crush Saga - in two ways.

Firstly, because they got lucky. Not even King.com foresaw how successful the game would be. According to an interview with King.com games guru Tommy Palm, in ValleyGawker, Candy Crush Saga's success was "nothing that we anticipated originally". They clearly thought it was promising, based on initial play testing and monitoring of the arcade version on King.com, but its reception in the market when released as a saga version on Facebook was even better than anticipated.

They then successfully invested in further growth via advertising, and via continued development of additional levels and monetisation inducements in the game. I have a mental image of teams of people at King.com kitted out in brightly coloured knitted curling team outfits, all sweeping quickly with brooms, trying to make the Candy Crush Saga stone go further, further, further along the ice....

Secondly, the game itself requires luck as well as skill to succeed. This can be easily misunderstood. One analysis of Candy Crush's design on Gamasutra, by a professional game economist, claims that early on, when you cross your first river in the saga map, the game changes from a skill game to a "money game", in which the dominant factor is luck, and you need to buy more and more chances in order to progress.

But CCS is not a game in which you can reliably buy your way to happiness by bribing Lady Luck. It is unlikely you could win a game without trying your best. And this requires skill. But mere skill is not enough, either. Luck is also required.

This skill/luck symbiosis is one place in Candy Crush Saga where the secret sauce really bubbles and boils. You can, if you wish, buy more chances at beating a level. And you can buy more tries at a game. These pay gates will give you more shots at being lucky - and more shots at exercising your skill. And I suspect it will sometimes give you the desired result - but not always, or possibly even not usually.

At many levels of Candy Crush, making an optimal-within-the-limits-of-human-cognitive-processing choice of moves will usually result in failure to clear a level. And it's my guess - as a non-purchaser - that although a pay gate is never far away, you usually can't buy your way out of failure. You can try to. But I think you can buy more unhappiness much more easily than you can buy happiness. Very scando. But clearly effective.

By contrast, success occasionally comes unexpectedly in extremely generous measure, in an orchestral crescendo and visual extravaganza of seemingly ever increasing and never-ending awesomeness. Wow. I don't smoke cigarettes but if I did I might want one.

This mixture of very tight level gating (both skill-based and chance-based) and unpredictable super-rewards is, I think, part of the appeal of Candy Crush's gameplay. The recipe is no where near as simple as making life difficult for people, and then offering them an easy way out by paying. It's much more interesting.

Skill is hugely important

Candy Crush is a game of skill. The specific skills required of the user are relatively straightforward in principle, but difficult to implement in practice. At least I find them so. Your mileage may differ. This is a quality which the co-founder, Ricardo Zacconi, calls "easy to learn and difficult to master".

The game's design is also skilful. Intentionally so - clearly. But possibly also in ways that its designers are still developing and enriching their understanding of.

Here's what I think it does right:

a. it is impossible to make an illegal move - the UI simply won't allow it - and if you don't move within a reasonably short amount of time, the UI will twinkle an option at you

  • this makes for easy on-boarding, and quickly builds a feeling of competence (which won't persist.... but hey ho one can always hark back to those glory days)

b. beautiful graphics and effects - what happens on the board as a result of most moves is simple, but looks and sounds just great

  • this makes "grinding" - playing repeatedly without winning - enjoyable in itself

c. irregular super-reward cascades - sometimes the effects of a move are just outrageously lovely and satisfying - not only in their consequences for your progress in the game, but in themselves, as effects

  • this is one of the best deployments of the powerful operant conditioning effect of variable reward intervals that I've yet seen

The rockstar designer and monetisation guru Michail Katkoff puts it this way, in his Game Analytics blog post: [the] "...graphical and audio feedback that follows these combinations is simply over the top. That massive fanfare of feedback is also particularly important for our casual gamers, as they aren’t traditionally good at playing games. With this kind of gameplay feedback we can make them feel good about themselves – we can make them feel like true masters". My own view is that the variability of this perceptual reward is just as important as its lovely over the top quality, in facilitating the desired outcome - lots of play. (And with that, perhaps, lots of pay.)

d. big variety of paygates on offer - just in case, just in time, special powers, more plays, extended games, no waiting between episodes...

  • I've no idea which ones are most successful - but whatever floats your boat, purchase-wise, you can probably do it

e. saga format - visualisation of level progress via a map-based progression story given structure by division into episodes

  • this is a very simple but effective way of creating a sense of visible progress and achievement out of an activity which is basically playing more and more (and usually - but not always -more difficult) variants of the same game over and over and over again

  • it is not in itself a defensible competitive advantage - it is too easy to copy - but there are experience engineering aspects of the saga that, being less obvious, might be more possible to retain early mover advantage on

f. social facilitation - much reference is made in reviews to the importance of the leaderboard as a social feature Facebook-connected games - but, bizarrely, nobody seems to make much reference to the positively reinforcing social mechanisms the game uses - my friends can be rewarded when I play, and when I succeed, and I can help them at no cost to myself when they request it. Very potlatch.

  • rewarding friends sets up both a reminder function for re-engagement for friends, as well as offering them gameplay benefits - and it shows a good understanding of human ethnomethodology, and the importance and power, across cultures, of reciprocal gifting arrangements.

g. luck/skill mix - neither luck nor skill will get you through the game's levels - you need both

  • the mixture, in combination with the basically attractive nature of play, and the intermittent use of super-rewards creates powerful psychological motivations - for some people - to persist in play. I hope to have time to talk more about this, as I think it's fascinating and poorly understood.

h. exquisite split-second comic timing - I laugh out loud love the mischievous way the game pauses to let me contemplate and prepare for my next move, before telling me my time is up, and offering me the chance to pay for further moves.

  • an action plan interrupted creates a tension - which can be relieved by paying

i. habit automation exploitation - when completing successive tries at a level, you select a green button to go to the next move, but when you have run out of tries, the pay button is in the same place and looks the same - except it's pay rather than play...

  • people can easily back out if "pay" isn't what they want to do - but my guess that this little kick onto the first step towards payment has helped more than a few people give it a go

j. varied level progression - in general, later levels are harder, but there are odd plateaus and even reductions in difficulty, given the accumulation of skill, where progress is rapid

  • my guess about the effect of this is a kind of psychic momentum, and build up of expectation about continued progress, which, when thwarted, induces a need to bring things back to plan

k. demanding yet snack-able

  • playing one game gives you a good workout - and an immersive power break - but it only takes a few minutes

l. the use of candies as game tokens, and gravity cascades as a board configuration change movement

  • feels good

If you copy all these design features wholesale, will you easily make another Candy Crush? Almost definitely not - for several reasons. For one thing, I'm sure there are many aspects of the game I haven't noticed. And it's more than possible I've noticed stuff that isn't actually there. It's not just me. People are really good at that.

For another thing, even King.com hasn't yet made another Candy Crush. One nordically blunt teardown from Michail Katkoff of one of their newer titles, Pet Rescue Saga, says it has "simply copied the mechanics from CCS without actually making sure that they fit the game". Ouch. But, even though it isn't a Candy Cross Saga, Pet Rescue Saga is a game many a studio would envy - it is successful by mortal standards. It's just not as jaw-droppingly successful as Candy Crush Saga.

The transplant of a mechanic from one game to another is not necessarily straightforward. I think that design success is an emergent feature of all the components I've called out, and more, working together in the right proportions. Transplanting them to a new context is something that is even more difficult than transposing the key of a piece of music. You have to expect it to feel different, in a different context. It's more like transplanting a peony: there's a big chance it won't work.


I think I understand some of what makes Candy Crush Saga tick. But there are lots of things that puzzle me.

One is payment. A biggie, eh? I started playing Candy Crush Saga out of a desire to reverse engineer my experience into an understanding of the game's finer points. And I have continued to play it for fun. But I haven't paid. In this I am not alone. According to oft-quoted Tommy Palm, King.com game guru, the game was designed so that it would be possible to work through it without paying, and indeed, of those people who make it to the top-most level, 70% do so without spending anything. Even a penny. (Except, perhaps if they are British.)

This means, of course, that 30% of the top-achieving players have spent money. And who knows what percentage of people who are toiling towards the top - but don't succeed - actually try to buy their way there? It could be more. It could be less.

The precise personality factors and situational triggers that work to inspire payment are a mystery to me. But that shouldn't stop me from having fun guessing. Mixing typologies with abandon, I would guess that achievers and completer/finishers would feel the pull more strongly. Situationally, I'd guess the almost-there-but-for-one-more-move situation would be a really strong trigger for payment, in-game, and a feeling of progress across tries before expiring free plays might tip me over to wanting to top up plays before they replenish with time. But my puzzle is that this situation arises so very very rarely in my game play. Usually I'm either miles out when I lose, or everything comes right. Very rarely would any of the offered powerups make a difference. And I like the fact that one's playtime is limited. I lose nothing by waiting.

Another puzzle is the massive dropoff they engineer into the tougher levels. Tommy Palm said in his ValleyWag interview that they needed to tweak difficulty of level 65 down to the point where "only" half the users dropped out at that point. That's one hell of a tough gate. Worse than Beecher's Brook at the Grand National. It's really puzzling that they'd think it a good thing to lose "only" half their users - when presumably they could, without straining themselves terribly, have made it easier still. The only way this could be a good result is if a goodly percentage of those players paid before quitting. More so, perhaps, than would be the case if a level was almost impossible. I can feel a graph coming on.

Another puzzle about the game for me is that I think there is some meaning to the scalloped pattern of level difficulty that I haven't quite understood. Tommy Palm said that last level in an episode is said to be the most difficult - but I'm not sure that's been my experience. Still chewing on this, and I need to think more about it.

And finally, I'm puzzled about the style of the graphics that go with the saga level map. The game board itself is bright and shiny, hyper-real, strongly lit with occasional sparkles. But the saga graphics and the character that lumber around it in an extravagantly 2d way are from a different mood board. A bit clown-sinister. A bit ironic. But not, perhaps, quite enough. I wonder what work that clash and tension between the two graphic styles does, in terms of affecting the feeling-tone of my experience. If I think I've figured it out I'll let you know.

[This piece was first published on my blog, insightanalysis.]

Read more about:

Featured Blogs
Daily news, dev blogs, and stories from Game Developer straight to your inbox

You May Also Like