COME HERE FOR EXPLANATION OF PUZZLE (Updated)

[OMGyouMuPPeT]

4 minutes ago, LuisLzn said:

In my experience with @OMGyouMuPPeT in the second part of the puzzle, it was a matter of finding out the match that you get all the time (you always get one light when you both enter your codes and click the button) and then moving around the rest till you find them all

1.- don't look at the symbols, look at both secuences like numbers from 1-6 so your 1-6 and your friend's 1-6.

2.-you have to move them in pairs all the time (at lest that worked for us), so to find the match you always get, choose 2 tiles and swap them;
let's say you choose 1 and 3, so you move 1 to the 3 and the 3 to the 1, and your friend does the same thing, then click the button to sync them, if your light goes off, that means either the pair in the position 1 or in the position 3 was your match

3.-try returning one pair of them to their original position so let's say  return the pair 1 to the position 1 switching it with another pair that is not the pair 3, if the light goes on then that's your match if not, move the pair 3 to its original position. 

4.- once you find your match then is all about moving the rest in pairs all the time, just switch them around till you get more lights.

5. or just move them around in any order but always synched with your friend, so you tell him let's swap 1 with 5, so you move 1 to 5 and 5 to 1 and your friend does the same, then press the button to sync them, try that till you get all the lights. 

 

Hope this helps someone. 

We spent so much time on it, but we got it, bless the potato cup, thanks still Luiz

[ButterStuffed]

I'm just about to lose my mind with this puzzle. Good 2 hours mindlessly putting in the codes with my friend and haven't gotten anywhere above 4.

[Pioneershark]

9 hours ago, 321579_1452782593 said:

really confused because this is all I get when I log into steam on the puzzle page , anyone else get this?

you haven't completed the cyclum puzzle, you have to complete the cyclum one first

 

[fsp98]

i can't do it, i just can't

 

[rezecib]

I've tried a scientific approach to this, that is, coming up with models and developing experiments to falsify them. Like with science, a model cannot be proven correct, but it can be disproven.

First I'll designate notation. Tile guy has symbols T = {t1, t2, ..., t6}. Symbol guy has symbols S = {s1, s2, ..., s6}. A submission consists of a sequence of 6 ti and a sequence 6 si, e.g. {t1, t1, t1, t1, t1, t1} and {s1, s1, s1, s1, s1, s1}. Each position is numbered p1, p2, p3, p4, p5, p6.

Science:

Spoiler

 

Model 1: each position pi has a scoring function; if si = sj and ti = tj, then you get a light. Note that this means that sj and tj are paired; there must be some shared correct position for them.

Experiment 1a: submit {t1, t1, t1, t1, t1} and {s1, s1, s1, s1, s1, s1}. Under Model 1, we should get one light, because s1 and t1 are paired, so one position should be correct.

Result 1a: no lights. Model 1 must be incorrect. Error discovered later: the scoring function uses the last submission from each side, so both sides must submit after any change that side makes (e.g. at the start both submit, if side A is the one making changes then only A needs to submit after that). Later results support Model 1.

Model 2: each position pi has a scoring function; if si = sj and ti = tk, then you get a light. Note that this means sj and tj are NOT paired, unless j = k.

Experiment 2a: submit {t1, t1, t1, t1, t1, t1} and, for each i={1..6}, submit {si, si, si, si, si, si}. Under Model 2, at only one such submission should result in a light (unless there are duplicate symbols), because one t1 must be in the correct position, and that position requires some particular s, which will only be satisfied by one of the submissions. If there are duplicates in both T and S, the experiment can't be run. If there are only duplicates in one, select one of the non-duplicates for submission and then run through all possibilities in the other, duplicate-free set.

Result 2a: I received the expected result, with only one submission causing a light.

Experiment 2b: My successful submission from before held T static, with t1, and succeeded with s1. Only one position should be correct for s1, so I will binary search for its position, using an extra symbol not in S, say, s7, to fill spots I am not testing. At each fork of the binary search, I will test both sides, in order to ensure it adheres with the model (if it does not adhere, there may be a case where both forks test positive).

Result 2b: Aside from a possible bug at the start, where I did not replicate 1 light with the full-s1 submission, once I was able to replicate the s1-submission light, the binary search proceeded as expected, and I localized the s1 to p6.

Experiment 2c: Repeat 2b with tile submissions; we should be able to localize the correct position of t1.

Result 2c: I ran out of tries at the start here, so I had to repeat the model-consistent procedures derived from 2a and 2b. However, this time T consisted of 3 pairs of duplicates. So 2a should result in 2 symbols that give 1 light; this happened, with s1 and s6. 2b proceeded as expected, localizing s1 to p2 and s6 to p4. Running 2c resulted in localizing t1 to the same two positions, p2 and p4. This aligns with Model 2. It also aligns with the more restricted Model 1, raising the possibility of experimental error in 1a (both sides must submit at the start, and each side must submit after a change; I did not realize this at the beginning).

Experiment 2d: Fill the remaining slots with the next t and s, t2 and s2. 3 lights should result under Model 1, and for Model 2 with one of the remaining s options, and we should be able to do a binary search for their positions. Their positions should be the same.

Result 2d: We did get 3 lights after filling in t2 and s2. This supports Model 1. Binary search proceeded as expected, and positions matched again: t2 and s2 in p6.

 

Okay, so based on Model 1 being accurate-- that is, s1 is paired with t1, and they both belong in the same position pi-- I can derive an algorithm. First let me define binary search, so we can use that as a subroutine.

Binary search (customized for this domain):

1. We know t1 has some correct position in p1..p6. So place it in p1, p2, p3. If you get a light, then you know it's in those three. Otherwise, it must be in p4, p5, p6.
2. Let's say it was in p1, p2, p3. Take t1 out of p2 and p3, and fill in t2 (your next symbol filler). If you get a light, yay, t1 goes in p1. Otherwise, it goes either in p2 or p3.
3. If it wasn't p1, put t2 in p1 and t1 in p2. If you get a light, yay, it's in p2. Otherwise, you KNOW it's in p3. Submit the correct layout so the other player can do their binary search.

Solution algorithm:

1. Since (ti,si) are paired, player T should first submit all s1 so that player S can binary search for the correct position of (t1,s1). Player S first submits {t1, t1, t1, t2, t2, t2}-- using t2 as filler allows you to use fewer submissions when switching to the next symbol. Complete the binary search to find pi=(t1,s1).
2. Let's say it was p1. Now that player S has submitted {t1, t2, t2, t2, t2, t2}, player T can binary search for the position of s2. So, for their first submission they try {s1, s2, s2, s2, s3, s3} (using the next symbol, s3, as filler for areas they are not testing). If they get a light, divide the area of s2's and continue the binary search. When the final position is found, submit and let player S binary search for t3.
3. Trade off like this until you solve the whole puzzle.

[Donke60]

3 hours ago, rezecib said:

I've tried a scientific approach to this, that is, coming up with models and developing experiments to falsify them. Like with science, a model cannot be proven correct, but it can be disproven.

First I'll designate notation. Tile guy has symbols T = {t1, t2, ..., t6}. Symbol guy has symbols S = {s1, s2, ..., s6}. A submission consists of a sequence of 6 ti and a sequence 6 si, e.g. {t1, t1, t1, t1, t1, t1} and {s1, s1, s1, s1, s1, s1}. Each position is numbered p1, p2, p3, p4, p5, p6.

Science:

  Reveal hidden contents

 

Model 1: each position pi has a scoring function; if si = sj and ti = tj, then you get a light. Note that this means that sj and tj are paired; there must be some shared correct position for them.

Experiment 1a: submit {t1, t1, t1, t1, t1} and {s1, s1, s1, s1, s1, s1}. Under Model 1, we should get one light, because s1 and t1 are paired, so one position should be correct.

Result 1a: no lights. Model 1 must be incorrect. Error discovered later: the scoring function uses the last submission from each side, so both sides must submit after any change that side makes (e.g. at the start both submit, if side A is the one making changes then only A needs to submit after that). Later results support Model 1.

Model 2: each position pi has a scoring function; if si = sj and ti = tk, then you get a light. Note that this means sj and tj are NOT paired, unless j = k.

Experiment 2a: submit {t1, t1, t1, t1, t1, t1} and, for each i={1..6}, submit {si, si, si, si, si, si}. Under Model 2, at only one such submission should result in a light (unless there are duplicate symbols), because one t1 must be in the correct position, and that position requires some particular s, which will only be satisfied by one of the submissions. If there are duplicates in both T and S, the experiment can't be run. If there are only duplicates in one, select one of the non-duplicates for submission and then run through all possibilities in the other, duplicate-free set.

Result 2a: I received the expected result, with only one submission causing a light.

Experiment 2b: My successful submission from before held T static, with t1, and succeeded with s1. Only one position should be correct for s1, so I will binary search for its position, using an extra symbol not in S, say, s7, to fill spots I am not testing. At each fork of the binary search, I will test both sides, in order to ensure it adheres with the model (if it does not adhere, there may be a case where both forks test positive).

Result 2b: Aside from a possible bug at the start, where I did not replicate 1 light with the full-s1 submission, once I was able to replicate the s1-submission light, the binary search proceeded as expected, and I localized the s1 to p6.

Experiment 2c: Repeat 2b with tile submissions; we should be able to localize the correct position of t1.

Result 2c: I ran out of tries at the start here, so I had to repeat the model-consistent procedures derived from 2a and 2b. However, this time T consisted of 3 pairs of duplicates. So 2a should result in 2 symbols that give 1 light; this happened, with s1 and s6. 2b proceeded as expected, localizing s1 to p2 and s6 to p4. Running 2c resulted in localizing t1 to the same two positions, p2 and p4. This aligns with Model 2. It also aligns with the more restricted Model 1, raising the possibility of experimental error in 1a (both sides must submit at the start, and each side must submit after a change; I did not realize this at the beginning).

Experiment 2d: Fill the remaining slots with the next t and s, t2 and s2. 3 lights should result under Model 1, and for Model 2 with one of the remaining s options, and we should be able to do a binary search for their positions. Their positions should be the same.

Result 2d: We did get 3 lights after filling in t2 and s2. This supports Model 1. Binary search proceeded as expected, and positions matched again: t2 and s2 in p6.

 

Okay, so based on Model 1 being accurate-- that is, s1 is paired with t1, and they both belong in the same position pi-- I can derive an algorithm. First let me define binary search, so we can use that as a subroutine.

Binary search (customized for this domain):

1. We know t1 has some correct position in p1..p6. So place it in p1, p2, p3. If you get a light, then you know it's in those three. Otherwise, it must be in p4, p5, p6.
2. Let's say it was in p1, p2, p3. Take t1 out of p2 and p3, and fill in t2 (your next symbol filler). If you get a light, yay, t1 goes in p1. Otherwise, it goes either in p2 or p3.
3. If it wasn't p1, put t2 in p1 and t1 in p2. If you get a light, yay, it's in p2. Otherwise, you KNOW it's in p3. Submit the correct layout so the other player can do their binary search.

Solution algorithm:

1. Since (ti,si) are paired, player T should first submit all s1 so that player S can binary search for the correct position of (t1,s1). Player S first submits {t1, t1, t1, t2, t2, t2}-- using t2 as filler allows you to use fewer submissions when switching to the next symbol. Complete the binary search to find pi=(t1,s1).
2. Let's say it was p1. Now that player S has submitted {t1, t2, t2, t2, t2, t2}, player T can binary search for the position of s2. So, for their first submission they try {s1, s2, s2, s2, s3, s3} (using the next symbol, s3, as filler for areas they are not testing). If they get a light, divide the area of s2's and continue the binary search. When the final position is found, submit and let player S binary search for t3.
3. Trade off like this until you solve the whole puzzle.

Can you say that in layman terms or simplier terms

my brain is dead

 

[rezecib]

Um... I can try.

Basically, both of the sets of symbols pair up. So we can use the same term for them. Let's call the players Player 1 and Player 2. They both get ABC,DEF, just in a different font or something. Just to show you how I'm listing the tiles, if they put in ABC,DEF it would look like this:

A B C
D E F

First, Player 1 puts in AAA,AAA; one of these is in the correct position.

A A A
A A A

Player 2 tries to find out where A belongs; they will fill A into the area they want to check, and fill the rest with the next symbol, in this case B. First they check the top half by putting  AAA,BBB, if that gives a light then they know A was in one of the top 3, otherwise it's in the bottom 3. Then they divide that area in two and test the split. Let's say it was the top; they do ABB,BBB. If that gives a light, then they know A belongs in the first spot. Otherwise, it's in the next two; try each of those until they get it. Let's say Player 2 ends up with ABB,BBB.

Now Player 1 can try to find the location of B; it must be in one of the slots where Player 2 has put B. Player 1 tests the first two slots by putting in ABB,CCC. If another light turns on, then they know it's in the two remaining top spots. Otherwise, it's in the bottom 3. They then divide the area where it is into two parts and test, just like above, until they find it and submit. Let's say they arrive at ABC,CCC.

Now Player 2 can search for C's correct position. They try ABC,CDD. If that gives a third light, then it's in top right or bottom left. Continue dividing in two and testing until they find it.

... repeat until you get all of them ...

Does that make sense?

[Brickfest]

Me and my buddy have been giving it a go... but the puzzle seems to reset after few minutes. Anyone knows if it's time or click limited? Or just a bug?

[Donke60]

3 hours ago, Brickfest said:

Me and my buddy have been giving it a go... but the puzzle seems to reset after few minutes. Anyone knows if it's time or click limited? Or just a bug?

its not timed though you only have a set amount of tries

 

[Brickfest]

Just now, Donke60 said:

its not timed though you only have a set amount of tries

 

Thanks man... I'm starting to feel a brain tumor growing already so this clears it up a little

[Donke60]

3 hours ago, Brickfest said:

Thanks man... I'm starting to feel a brain tumor growing already so this clears it up a little

Join the club me and my friend already are stressed out thanks to this guessing game.

[fsp98]

8 minutes ago, rezecib said:

Um... I can try.

Basically, both of the sets of symbols pair up. So we can use the same term for them. Let's call the players Player 1 and Player 2. They both get ABC,DEF, just in a different font or something. Player 1 puts in AAA,AAA; one of these is in the correct position. Just to show you how I'm listing the tiles, if they put in ABC,DEF it would look like this:

A B C
D E F

Player 2 then tries to divide and conquer the area to find out which position; they put AAABBB, if that gives a light then they know A was in one of the top 3, otherwise it's in the bottom 3. So then they divide that area in two and test some split. Let's say it was the top; they do ABB,BBB. If that gives a light, then they know A belongs in the first spot. Otherwise, it's in the next two; try each of those until they get it. Let's say Player 2 ends up with ABB,BBB.

Now Player 1 can try to find the location of B. They put in ABB,CCC. If another light turns on, then they know it's in the two remaining top spots. Otherwise, it's in the bottom 3. They then divide the area where it is into two parts and test, just like above, until they find it and submit. Let's say they arrive at ABC,CCC. (they're using the Cs as filler to prepare for the next search)

Now Player 2 can search for C's correct position. They try ABC,CDD. If that gives a third light, then it's in top right or bottom left. Continue dividing in two and testing until they find it.

... repeat until you get all of them ...

Does that make sense?

understood the first explanation, but this is a nice confirmation

I think others will get it, written like that

thanks rezecib

[Szyszyn]

Guys this lit up golden thing in the second puzzle looks like a chest.

Actually i've found this kind of chest in ruins, right next to the pseudoscience station. Sorry i don't have a screenshot... I didn't know what it was but it had 6 slots just like in the puzzle.

[fsp98]

1 hour ago, Szyszyn said:

Guys this lit up golden thing in the second puzzle looks like a chest.

Actually i've found this kind of chest in ruins, right next to the pseudoscience station. Sorry i don't have a screenshot... I didn't know what it was but it had 6 slots just like in the puzzle.

this is being discussed in the puzzle thread, you're supposed to fill the chest with items corresponding to your tile puzzle formation, and how many depends on the number of dents in the correct tiles

 

[Botaxalim]

i remote my computer friend using teamviewer, i play both side alone. more easy to manage, done this in less 10 min, first try made mistake and reset the game. second try finishe easily
all u need after changing position to click the button both side, and wait several second because server is checking sync both party

[Rasen_BR]

6 hours ago, rezecib said:

Um... I can try.

Basically, both of the sets of symbols pair up. So we can use the same term for them. Let's call the players Player 1 and Player 2. They both get ABC,DEF, just in a different font or something. Player 1 puts in AAA,AAA; one of these is in the correct position. Just to show you how I'm listing the tiles, if they put in ABC,DEF it would look like this:

A B C
D E F

Player 2 then tries to divide and conquer the area to find out which position; they put AAABBB, if that gives a light then they know A was in one of the top 3, otherwise it's in the bottom 3. So then they divide that area in two and test some split. Let's say it was the top; they do ABB,BBB. If that gives a light, then they know A belongs in the first spot. Otherwise, it's in the next two; try each of those until they get it. Let's say Player 2 ends up with ABB,BBB.

Now Player 1 can try to find the location of B. They put in ABB,CCC. If another light turns on, then they know it's in the two remaining top spots. Otherwise, it's in the bottom 3. They then divide the area where it is into two parts and test, just like above, until they find it and submit. Let's say they arrive at ABC,CCC. (they're using the Cs as filler to prepare for the next search)

Now Player 2 can search for C's correct position. They try ABC,CDD. If that gives a third light, then it's in top right or bottom left. Continue dividing in two and testing until they find it.

... repeat until you get all of them ...

Does that make sense?

 

Am I the only one who still doesn't understand this? I wish there was an accurate video...

[yujinthegamer]

20 minutes ago, Rasen_BR said:

Am I the only one who still doesn't understand this? I wish there was an accurate video...

if i understand correctly you and your partner mirror each other any move they make you make and vise versa you keep trying to move till the tiles are in the right spot the dots will glow yellow showing how close you are to what pieces are in the right place also something i see alot of people confused about is they think all 12 pieces go in a different position but no you need to treat you and your partners pieces as if it was one set and move them the same way i hope this maybe helps?

[rezecib]

@yujinthegamer That's a valid way to "stumble" on the solution, but it's not a very good way for always getting it.

The algorithm I was trying to explain in the quote involves one person keeping it static while the other person uses a divide-and-conquer approach to find the correct position for a particular symbol.

I'm not sure how else I can explain it better... Although I could elaborate more on how you get to the ABCDEF representation?

One player gets blocky symbols, and the other player gets letter-like symbols. The blocky symbols always have some duplicates, but for the purpose of the algorithm you should still label them separately. So A corresponds to both the first blocky symbol and the first letter symbol, B to the second of each, and so on. Like yujin said, A should be in the same position for both players in the final solution. But if you put A everywhere, then the other player can put A in one half to determine if A belongs in that half or the other half.

[yujinthegamer]

1 minute ago, rezecib said:

@yujinthegamer That's a valid way to "stumble" on the solution, but it's not a very good way for always getting it.

The algorithm I was trying to explain in the quote involves one person keeping it static while the other person uses a divide-and-conquer approach to find the correct position for a particular symbol.

I'm not sure how else I can explain it better... Although I could elaborate more on how you get to the ABCDEF representation?

One player gets blocky symbols, and the other player gets letter-like symbols. The blocky symbols always have some duplicates, but for the purpose of the algorithm you should still label them separately. So A corresponds to both the first blocky symbol and the first letter symbol, B to the second of each, and so on. Like yujin said, A should be in the same position for both players in the final solution. But if you put A everywhere, then the other player can put A in one half to determine if A belongs in that half or the other half.

i was just trying to write it as simple as i can from what i know and ya me and my partner did something similar

[Botaxalim]

simple word , its like sudoku game. both side use same move
carefull blocky side tend to overlap/overwrite, so i suggest let blocky side (person A) command and symbols side (person B) follow the command
everytime changing position remember to click the button both side
check the lamps till u find the correct position for each of blocks

[Kickapoo]

Can I post links here? I've made a video teaching how to pass the second phase.. Check it out here (I'm working on English subs)

METHEUS SOLVED PUZZLE

[rezecib]

@Kickapoo Nice video. You can make it a lot faster (use fewer tries) if instead of testing positions one by one you divide the remaining positions in two and fill one half.

Also, instead of filling the rest with an unused symbol, fill it with the next symbol-- although this gives only small gains, so is probably not necessary.

[Kickapoo]

8 minutes ago, rezecib said:

@Kickapoo Nice video. You can make it a lot faster (use fewer tries) if instead of testing positions one by one you divide the remaining positions in two and fill one half.

Also, instead of filling the rest with an unused symbol, fill it with the next symbol-- although this gives only small gains, so is probably not necessary.

This is a very good alternative. I'll try the next one, thank you! 

[Leonardo Cox]

Something seems to be off, my sister and I put our codes in (she put in my black code, I put in her yellow code) and for a split second the 2nd side lights up and then dissappears again.

[Rasen_BR]

1 minute ago, DarkKingBoo said:

Something seems to be off, my sister and I put our codes in (she put in my black code, I put in her yellow code) and for a split second the 2nd side lights up and then dissappears again.

 

That just happened to me...