PDA

View Full Version : Brute Forcing the Monolith Puzzle


SnapDragon
04-16-2012, 01:23 AM
Ok, well, there are rumblings around the 'Net that there isn't a proper solution to the Monolith puzzle, and the only way to solve it is to brute force. I created a spreadsheet to let us crowd-source the solution:
https://docs.google.com/spreadsheet/ccc?key=0ArBHYlyQxJXedHFubTU2N3N2eUZWY3Z5c0JPRnJ0M 1E
EDIT: Based on some (word-of-mouth) hints from Treyver we've created a different sequence to try, at https://docs.google.com/spreadsheet/ccc?key=0ArBHYlyQxJXedHFubTU2N3N2eUZWY3Z5c0JPRnJ0M 1E#gid=8

I made the following assumptions to bound the search space:
a) The sequence is 8 long
b) Only LT, RT, Up, Down, and A are used in the sequence
c) No moves appear twice in succession (many of the sequences in the game seem to satisfy this)

The sheet contains about 82,000 moves which, if done in succession, will try out every single possible sequence of this form. That's a lot for any one person to do, but maybe not for all of us. :)

If you want to contribute, follow these steps:
- Go to the Monolith room and MAKE SURE YOU ARE STANDING ON THE SQUARE WITH A LINE IN IT. Use first-person view, look down, and check from all 4 directions to make sure you're inside.
- Pick a starting row in the table, say "22221".
- Go to the row ABOVE that, and start entering in the sequence. (This is because the 7 moves preceding that row must also be entered.)
- Keep going until you get tired. Any time you make a mistake, be sure to restart the sequence at least 8 moves before your mistake.
- Note the last row you finished completely, say "22321".
- Go to the second sheet on the docs page, labeled "Claimed Rows". Anyone can edit this one, so be careful not to ruin it.
- Type your name in next to your starting row (eg "22221"). Copy (CTRL-C) your name and paste (CTRL-V) it all the way down to your ending row (eg "22321").
- Yay! You've now contributed. :) Your rows should be claimed on the spreadsheet.

Now, if we somehow manage to cover this entire spreadsheet and nothing has worked, I can create another one that removes assumption c) above. Removing assumption b) is probably not going to help, because 7^8 is a LOT of moves to try...

Z Force Clemson
04-16-2012, 01:26 AM
I have no idea if this will work or if your presumptions are even correct, but either way, I applaud your effort.

ll X ll
04-16-2012, 01:35 AM
Great idea. In a week's time we should have done most of this if not all.

Auburok
04-16-2012, 02:49 AM
I'll attempt to help with this in the next few days if the solution hasn't been found. Got class tomorrow, so I can't stay up doing this like I'd prefer to do. :/

Tsukibara
04-16-2012, 03:15 AM
Isn't there a faster way to check codes? When inputting 11221, you are trying every single combination of 1 and 2. You can use the same principle on this. We use 1, 2, 3, 4, 5 to represent A and the L and R buttons. And combination thereof in up to 8 characters. Uhm, watch this video because I can't explain it well. Can you crack the combination lock? - Solution - YouTube .

Fenrakk101
04-16-2012, 04:02 AM
Isn't there a faster way to check codes? When inputting 11221, you are trying every single combination of 1 and 2. You can use the same principle on this. We use 1, 2, 3, 4, 5 to represent A and the L and R buttons. And combination thereof in up to 8 characters. Uhm, watch this video because I can't explain it well. Can you crack the combination lock? - Solution - YouTube (http://www.youtube.com/watch?v=iPLQgXUiU14) .

Yay, someone else who watches the numberphile guys.

Tsukibara
04-16-2012, 04:12 AM
Yay, someone else who watches the numberphile guys.There's few people who seem to care about mathematics these days. It's really sad.

I came back to say the we might also potentially need to include not only L/R triggers and bumpers with A, but also the directional pad. Those were used in some of the input codes for anticubes.

Fenrakk101
04-16-2012, 04:25 AM
There's few people who seem to care about mathematics these days. It's really sad.

I came back to say the we might also potentially need to include not only L/R triggers and bumpers with A, but also the directional pad. Those were used in some of the input codes for anticubes.

We've figured this out already. We have RT, LT, A, up, and down. We can't use left and right because we need to stay in one spot.

SnapDragon
04-16-2012, 05:09 AM
There are 81,920 (5 * 4^7) possible combinations, and my sequence is 81,952 buttons long. So an "optimal" sequence could knock 20 off that number. Big whoop. :)

Fenrakk101
04-16-2012, 05:12 AM
There are 81,920 possible combinations, and my sequence is 81,952 buttons long. So an "optimal" sequence could knock 20 off that number. Big whoop. :)

There are six unique letters in PHIL FISH, actually...

ImAlann
04-16-2012, 06:09 AM
maybe could be in a new DLC lol

Fenrakk101
04-16-2012, 06:52 AM
maybe could be in a new DLC lol

...no.

Someone already figured it out, but won't share.

JetpackJeezus
04-17-2012, 06:10 AM
You may want to reconsider assumption c based on a hint from treyher:

How goes the brute forcing? I'm surprised I haven't seen more organization around dividing up the segments.

I've been working on backsolving the code based on what the community has been doing but haven't seen anything that works yet. However there's a bunch of words above that I haven't tried. Can you narrow them down to words that start with two letters which are rotations of one another? That ought to reduce the search time further.

If you assume that the word having two letters that are rotations of each other means that they would be the same input, then the current list wouldn't contain the answer.

If there is a way to create a similar spreadsheet that takes into account double input and (if possible) the double input would be at the start, I would personally help test a couple hundred of those rows.

SnapDragon
04-17-2012, 08:18 AM
Where did he post this? Without any context, I don't understand why you think that implies that the code begins with a repeated move. If the current effort finishes without a hit, I'll do another spreadsheet for the remaining 300k sequences that have a repeated move. Right now, I'm just trying to target the search so there's a higher chance we'll hit it early.

We need more people to contribute. :) I'm surprised more people aren't jumping at the chance to be "the one" who discovered the code! It only takes like 10-15 minutes to enter 1,000 symbols.

mas4mune
04-17-2012, 08:27 AM
Where did he post this? Without any context, I don't understand why you think that implies that the code begins with a repeated move. If the current effort finishes without a hit, I'll do another spreadsheet for the remaining 300k sequences that have a repeated move. Right now, I'm just trying to target the search so there's a higher chance we'll hit it early.

We need more people to contribute. :) I'm surprised more people aren't jumping at the chance to be "the one" who discovered the code! It only takes like 10-15 minutes to enter 1,000 symbols.
i'd be more than happy to if i we had a concrete way of doing it. this is all theory based on conflicting hearsay.

Nquoid
04-17-2012, 09:31 AM
If we are going to brute force this thing, then I think that at least 2 people should check each sequence. A hassle, but at least then we'll know that no mistakes were made and we won't slap ourselves that the code was actually in the early 2000s on the spreadsheet but the person didn't quite press the Up button hard enough.

mas4mune
04-17-2012, 09:34 AM
If we are going to brute force this thing, then I think that at least 2 people should check each sequence. A hassle, but at least then we'll know that no mistakes were made and we won't slap ourselves that the code was actually in the early 2000s on the spreadsheet but the person didn't quite press the Up button hard enough.
a good idea. especially since the RT/LT movements fail if you go too fast. who knows if the game still registers the button hit or not.

DEADSUN PRIME
04-17-2012, 09:57 AM
There's few people who seem to care about mathematics these days. It's really sad.

I came back to say the we might also potentially need to include not only L/R triggers and bumpers with A, but also the directional pad. Those were used in some of the input codes for anticubes.
considering they are playing games on that very math is kinda depressing.

madninjaskillz
04-17-2012, 12:45 PM
As much as it saddens me to go down that route, I believe i can automate the bruteforce. I have an FPS adapter that lets me use a keyboard as a gamepad - and i have a teensy+ which is a little programmable chip that can pretend to be a keyboard. the combo of the two means a fully automatable gamepad
to make something to battle out every 64 bit code with lt+rt wouldnt be hard. ignoring rolling duplicates, if my maths is right its 2 to the power of 64 possible binary combos, so 9223372036854775808 key presses.

Those of you still following will realise that
a) ill be dead before it finishes.
b) it wont tell us which code worked.

however, if someone wants to give me a smaller vector to attack, im willing to try anything - but again, we wont know what the code is, just that yes, its a code and yes, we entered it in the right place.

Auburok
04-17-2012, 01:11 PM
Isn't there a faster way to check codes? When inputting 11221, you are trying every single combination of 1 and 2. You can use the same principle on this. We use 1, 2, 3, 4, 5 to represent A and the L and R buttons. And combination thereof in up to 8 characters. Uhm, watch this video because I can't explain it well. Can you crack the combination lock? - Solution - YouTube (http://www.youtube.com/watch?v=iPLQgXUiU14) .

I don't see you rushing to create every two button combination in an Eulerian path. :p

The result would be just as fast as what he has there, though. This guy's point is that figuring out the paths is efficient. That part's already done. He's not suggesting that using an Eulerian path will somehow make putting the combinations in the the lock physically quicker. The end result is similar to what's in the GoogleDoc, if you haven't read it, except on steroids.

the combo of the two means a fully automatable gamepad
to make something to battle out every 64 bit code with lt+rt wouldnt be hard. ignoring rolling duplicates, if my maths is right its 2 to the power of 64 possible binary combos, so 9223372036854775808 key presses.

Those of you still following will realise that
a) ill be dead before it finishes.
b) it wont tell us which code worked.

however, if someone wants to give me a smaller vector to attack, im willing to try anything - but again, we wont know what the code is, just that yes, its a code and yes, we entered it in the right place.

Where are you getting the 64? There is a smaller vector to attack. Look at the GoogleDoc. Read his instructions at the top post.

madninjaskillz
04-17-2012, 01:14 PM
Where are you getting the 64?

the 64bit name of god stuff.

Auburok
04-17-2012, 01:15 PM
the 64bit name of god stuff.

The input code is apparently 7-8 button inputs long. Not 64. The answer isn't derived from in game stuff, as in the name of god in 64 bit.

madninjaskillz
04-17-2012, 01:38 PM
ignoring 7 (as will get done by accident during) at 8 inputs - if we assume up+down+lt+rt+A (as no left or right) thats 32k combos. now we have been told its 7, which is 5^7 - 16807.

If i play it safe and do one button every 500 ms, thats under 2 and a half hours to try every code. If this time tomorrow, we have not sussed it, ill try it (and share the code i use so others can do it.) (ill probably make it do the 8 digit long as its only 2 hours longer and will do both sets)

Nquoid
04-17-2012, 01:47 PM
ignoring 7 (as will get done by accident during) at 8 inputs - if we assume up+down+lt+rt+A (as no left or right) thats 32k combos. now we have been told its 7, which is 5^7 - 16807.

If i play it safe and do one button every 500 ms, thats under 2 and a half hours to try every code. If this time tomorrow, we have not sussed it, ill try it (and share the code i use so others can do it.) (ill probably make it do the 8 digit long as its only 2 hours longer and will do both sets)

Only way we'll get it solved is if people actually try with the spreadsheet, but at the moment only about 5-6 people have actually bothered to put there names down, but there seem to be quite a few lurkers based on number of viewers.

madninjaskillz
04-17-2012, 02:04 PM
i have had nothing but bad experience with google docs so i avoided it. so to prove i want to help rather than derail your thread, i thought id take on a hundred or so....it wont let me write in the 'tried by' column.

madninjaskillz
04-17-2012, 02:08 PM
i sussed it out, but i cant claim it AND look at the keys. thats hella awkard

Mechanikill
04-17-2012, 02:13 PM
I'll rattle through as many as I can tonight after work! A few hours worth if I get chance!

Hexahedronaut
04-17-2012, 03:12 PM
Just got through 100.. and nothing :(

JetpackJeezus
04-17-2012, 04:04 PM
Where did he post this? Without any context, I don't understand why you think that implies that the code begins with a repeated move. If the current effort finishes without a hit, I'll do another spreadsheet for the remaining 300k sequences that have a repeated move. Right now, I'm just trying to target the search so there's a higher chance we'll hit it early.

We need more people to contribute. :) I'm surprised more people aren't jumping at the chance to be "the one" who discovered the code! It only takes like 10-15 minutes to enter 1,000 symbols.

Sorry. The context is that treyher is a guy who was supposedly given the code from the developers. He has made a couple posts int the GameFaqs forums (like the one I quoted) that made it seem like he was trying to take the code and work backwards to derive what the solution is. It follows certain logic that he would only be looking for words that that started with letters that were rotations of themselves (in Fez language) if he knew that the first two inputs for the code were the same.

Another post that might explain it better: http://www.xbox360achievements.org/forum/showpost.php?p=5019062&postcount=374

Like I said, I would love to help with the brute force, but I honestly don't think the code is within the current list. I could be wrong, but just trying to put in my two cents. :)

Nquoid
04-17-2012, 04:31 PM
Sorry. The context is that treyher is a guy who was supposedly given the code from the developers. He has made a couple posts int the GameFaqs forums (like the one I quoted) that made it seem like he was trying to take the code and work backwards to derive what the solution is. It follows certain logic that he would only be looking for words that that started with letters that were rotations of themselves (in Fez language) if he knew that the first two inputs for the code were the same.

Another post that might explain it better: http://www.xbox360achievements.org/forum/showpost.php?p=5019062&postcount=374

Like I said, I would love to help with the brute force, but I honestly don't think the code is within the current list. I could be wrong, but just trying to put in my two cents. :)

Thing is, if it is merely a 7 button code (with only 5 inputs) then that means the correct combination must be in that list somewhere. Unless we go back to the binary solution and it only uses LT and RT which I highly doubt.

madninjaskillz
04-17-2012, 05:11 PM
but where do we type this code? we assume the other circle center - but why?

Nquoid
04-17-2012, 05:25 PM
but where do we type this code? we assume the other circle center - but why?

Because that's the half of the map that's been burnt off. And the answer is more than likely in this room because every other room has become gold by this point. Meaning that the code needs to be entered on the opposite circle

mas4mune
04-17-2012, 05:32 PM
in theory.

madninjaskillz
04-17-2012, 06:43 PM
i dont get using theory to tell us WHERE to type the code - but then do an organised version of closing ur eyes and bashing it till it does what you want. if we are going to use clues to work out the where, why not the what?

Nquoid
04-17-2012, 06:50 PM
because the only clues we have amount to the number 7 (or 8) and possibly the 64bit name of God. As far as we've been told by the actual guy who finished, he was told by the developer and there is no way to figure out the code in game. Therefore, we're doing this based on the assumption that we don't have anyway of finding the code. He even commented on the fact that the brute force effort hasn't really gotten off the ground, almost like it might actually work.

Kibblre
04-17-2012, 07:25 PM
So, the spreadsheet is read only for me. I want to add to the effort, but I can't put my name in any rows.

johnnyspliff
04-17-2012, 07:30 PM
See at the bottom of your screen? Where it says "Sequence Row (Read Only)" and "Claimed Rows"? Click on claimed rows, and enter your name on the row numbers you've tried.

Kibblre
04-17-2012, 07:32 PM
Ah, I assumed I would just put it on the sheet I was looking at. I've never used Google docs spreadsheets.

SnapDragon
04-17-2012, 08:47 PM
Yeah, it's a bit confusing, sorry. I just did it that way because I was worried about people mucking with the sequence. :)

Kibblre
04-17-2012, 09:12 PM
I managed to knock out some, but I got distracted by the clock room since my last to anticubes are in there.

valcfield
04-17-2012, 10:42 PM
sooo am i crazy or does this brute force not have any repeated inputs? i haven't come across any and i've done a lot, and control f doesn't seem to see them (i tried control f'ing "dn rt" and that worked fine...)... kind of a problem since trayher claims the sequence starts off with a repeated button? hopefully i'm just crazy.

Scapetti
04-17-2012, 10:45 PM
sooo am i crazy or does this brute force not have any repeated inputs? i haven't come across any and i've done a lot, and control f doesn't seem to see them (i tried control f'ing "dn rt" and that worked fine...)... kind of a problem since trayher claims the sequence starts off with a repeated button? hopefully i'm just crazy.

You're not crazy and a lot of people have wasted their time with this I am sorry to say... this spreadsheet desperately needs to be updated due to new information! 7 inputs, beginning with two the same

valcfield
04-17-2012, 10:47 PM
the 7 part i'm less worried about- the spreadsheet is nice by optimizing into one long chain, no reason to treat it as discrete codes. the double part though is disconcerting. and, also, i have grains of salt for any info we've gotten so i don't think its a waste of time per se. i do think it would be optimal to create a new spreadsheet that capitalizes on the double input though

zingrook
04-17-2012, 10:56 PM
http://i1260.photobucket.com/albums/ii563/zingrook/Decoder-1.png

Kibblre
04-17-2012, 11:35 PM
I think someone is trolling our spreadsheet. There's a chunk claimed by "Polytron Corporation".

valcfield
04-17-2012, 11:37 PM
ha is that trolling? obviously not me but i thought that was amusing, not misanthropic...

Kibblre
04-17-2012, 11:38 PM
There's also "Phillip Fisherman".

MrMiyamoto
04-17-2012, 11:40 PM
I want to help brute-forcing, but I don't understand how the numbers map to button presses.

So, if I wanted to do the line 44041, I understand starting one line before, but after that what numbers map to what buttons?

There's five possibilities, right? So something like U = 1, D = 2, LB = 3, RB = 4, and A = 5 would be a key, but then what do I press when I reach a 6, 7, or 8? Is there a standard key we're supposed to be using that isn't included in the original post?

Kibblre
04-17-2012, 11:42 PM
Just start at the row above it and just press the buttons as you read them until you get tired/bored.

MrMiyamoto
04-17-2012, 11:43 PM
Just start at the row above it and just press the buttons as you read them until you get tired/bored.

Right, but which buttons for which numbers?

Kibblre
04-17-2012, 11:44 PM
The buttons are laid out plainly, I don't know what you're talking about.

MrMiyamoto
04-17-2012, 11:46 PM
The buttons are laid out plainly, I don't know what you're talking about.

Well, for example, I don't have a button labelled "8" on my Xbox controller.

SnapDragon
04-17-2012, 11:47 PM
Sigh. People have taken an unreliable deduction (rotated letters = same inputs) from an unreliable source (trayher isn't really trying to help us) and built it up into a certainty in their minds. This is a pretty common psychological effect: the more you repeat unreliable information to yourself, the more certain you become, even without any outside confirmation. :(

For some interesting reading, check out http://en.wikipedia.org/wiki/Masquerade_(book):
Tens of thousands of letters from Masqueraders have convinced me that the human mind has an equal capacity for pattern-matching and self-deception. While some addicts were busy cooking the riddle, others were more single-mindedly continuing their own pursuit of the hare quite regardless of the news that it had been found. Their own theories had come to seem so convincing that no exterior evidence could refute them.

But ok, in this case it's pretty harmless to act on this information; we might as well start searching all patterns with repeated inputs, too. (For technical reasons, it doesn't save much time to narrow the search to only patterns STARTING with repeated inputs; we might as well try them all while we're at it.) There are around 300k of them. I'll add them tonight... assuming the spreadsheet can handle it. :)

Kibblre
04-17-2012, 11:47 PM
Did you click the "Sequence Rows" sheet on the bottom? There's actual buttons on that page.

SnapDragon
04-17-2012, 11:48 PM
Well, for example, I don't have a button labelled "8" on my Xbox controller.

Are you on the right sheet? There's one editable sheet (with 2 columns) for contributors, and another one (with 3 columns) that lists the actual sequence to press.

MrMiyamoto
04-17-2012, 11:48 PM
Did you click the "Sequence Rows" sheet on the bottom? There's actual buttons on that page.

Okay, thanks. That's all I needed to know.

valcfield
04-17-2012, 11:49 PM
thanks man- and based on the clever set up i understand why it makes sense not to change it to starting with them... fyi, there have been repeated inputs in other sequences right? not many, but there hasn't been some categorical rule against them that would mean it couldn't be there?

Kibblre
04-17-2012, 11:51 PM
There's one sequence in what I called "caterpillar room" that is a tetromino code that can only be seen in first person. The button combo is 4 sets of repeats (L L D D R R U U).

SnapDragon
04-17-2012, 11:54 PM
I think someone is trolling our spreadsheet. There's a chunk claimed by "Polytron Corporation".

LOL, that's awesome. I see somebody who knows spreadsheets better than me also added a "% completed" count... thanks! :Bounce:

Scapetti
04-18-2012, 12:18 AM
LOL, that's awesome. I see somebody who knows spreadsheets better than me also added a "% completed" count... thanks! :Bounce:

soundofjw did that, he also confirmed from treyher (through twitter) that we should be expecting repeating inputs... sad times

Kibblre
04-18-2012, 12:23 AM
As of this post, we're roughly 62% done with the current spreadsheet. Can you start making one for us that contains all (and possibly only if you can do it) the doubles so we can start testing those?

johnnyspliff
04-18-2012, 12:47 AM
Treyher just posted on GameFAQ that the code might have multiple repeated inputs.
http://www.gamefaqs.com/boards/961239-fez/62566926?page=10

Kibblre
04-18-2012, 01:15 AM
New spreadsheet with new constraints.

https://docs.google.com/spreadsheet/ccc?key=0AlO_qKx0yN5FdDQ2MHl1cXhobkJSd0hIeHlzbnhvT FE#gid=3

SnapDragon
04-18-2012, 03:24 AM
Ugh. Somebody deleted the sequence off that spreadsheet. I've added two new sheets (one read-only) to the original spreadsheet for the newly-constrained sequence, based on Treyher's hints. It's only 52,000 buttons long, so it shouldn't take us long to complete!
https://docs.google.com/spreadsheet/ccc?key=0ArBHYlyQxJXedHFubTU2N3N2eUZWY3Z5c0JPRnJ0M 1E#gid=8

JetpackJeezus
04-18-2012, 04:27 AM
Ugh. Somebody deleted the sequence off that spreadsheet. I've added two new sheets (one read-only) to the original spreadsheet for the newly-constrained sequence, based on Treyher's hints. It's only 52,000 buttons long, so it shouldn't take us long to complete!
https://docs.google.com/spreadsheet/ccc?key=0ArBHYlyQxJXedHFubTU2N3N2eUZWY3Z5c0JPRnJ0M 1E#gid=8

Finally! Anyway, I'll help work on this new spreadsheet tomorrow.

Kibblre
04-18-2012, 04:42 AM
Finally! Anyway, I'll help work on this new spreadsheet tomorrow.

Someone deleted the free sheets of the most recent document as well.

SnapDragon
04-18-2012, 09:03 AM
Grr. There's always some asshole. I recovered the sheets. If this happens again, just keep track of a range on your own and enter it in later.

Kibblre
04-18-2012, 05:24 PM
Snap, check this out. I made a private one where people need to be logged in to access it, so when someone changes it I can recover and boot them.

http://www.xbox360achievements.org/forum/showthread.php?t=366675