Candy Crush Saga

Discussions
Chat about Candy Crush Saga, share thoughts, news, or anything sweet!
Support
Support - Stuck or having an issue? Ask a question and get help from other players!
Contests
Enter for a chance to win Gold Bars and sweet in-game rewards in Candy Crush Saga!
Ideas
Have a sweet suggestion? Share your ideas to help improve Candy Crush Saga.
Off-Topic
This is your chill zone! Talk about anything that’s not Candy-related with other Candy players!.

All You Need to Know About All Stars 2026

Learn More

The Eureka moment

Hello friends. I am posting this here away from view in the off topic section because I really don't want to take the fun out of the competition for all the newbies out there who would not understand my post and only get confused by it all. I'm not sure I will add all my data, but I have loads of it and it all leads to the same conclusion:

We are playing against recordings in the first three All Stars rounds.

I can't believe it's taken me so long to realise this. I started off about a year ago and I have repeatedly checked out previous leaderboard events and always seen the same patterns. One example is this post from August 2025:

https://community.king.com/en/candy-crush-saga/discussion/506138/suspicious-player-behaviour-in-leaderboards

Since the AI took over running the game we have seen the same patterns again and again and again in live leaderboard competitions.

I have taken extensive notes on all my accounts and found that players just stop playing for no good reason. There can be no other conclusion than that time is being manipulated by the app. I cannot work out the exact mechanics of it. But the only way I see that this can be done is by making us play against recordings, in much the same way that we know we are playing against recordings in the Episode Races.

The pattern is almost identical as well: One player speeds ahead, often then to be overtaken by a different player, to make it look more realistic. Two players (about 25% of the leaderboard) fail to make any progress at all, or just one or two levels. And then near the end we are allowed to catch up if we want to and get the "win". In the Episode Race this is the top spot. In the Weekly Contest this is a podium place. In All Stars it is a qualifying spot.

In the Weekly Contest and in live leaderboards usually one player is so far ahead we would have to try extremely hard to beat that one player. This one is for the "mad crushers" to pit their skills against and in my two newbie games I am not seeing this in All Stars, probably because I have been playing those accounts slowly: the app thinks I am a genuine newbie who cannot crush any faster so it's giving me all easy opponents I can beat this round. It looks to me like King want every single player who really cares about it to get to the final knockout rounds this year.

The mechanics of how this feat is achieved is not clear but my guess is this: I have noted that it takes two levels to be played for anyone to show up on our leaderboards. When I was first on my leaderboard last night it filled up ever so slowly until I had played my second level as well. So I think the app calculates our own speed and that of others in the first two new levels we play as we join. Those players who show up and then never play another level seem to often have a score of over 200, so it looks to me like they played two levels and then got stuck on the third level. I have also found that near the beginning of the joining process the leaderboard takes a long time to load but after that updates very quickly. This makes me think that data of other players is being loaded and stored onto our devices at the beginning of the round. Possibly at this point the speed of the other players is decided as well, according to parameters of our own playing tactics and level. I suspect that the eventual outcome (the score we will have to beat) is pre determined in the same way that the outcome of the gold we will can win in Candy Royale is pre determined. (We have confirmation from a post made by the CM about two years ago that Candy Royale is recordings). This is my theory and I have ample evidence to support it. I'd be happy for anyone here to disagree. I really don't like the findings but I cannot draw any other conclusion.

What complicates matters is that some of these players on our leaderboards can become our friends, so this shows that they are real and actively playing. However, the recordings may be from an earlier round. Or maybe it's a time-lapse of some sort. That would mean the server would have to update their progress periodically and recalculate the speed of the others - possibly according to how many levels we ourselves have passed or if we have played a multiplier ourselves-, which may be what is happening. Some of the players may be real and some may be older recordings from years ago (such as the ones who stop playing altogether).

I just hope that the final knockout rounds won't be run in this way. I definitely hope that final knockout round 3/3 will be run as fairly as it was last year, with ten leaderboards and players who see one another. We know for a fact this was the case last year. Let's hope it's the same this year.

Good luck to you all!

I'll tag those who might be interested and not find this thread otherwise:

To join the Conversation:

Comments

teenybee

Thank you for your fascinating research and great analysis, @christinewupp. It was quite spooky to read it as you had predicted exactly what had happened on my board yesterday evening. The runaway leader had been overtaken by the 10th placed ‘player’, who had played a token. I followed suit to go top (since I’m only in this to try to win as many boosters as possible in the early rounds) and then the original leader played a token to regain first. Exactly as you said. As of now there is a sizeable gap between 3rd and 4th.

It also makes sense of a lot of oddities eg everyone having the same initial experience as regards loading, ‘players’ playing at strange times etc.

Well done. I wouldn’t worry about shattering anyone’s illusions, by the way. Most experienced players aren’t surprised by anything in this game anymore. What’s fascinating is your working out how it’s been done.

Good luck with beating the system!

MannyFae

In fact I cannot see any way that final knockout round 1 could function unless it was also a fake round like that second chance round. The aim must be to get as much money from as many players trying to beat the one player in first place.

Agree with this. After reading Testy's story, I thought it was a sort of a preview of the final knockouts. In fact, both rounds 4 and 5 can possibly be organized in this manner.

christinewupp

I agree. I suspect we might all have to use our remaining tokens. I already have experienced one such leaderboard in part two of the second chance round. The Testy account had to definitely burn four tokens to get past that round, it was fixed that way. The score in second place was over 68,000. That is equal to 34 episodes without tokens and it was a 24 hour round! In fact I cannot see any way that final knockout round 1 could function unless it was also a fake round like that second chance round. The aim must be to get as much money from as many players trying to beat the one player in first place. It may be that players who use a token will see opponents that also use tokens and those who do not use tokens might get easier opponents that they could possibly catch. In the end though (after enough players have spent enough money to try and compete) there would have to be many who still fail to beat the leader. Only if the actual winner is fake could this work.

MannyFae

Okay, let's assume that they are using recordings (or bots, or whatever) to get as many players as possible to the later rounds. If every leaderboard is fake (or just bespoke, meaning 'Jennifer can see Paul but Paul cannot see Jennifer'), there can be as many leaderboards as there are players, so everyone who makes an effort will progress. This makes business sense as more players participating for a longer time means more money for King. In that regard, I'm curious how round 6 will be set up. If those leaderboards are also fake, then screenshots that will inevitably be shared will make it apparent that the whole competition is fake. I don't think King will want to acknowledge that, so I tend to think the last round will have real leaderboards. If the final 10 leaderboards are real, King will have to reduce the huge number of players currently participating to just 1000 participants in round 6. If this is the case, the final knockouts will be sheer hell.

christinewupp

https://community.king.com/en/candy-crush-saga/discussion/comment/3501820#Comment_3501820

Thank you for that @treeclimber and I'm glad you found my post.

I would like to add here that with the new multiplier token feature it seems doable that I could burn tokens to get several accounts all the way to the last round. My dormant Echinacea account with 14,000 gold bars is being invited to the last second chance round as well. It's not using a valid email so I won't play it into the last round, but I can use it to spy on the rounds. The fact we can burn multiplier tokens on spare accounts plays hugely into the hands of those who are trying to find out how the rounds play. So what I'm trying to do is send one of my games out as a spy early on for each of the three final knockout rounds and I will post about it here.

I had set out wanting to give it a real go this year to win All Stars, but since playing the event I have come to realise that it would give me just as much pleasure if someone else would win it using the information I am able to gather. I learned this also from last year's competition. I don't need the money and actually have other plans right now for that live finals weekend (which oddly enough involves a trip to London), which I would have to rearrange.

treeclimber

I think you are correct. I joined the knockouts tonight with about 48 hours left. After two games I was #3 on the board. After one episode race at a normal pace I was #1 on the board. Then suddenly one of the players sped up and overtook me. I decided to play another episode race just for fun and I could not keep up with this player. They did not appear to be using a token, but jumped ahead fast and I could not retake the lead. There was one other player (#3)“actively playing” and all the remaining players (4-10) stopped playing. At one point they had all stopped playing exactly one hour earlier, even though I’d only been playing about an hour. Just highly unlikely behavior. I took a break and two hours later I remain #2 and everyone has stopped playing.

I was suspecting the boards aren’t real based on your previous posts (which thanks very much have made me rather unconcerned about getting thru the first three rounds). Since I’m newer to the game, I don’t spend money, and this is my first all stars I’m not bothered by this and just consider it part of the game. I do appreciate your insights though because I now know I don’t have to kill myself or waste boosters to get thru this round. The AI will likely let me finish #2 or #3 with minimal effort.

Comments

teenybee

It also makes sense of a lot of oddities eg everyone having the same initial experience as regards loading, ‘players’ playing at strange times etc.

Good luck with beating the system!

MannyFae

In fact I cannot see any way that final knockout round 1 could function unless it was also a fake round like that second chance round. The aim must be to get as much money from as many players trying to beat the one player in first place.

Agree with this. After reading Testy's story, I thought it was a sort of a preview of the final knockouts. In fact, both rounds 4 and 5 can possibly be organized in this manner.

christinewupp

MannyFae

christinewupp

https://community.king.com/en/candy-crush-saga/discussion/comment/3501820#Comment_3501820

Thank you for that @treeclimber and I'm glad you found my post.

treeclimber