How does the game determine level difficulty? Some conclusions and answers (Nov 2025)

christinewupp

I've been trying to get to the bottom of this question for some time now and have started running some tests.

I would like to specify what I now consider the definition of "level difficulty" in Candy Crush:

Level difficulty is the number of attempts needed until the game will allow a player to pass*

My questions were:

1. Are we all playing the same difficulty? (for each specific level)
2. How is difficulty set and managed by the AI?
3. Is there a pattern?
4. Are labels correct and are they still being changed?
5. Do starting boosters still help get luck?

Some evidence for my findings and conclusions is posted in the posts that follow.

This is what I have found:

Answering questions 1., 2. and 3:

No, we are not all playing the same difficulty.

Difficulty is determined by the AI using these factors:

1. Overall the game manipulates difficulty by features and events now. This includes the super charger, timed boosters from various events, extra moves from friends, absence of ads etc. Events are a way for King to make the game easier for some test groups and harder for others. The individual level board algorithms include some boards that either require the use of certain boosters or +1, +3 or +5 moves, and the events will then boost specific player groups to pass those boards.
2. The overall basic level pass rate algorithm is quite simple: It includes four types of outcomes: an easy pass, a pass with skill or on last two moves, a bad fail, a near miss (one jelly or few moves short)
3. I think there is evidence of a "first attempt" pass rate, (or maybe a "first five attempts" pass rate) being written into each level in a bespoke manner and this is superimposed on top of the basic level pass rate, which is pretty constant for each test group. I have seen some signs that this will get renewed each time we load the game after closing the app.

The answer to question 4: Are labels correct?

No.

Labels are not being updated and are blurred and overlap.

All of the labels hard/super hard/nightmarishly hard can be impossible for some players to pass without boosters (and also with starting boosters in many cases) for so many attempts that these labels have become meaningless.

All of these labels can be impossible even with starting boosters plus 5 extra moves on a first attempt or possible without boosters on a first attempt, blurring the distinction between labels.

Legendary levels probably are different by not containing any lucky board algorithms at all. For these the best luck is a near miss, regardless of the number of replays. But for obvious reasons I am not going to test this out. I go by reports I read here in the community to support this assumption.

The answer to question 5: Do starting boosters still help?

This varies.

Not nearly as much as they used to before the AI took over. Some levels they do help improve pass rate. Some levels they have no effect whatsoever. The game difficulty has seemed to have increased to many players since the AI took over. I think what underpins this is probably the loss of many "pass with boosters" boards.

Lucky boards are systematically being removed from Candy Crush Saga. This seems to apply across all level ranges and all four difficulty labels.

The result has been a game with a stark contrast between the easy (lucky) levels and the hard+ levels.

*"Allow a player to pass" is not the same as a lucky board. It includes boards where players need to use some expertise, appropriate for the level range they are at, to spot a possible pass. A lucky board will actively help a player to pass, either with hints or lots of opportunities, and then the jelly fish finishing off the job. For players at the highest levels "difficulty" will also factor in use of strong boosters, so might be more accurately defined by the number of boosters needed to pass

christinewupp

Now here is an interesting example of how the level difficulty is plain wrong:

I've taken a look at this episode this morning and noted that there are an unusually high number of legendary levels. This one is in the middle of the episode, so it seems to be an outlier. I played it to see if the new jelly fish that won't target the cakes ruin the level. But far from it! With both starting boosters I have found that all I had to do is follow the suggested moves and it led me to victory. And this was also true on repeated efforts. Starting boosters are enough to pass this legendary level, even for players who do not understand the strategy (just follow the hints!) and even with the now simplified fish.

I think the game is deliberately helping players who cannot figure out these complex levels. Because this level is in the "easy" section of the episode the game has made the level passable for all. But so far the label has not changed but is clearly not correct.

christinewupp

Now here is an interesting example of how the level difficulty is plain wrong:

I've taken a look at this episode this morning and noted that there are an unusually high number of legendary levels. This one is in the middle of the episode, so it seems to be an outlier. I played it to see if the new jelly fish that won't target the cakes ruin the level. But far from it! With both starting boosters I have found that all I had to do is follow the suggested moves and it led me to victory. And this was also true on repeated efforts. Starting boosters are enough to pass this legendary level, even for players who do not understand the strategy (just follow the hints!) and even with the now simplified fish.

I think the game is deliberately helping players who cannot figure out these complex levels. Because this level is in the "easy" section of the episode the game has made the level passable for all. But so far the label has not changed but is clearly not correct.