Thread Rating:

Poll

25 votes (49.01%)
16 votes (31.37%)
7 votes (13.72%)
4 votes (7.84%)
12 votes (23.52%)
3 votes (5.88%)
6 votes (11.76%)
5 votes (9.8%)
12 votes (23.52%)
10 votes (19.6%)

51 members have voted

charliepatrick
charliepatrick
  • Threads: 39
  • Posts: 3099
Joined: Jun 17, 2011
Thanked by
Wizard
August 9th, 2025 at 1:45:00 PM permalink
The logic I used was N210...010...0 works if you know how many zeroes there are, so there is one N. two 1s, and one 2.
6210001000.
ThatDonGuy
ThatDonGuy 
  • Threads: 128
  • Posts: 7270
Joined: Jun 22, 2011
Thanked by
Wizard
August 9th, 2025 at 2:34:13 PM permalink
Quote: Wizard

Can you find the one 10-digit "autobiographic number." Such a number describes itself with the number in each position, starting with 0, equaling the count of that digit in the whole number.
link to original post


To be fair, I have heard this one before - the number is 6210001000
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 9th, 2025 at 3:28:33 PM permalink
Quote: charliepatrick

The logic I used was N210...010...0 works if you know how many zeroes there are, so there is one N. two 1s, and one 2.
6210001000.

link to original post



I agree!
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 10th, 2025 at 3:32:29 PM permalink
Quote: Wizard

This one is an old classic that may have been asked before.

Can you find the one 10-digit "autobiographic number." Such a number describes itself with the number in each position, starting with 0, equaling the count of that digit in the whole number.

For example, 1210 is one because:

There is 1 in in the 0 position and there is 1 numeral 0 in the entire number.
There is 2 in in the 1 position and there are 2 numeral 1's (that's an improper use of an apostrophe, Wiz) in the entire number.
There is 1 in in the 2 position and there is 1 numeral 2 in the entire number.
There is 0 in in the 3 position and there are 0 numeral 3's in the entire number.

Yes, the number is in base-10.
link to original post



I think this works:


6210001000
6x0, 2x1 1x2, 1x6


Is the answer really unique?

ETA: I guess I'm late with this answer. But I'm curious to see a uniqueness proof.
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 10th, 2025 at 7:09:42 PM permalink
Quote: SkinnyTony

ETA: I guess I'm late with this answer. But I'm curious to see a uniqueness proof.
link to original post



The source I got it from indicated it was the only 10-digit answer. I think I could prove no other first digit works.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
charliepatrick
charliepatrick
  • Threads: 39
  • Posts: 3099
Joined: Jun 17, 2011
August 11th, 2025 at 6:32:03 AM permalink
There are two equations that I can see need to be satisified.
D: 0 1 2 3 4 5 6 7 8 9
N: 6 2 1 0 0 0 1 0 0 0
P: 0 2 2 0 0 0 6 0 0 0
D are the digits 0 thru 9. N are the number of times each digit appears. P is the product of D*N.
Note that the sum of Row N is 10 because it says how often each number appears, and so there are ten numbers in total.
Working along the N row , there will be six 0's, then two 1's, then one 2 etc. This also has to total to the ten digits. Hence the sum of the products, as in Row P, also has to be 10.
I suspect there's some kind of contradictory proof, for instance there cannot be 5 zeroes as that would mean there's a 1 under 5 and a total of 4 under numbers 1-4, so that the product is 5: 5xyz010000 etc. don't work.
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 11th, 2025 at 1:23:35 PM permalink
I'll call this one the Tunnel of Unknown Length Puzzle.

You are trapped in a circular tunnel of unknown length. Equally spaced apart are at least two switches, which can be either up or down. You may flip the switches as much as you like. You may not leave any markings other than adjusting the switches. How can you tell with certainty how many switches are in the tunnel?
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 11th, 2025 at 1:37:07 PM permalink
Quote: Wizard

I'll call this one the Tunnel of Unknown Length Puzzle.

You are trapped in a circular tunnel of unknown length. Equally spaced apart are at least two switches, which can be either up or down. You may flip the switches as much as you like. You may not leave any markings other than adjusting the switches. How can you tell with certainty how many switches are in the tunnel?
link to original post



For i = 2 to infinity:
flip current switch up
walk cw i switches
flip current switch down
walk ccw i switches
if current switch is down return i
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 11th, 2025 at 2:31:29 PM permalink
Quote: SkinnyTony


For i = 2 to infinity:
flip current switch up
walk cw i switches
flip current switch down
walk ccw i switches
if current switch is down return i

link to original post



I agree. Welcome to the forum and to this thread in particular! We could use some new blood.

BTW, I think I should have said there are at least two switches in the up position.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
Thanked by
MichaelBluejay
August 11th, 2025 at 3:30:43 PM permalink
I think I'll make this one a newsletter puzzle. Here is my revised wording.

You are trapped in a circular tunnel. At regular intervals are switches. The switches are far apart enough to be unable to see two at once. The switches can be in either position, except you're told at least two are in the "on" position. You may switch the switches as much as you wish. Otherwise, you may not leave markings of any kind. How can you determine with certainty how many switches are in the tunnel?


The following is not the most efficient solution, but probably the easiest to explain.

1. Go either direction until you arrive at the first switch. If it is in the OFF position, turn it ON.
2. Walk clockwise until you get to the first switch in the ON position. Then turn it OFF. Keep track of how many switches you travel (not counting the starting switch).
3. Turn back the opposite direction the same number of switches from step 2.
4. If you encounter the switch you return to in the ON position, then go back to step 2.
5. If you encounter the switch you return to in the OFF position, then the number of switches equals the number of switches you counted from step 2.


I'm open to any challenges of my solution or comments if it wasn't clear.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 12th, 2025 at 12:44:30 AM permalink
Thanks!

Does the starting state (at least two switches up) matter? Is there a nicer (faster, less brute force) solution?
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 12th, 2025 at 5:28:49 PM permalink
This comes from a Japanese puzzle book. I searched and didn't find it posted here before.

You're mailing a package to a recipient in another country. Because the mail system is corrupt, the contents of any unlocked package will be stolen. You can definitely prevent contents theft by putting a lock on the box. However, the recipient can't open your lock without the key. How can you get the contents to the recipient?

Here are clarifications about thingsI wondered about:
• There is nothing special about the lock or the key. It's not a combination lock. It's not a puzzle lock. The recipient has no ability to manufacture a key.
• You cannot carry the package to the recipient yourself nor use a courier.

My whole family failed this one. I'm not a good problem-solver. I imagine forum regulars will figure it out quickly, though.

[edited for clarity]
Last edited by: MichaelBluejay on Aug 13, 2025
I run Easy Vegas ( https://easy.vegas )
AutomaticMonkey
AutomaticMonkey 
  • Threads: 14
  • Posts: 865
Joined: Sep 30, 2024
August 12th, 2025 at 6:04:54 PM permalink
Quote: MichaelBluejay

This comes from a Japanese puzzle book. I searched and didn't find it posted here before.

You're mailing a package to a recipient in a corrupt country. If the package is unlocked, it's likely the contents will be stolen. You can definitely prevent contents theft by putting a lock on the box. However, the recipient can't open your lock without the key either. How can you get the contents to the recipient?

Here were the clarifications I wondered about:
• If you mail the key separately in an unlocked package, you can presume the key will be stolen.
• There is nothing special about the lock or the key. It's not a combination lock. It's not a puzzle lock. The recipient has no ability to manufacture a key.
• You cannot carry the package to the recipient yourself nor use a courier.

My whole family failed this one. I'm not a good problem-solver. I imagine forum regulars will figure it out quickly, though.
link to original post



Is there communication between the two parties?

The wording of the problem implies it is only the corruption in my recipient's country that will result in theft, not mine, and it doesn't specify that the theft will happen with outbound packages. So I would have my recipient send me an unlocked padlock to which he already has a key.
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 12th, 2025 at 7:04:33 PM permalink
Clarifying AutomaticMonkey's question / idea.

The two parties can communicate as much as they like. I hear they're partial to WhatsApp.

Since AM put an idea in a spoiler (which doesn't work), I'll clarify that one in a spoiler.

If a package is sent in either direction, it's still handled by the mail service in each country, and subject to theft in the corrupt country. So, the customer cannot send a padlock and a key in an unlocked package to the seller, because the contents (the key and the lock) would be stolen.
I run Easy Vegas ( https://easy.vegas )
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 12th, 2025 at 7:14:38 PM permalink
I asked ChatGPT and it got the right answer.

God. Damn.

About 20 years ago when IBM's early AI "Watson" played as a contestant on Jeopardy and won, I told the Wizard that that marked the beginning of the end. He replied, "When the robot overlords come for me, my dying words will be, 'Bluejay was right.' "
I run Easy Vegas ( https://easy.vegas )
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 12th, 2025 at 8:04:02 PM permalink
Quote: MichaelBluejay

You're mailing a package to a recipient in a corrupt country. If the package is unlocked, it's likely the contents will be stolen. You can definitely prevent contents theft by putting a lock on the box. However, the recipient can't open your lock without the key either. How can you get the contents to the recipient?
link to original post




1. Put your own lock on the box and mail it to the recipient. Hold onto the key.
2. Recipient adds his own lock to the box, keeping the key, and mails the box back to you.
3. You take off your own lock and mail it back to recipient.
4. Recipient removes his own lock and reveals the contents.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 12th, 2025 at 8:20:31 PM permalink
Wizard got it. Of course.

ChatGpt: 1
Wizard: 1
AutomaticMonkey: 0.5
Bluejay: 0
I run Easy Vegas ( https://easy.vegas )
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 1:29:24 AM permalink
Quote: MichaelBluejay

I asked ChatGPT and it got the right answer.

God. Damn.

About 20 years ago when IBM's early AI "Watson" played as a contestant on Jeopardy and won, I told the Wizard that that marked the beginning of the end. He replied, "When the robot overlords come for me, my dying words will be, 'Bluejay was right.' "
link to original post



That's not so surprising. It's just an advanced search engine plus a very good language engine.

ChatGPT didn't figure out the answer. It found the answer online somewhere. The impressive part (which ChatGPT is very good at) is that it was able to interpret your question and figure out which problem it was referring to.

But it didn't solve it in the way that you would expect an intelligent entity to. It has no intelligence. It just looked the answer up. It doesn't understand the question or the answer.
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 1:38:41 AM permalink
Quote: MichaelBluejay

This comes from a Japanese puzzle book. I searched and didn't find it posted here before.

You're mailing a package to a recipient in a corrupt country. If the package is unlocked, it's likely the contents will be stolen. You can definitely prevent contents theft by putting a lock on the box. However, the recipient can't open your lock without the key either. How can you get the contents to the recipient?

Here were the clarifications I wondered about:
• If you mail the key separately in an unlocked package, you can presume the key will be stolen.
• There is nothing special about the lock or the key. It's not a combination lock. It's not a puzzle lock. The recipient has no ability to manufacture a key.
• You cannot carry the package to the recipient yourself nor use a courier.

My whole family failed this one. I'm not a good problem-solver. I imagine forum regulars will figure it out quickly, though.
link to original post



I haven't read the Wizard's answer. I'm a little unsure about what will and won't be stolen though. A key will be stolen (despite it being worthless without the lock that it opens). A lock without a key (also worthless) will also be stolen. But a locked package (worthless without the key) will not be.

So my question is, how do I determine what will and won't be stolen? It seems kind of random. Is anything that's not specifically a locked box stolen?
Dieter
Administrator
Dieter
  • Threads: 17
  • Posts: 6945
Joined: Jul 23, 2014
August 13th, 2025 at 1:42:29 AM permalink
Quote: SkinnyTony


But it didn't solve it in the way that you would expect an intelligent entity to. It has no intelligence. It just looked the answer up. It doesn't understand the question or the answer.
link to original post



Most people just want an adequately correct answer, and don't care how it is achieved. If they could look up all the answers to life's problems in the back of the teacher's book, they probably would.

That approach to problems certainly leaves more time to watch reality television.
May the cards fall in your favor.
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 4:55:18 AM permalink
Quote: Dieter

Quote: SkinnyTony


But it didn't solve it in the way that you would expect an intelligent entity to. It has no intelligence. It just looked the answer up. It doesn't understand the question or the answer.
link to original post



Most people just want an adequately correct answer, and don't care how it is achieved. If they could look up all the answers to life's problems in the back of the teacher's book, they probably would.

That approach to problems certainly leaves more time to watch reality television.
link to original post



To be clear, I'm not questioning ChatGPT's usefulness. I'm questioning its intelligence (artificial or otherwise).

It's severely limited in that it can only answer questions that have already been answered (as opposed to an actually intelligent entity, who is capable of reasoning and answering new problems). So, maybe this is also about how useful it is.
AnotherBill
AnotherBill
  • Threads: 1
  • Posts: 28
Joined: Jul 27, 2025
August 13th, 2025 at 5:01:32 AM permalink
Quote: SkinnyTony

To be clear, I'm not questioning ChatGPT's usefulness. I'm questioning its intelligence (artificial or otherwise).

It's severely limited in that it can only answer questions that have already been answered (as opposed to an actually intelligent entity, who is capable of reasoning and answering new problems). So, maybe this is also about how useful it is.
link to original post



ChatGPT is less that 3 years old. It's still a baby.
AnotherBill
AnotherBill
  • Threads: 1
  • Posts: 28
Joined: Jul 27, 2025
August 13th, 2025 at 5:15:44 AM permalink
I've seen this puzzle great many years ago in some printed book. (I asked ChatGPT to translate and adapt it into English (from memory).)

Problem:
You have 3 burger patties and a frying pan that can hold only 2 patties at a time.
Each side of a patty takes 1 minute to fry.
You need to fry both sides of each patty.

Question:
What is the minimum time required to fully fry all the patties?
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 5:54:24 AM permalink
Here's one that I got from a YouTube video.

We are going to roll a single 6-sided die repeatedly until a 6 is rolled, and keep track of how many rolls it takes.

So for example, if you rolled a 2, 3, 1, 6, that would be 4 rolls. If you rolled a 6 on the first roll, that would be 1 roll.

What is the average number of rolls needed, conditioned on the fact that all rolls are an even number? In other words, if we throw out all sequences that contain an odd number, what's the average length of the remaining sequences?

So for example, 2, 2, 6 would be 3 rolls. 2, 2, 3, 6 would not count since it includes an odd number (3). Similarly 2, 1, 3, 6 would not count.
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 6:25:13 AM permalink
Quote: AnotherBill

I've seen this puzzle great many years ago in some printed book. (I asked ChatGPT to translate and adapt it into English (from memory).)

Problem:
You have 3 burger patties and a frying pan that can hold only 2 patties at a time.
Each side of a patty takes 1 minute to fry.
You need to fry both sides of each patty.

Question:
What is the minimum time required to fully fry all the patties?
link to original post



3 minutes

Number the patties 1, 2, 3, and label the sides A and B. Then fry:

1st minute: 1A & 2A
2nd minute: 1B & 3A
3rd minute: 2B & 3B
AnotherBill
AnotherBill
  • Threads: 1
  • Posts: 28
Joined: Jul 27, 2025
August 13th, 2025 at 6:32:50 AM permalink
Quote: SkinnyTony

3 minutes

Number the patties 1, 2, 3, and label the sides A and B. Then fry:

1st minute: 1A & 2A
2nd minute: 1B & 3A
3rd minute: 2B & 3B

link to original post



Well done.
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 13th, 2025 at 6:42:12 AM permalink
Quote: SkinnyTony

That's not so surprising. [AI] is just an advanced search engine plus a very good language engine. ChatGPT didn't figure out the answer. It found the answer online somewhere. link to original post

I'm not so sure about that. From what I've read, AI can indeed solve original problems without having seen a solution.

Quote: SkinnyTony

I'm a little unsure about what will and won't be stolen though....It seems kind of random.

(1) The contents of an unlocked package will always be stolen.

(2) The contents of a locked package will never be stolen.

(3) A locked package itself will never be stolen.
I run Easy Vegas ( https://easy.vegas )
AnotherBill
AnotherBill
  • Threads: 1
  • Posts: 28
Joined: Jul 27, 2025
August 13th, 2025 at 6:59:29 AM permalink
Quote: MichaelBluejay

(1) The contents of an unlocked package will always be stolen.
link to original post



Or, the contents of an unlocked package may be stolen but you trust no one and don't want to take any risk :)
AnotherBill
AnotherBill
  • Threads: 1
  • Posts: 28
Joined: Jul 27, 2025
August 13th, 2025 at 7:03:35 AM permalink
Quote: SkinnyTony

What is the average number of rolls needed, conditioned on the fact that all rolls are an even number? In other words, if we throw out all sequences that contain an odd number, what's the average length of the remaining sequences?
link to original post



I'm not a native speaker, but isn't there any contradiction?
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 13th, 2025 at 7:22:23 AM permalink
Quote: SkinnyTony

Here's one that I got from a YouTube video.

We are going to roll a single 6-sided die repeatedly until a 6 is rolled, and keep track of how many rolls it takes.

So for example, if you rolled a 2, 3, 1, 6, that would be 4 rolls. If you rolled a 6 on the first roll, that would be 1 roll.

What is the average number of rolls needed, conditioned on the fact that all rolls are an even number? In other words, if we throw out all sequences that contain an odd number, what's the average length of the remaining sequences?

So for example, 2, 2, 6 would be 3 rolls. 2, 2, 3, 6 would not count since it includes an odd number (3). Similarly 2, 1, 3, 6 would not count.
link to original post



I'm having some trouble understanding what is being asked. Wouldn't 2-3-1-6 not count, since there is an odd number in there?

Could this be reworded to say that if an odd number precedes the six, then that six doesn't end rolling, but resets the roll history?

For example 2-3-1-6-1-2-3-4-5-6-2-4-6 would be an example of a concluded trial. Note how there are all even number between the last and second to last 6.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 13th, 2025 at 7:26:48 AM permalink
Quote: SkinnyTony

Thanks!

Does the starting state (at least two switches up) matter? Is there a nicer (faster, less brute force) solution?
link to original post



I don't think I needed the condition about two up switches, but instead could have just said any two switches. Let me think about it.

Also, instead of each time going to the next up switch, you could double the number of switches you passed the last time.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 7:47:28 AM permalink
Quote: Wizard

Quote: SkinnyTony

Similarly 2, 1, 3, 6 would not count.



I'm having some trouble understanding what is being asked. Wouldn't 2-3-1-6 not count, since there is an odd number in there?



That's what I said. 2,3,1,6 would not count.

Quote:

Could this be reworded to say that if an odd number precedes the six, then that six doesn't end rolling, but resets the roll history?

For example 2-3-1-6-1-2-3-4-5-6-2-4-6 would be an example of a concluded trial. Note how there are all even number between the last and second to last 6.



But that would be a very long sequence, and we are concerned with the average length of the sequences.

What we do is roll until we get a 6, and record that as a sequence. We can do this repeatedly. In this case the expected length of a sequence is 6. (This is well known I think)

But now we throw out all the sequences that contain at least one odd number. What's the expected length of a sequence that remains?
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 13th, 2025 at 11:10:39 AM permalink
Quote: SkinnyTony


That's what I said. 2,3,1,6 would not count.

Quote:

Could this be reworded to say that if an odd number precedes the six, then that six doesn't end rolling, but resets the roll history?

For example 2-3-1-6-1-2-3-4-5-6-2-4-6 would be an example of a concluded trial. Note how there are all even number between the last and second to last 6.



But that would be a very long sequence, and we are concerned with the average length of the sequences.

What we do is roll until we get a 6, and record that as a sequence. We can do this repeatedly. In this case the expected length of a sequence is 6. (This is well known I think)

But now we throw out all the sequences that contain at least one odd number. What's the expected length of a sequence that remains?
link to original post



Thanks, but I still don't understand what's being asked. Maybe somebody smarter than me can rephrase it.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
camapl
camapl
  • Threads: 8
  • Posts: 679
Joined: Jun 22, 2010
August 13th, 2025 at 11:59:17 AM permalink
Quote: Wizard

Quote: SkinnyTony


That's what I said. 2,3,1,6 would not count.

Quote:

Could this be reworded to say that if an odd number precedes the six, then that six doesn't end rolling, but resets the roll history?

For example 2-3-1-6-1-2-3-4-5-6-2-4-6 would be an example of a concluded trial. Note how there are all even number between the last and second to last 6.



But that would be a very long sequence, and we are concerned with the average length of the sequences.

What we do is roll until we get a 6, and record that as a sequence. We can do this repeatedly. In this case the expected length of a sequence is 6. (This is well known I think)

But now we throw out all the sequences that contain at least one odd number. What's the expected length of a sequence that remains?
link to original post



Thanks, but I still don't understand what's being asked. Maybe somebody smarter than me can rephrase it.
link to original post



I’ll give it a try by starting the list…
SkinnyTony, let us know whether I’ve got the gist…
(rhyme unintentional!)

(partial) List of POSSIBLE sequences:
6
1,6
2,6
3,6
4,6
5,6
1,1,6
1,2,6
1,3,6
1,4,6
1,5,6
2,1,6
2,2,6
2,3,6
2,4,6
2,5,6
3,1,6
3,2,6
3,3,6
3,4,6
3,5,6
4,1,6
4,2,6
4,3,6
4,4,6
4,5,6
5,1,6
5,2,6
5,3,6
5,4,6
5,5,6


Now, we eliminate those with an odd number to make the…

(partial) List of INCLUDED sequences:
6
2,6
4,6
2,2,6
2,4,6
4,2,6
4,4,6


Now, I’ll attempt a new definition. Allowed sequences will include any number of 2’s and 4’s (including 0 of either or both) and exactly one 6, which is strictly at the end of the sequence.
I want to start wearing a T-shirt that reads, “Don’t feel sorry for me. I’m an AP!”
charliepatrick
charliepatrick
  • Threads: 39
  • Posts: 3099
Joined: Jun 17, 2011
Thanked by
camapl
August 13th, 2025 at 12:17:42 PM permalink
My feeling is the puzzle, where you ignore any sequence which has an odd numbered roll before the ultimate "6", is mathematically the same as rolling a die which only has 2, 4 and 6. In this case the chances of a "6" is 1/3, so the average length of the roll is 3.
ThatDonGuy
ThatDonGuy 
  • Threads: 128
  • Posts: 7270
Joined: Jun 22, 2011
Thanked by
charliepatrickMichaelBluejay
August 13th, 2025 at 12:40:13 PM permalink
Quote: SkinnyTony

Here's one that I got from a YouTube video.

We are going to roll a single 6-sided die repeatedly until a 6 is rolled, and keep track of how many rolls it takes.

So for example, if you rolled a 2, 3, 1, 6, that would be 4 rolls. If you rolled a 6 on the first roll, that would be 1 roll.

What is the average number of rolls needed, conditioned on the fact that all rolls are an even number? In other words, if we throw out all sequences that contain an odd number, what's the average length of the remaining sequences?

So for example, 2, 2, 6 would be 3 rolls. 2, 2, 3, 6 would not count since it includes an odd number (3). Similarly 2, 1, 3, 6 would not count.
link to original post


Hasn't this one been asked already? I remember providing an incorrect solution, but I can't remember if it was here or somewhere else.
My original answer was, it should be the same as if each die had only 3 sides, with numbers 2, 4, and 6, so the expected number of rolls would be 3. Turns out that this is wrong.


The probability for each number of rolls needed:
1 roll: 1/6 x 1
2 rolls: 1/3 x 1/6 x 2
3 rolls: (1/3)^2 x 1/6 x 3
...

The sum is 1/6 x (1 + 1/3 x 2 + (1/3)^2 x 3 + ...)
= 1/6 (1 + 1/3 + (1/3)^2 + ...)^2
= 1/6 x (3/2)^2
= 1/6 x 9/4
= 3/8

The probability of a valid sequences is 1/6 + 1/3 x 1/6 + (1/3)^2 x 1/6 + ...
= 1/6 (1 + 1/3 + (1/3)^2 + ...)
= 1/6 / (2/3)
= 1/4

The expected number of rolls over all valid sequences = 3/8 / 1/4 = 3/2.
This has been confirmed by simulation.



Update: this has been asked here before, in February - by me, in fact
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
Thanked by
MichaelBluejay
August 13th, 2025 at 1:50:48 PM permalink
I know this puzzle is getting stale, but I was asked for a more efficient way to get the answer to the circular tunnel puzzle.



You are trapped in a circular tunnel by an evil wizard. Inside the circular tunnel are evenly-spaced switches. You're told there are at least two switches. Each switch can be in either the On or Off position. You can at most one switch at a time only. You're allowed to change the position of the switches as much as you like, but may not leave any other markings in the tunnel.

The wizard will set you free if you correctly tell him the number of switches. However, if you're wrong, you'll be trapped forever in the tunnel. What do you do?

Simple Answer

Here is the simplest answer to explain, but not the most efficient.

1. Go to the nearest switch and turn it on.
2. Go clockwise to the nearest on switch and turn if off. Keep a count of the number of switches you travel to (not counting the starting switch).
3. Turn around and go counterclockwise the number of switches from step 2.
4. If the switch you come to is off, then the number of switches from step 2 is the answer.
5. Otherwise, if the switch you return to in step 4 is on, go back to step 2.

Here is a more efficient answer. Note the first four steps are the same as above.

1. Go to the nearest switch and turn it on.
2. Go clockwise to the nearest on switch and turn if off. Keep a count of the number of switches you travel to (not counting the starting switch).
3. Turn around and go counterclockwise the number of switches from step 2.
4. If the switch you come to is off, then the number of switches from step 2 is the answer.
5. Otherwise, if the switch you return to in step 4 is on, then decide on a multiplier factor greater than 1. For example, 5.
6. Multiply the number of switches you traveled in one direction the last time by the multiplier in step 5.
7. Travel clockwise the number of switches calculated in step 6.
8. Turn off all up switches as you go.
9. As you go, remember the number of switches traveled between the starting switch and the last up switch seen.
10. When you get to the number of switches from step 6, turn around.
11. Go back the same number of switches from step 6.
12. If the switch you come to is off, the number of switches is the total from step 9.
13. If the switch you come to is on, then go back to step 6.


I think it makes for good discussion for the best multiplier in step 5. I think we need a mean distance to properly answer the question. Without that, but assuming the tunnel could be really long, like a google number of switches, the factor should be about 45.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
charliepatrick
charliepatrick
  • Threads: 39
  • Posts: 3099
Joined: Jun 17, 2011
August 13th, 2025 at 2:18:33 PM permalink
Yes it looks as if I fell into the trap and can sense a goat and car somewhere in the background. I've also managed to run simulations that show the average length of winners, and coincidentally losers, are 1.5.
As it's past wine o'clock here I haven't seen an obvious logical reason!
Overall Result: W: 3748231 2497716 Avg: 1.500663406087802 L: 11256251 7502284 Avg: 1.5003765519940329
Parms: sh:10000000 Time:21:39:42:422
SkinnyTony
SkinnyTony
  • Threads: 2
  • Posts: 60
Joined: Jul 22, 2025
August 13th, 2025 at 2:35:24 PM permalink
Quote: camapl

Quote: Wizard

Quote: SkinnyTony


That's what I said. 2,3,1,6 would not count.

Quote:

Could this be reworded to say that if an odd number precedes the six, then that six doesn't end rolling, but resets the roll history?

For example 2-3-1-6-1-2-3-4-5-6-2-4-6 would be an example of a concluded trial. Note how there are all even number between the last and second to last 6.



But that would be a very long sequence, and we are concerned with the average length of the sequences.

What we do is roll until we get a 6, and record that as a sequence. We can do this repeatedly. In this case the expected length of a sequence is 6. (This is well known I think)

But now we throw out all the sequences that contain at least one odd number. What's the expected length of a sequence that remains?
link to original post



Thanks, but I still don't understand what's being asked. Maybe somebody smarter than me can rephrase it.
link to original post



I’ll give it a try by starting the list…
SkinnyTony, let us know whether I’ve got the gist…
(rhyme unintentional!)

(partial) List of POSSIBLE sequences:
6
1,6
2,6
3,6
4,6
5,6
1,1,6
1,2,6
1,3,6
1,4,6
1,5,6
2,1,6
2,2,6
2,3,6
2,4,6
2,5,6
3,1,6
3,2,6
3,3,6
3,4,6
3,5,6
4,1,6
4,2,6
4,3,6
4,4,6
4,5,6
5,1,6
5,2,6
5,3,6
5,4,6
5,5,6


Now, we eliminate those with an odd number to make the…

(partial) List of INCLUDED sequences:
6
2,6
4,6
2,2,6
2,4,6
4,2,6
4,4,6


Now, I’ll attempt a new definition. Allowed sequences will include any number of 2’s and 4’s (including 0 of either or both) and exactly one 6, which is strictly at the end of the sequence.
link to original post



That's not wrong, but the method to get the sequences is important. We are repeatedly rolling a fair die, stopping on a 6, and eliminating those sequences that contain an odd number. There are other ways to obtain the same sequences that might change the answer.
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 13th, 2025 at 2:54:13 PM permalink
Here are two more from the Japanese puzzles book. These are much easier.

(1) It takes 4 workers 4 days to produce 4 units. How many workers are required to produce 100 units in 100 days?

(2) One person enters an empty concert venue 1 minute after the doors open, and every successive minute, the number of people in the venue doubles. After 12 minutes, the venue is full. After how many minutes is the venue half-full?

Thankfully, I got both of those right, so I kind of redeemed myself. Like a coupon.
I run Easy Vegas ( https://easy.vegas )
DogHand
DogHand
  • Threads: 3
  • Posts: 2209
Joined: Sep 24, 2011
August 13th, 2025 at 6:08:12 PM permalink
Answer to #2:

11 minutes

Dog Hand
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 13th, 2025 at 6:46:02 PM permalink
Yes, DogHand.

It's basically a concert Martingale.
I run Easy Vegas ( https://easy.vegas )
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 13th, 2025 at 9:02:26 PM permalink
Quote: MichaelBluejay

Here are two more from the Japanese puzzles book. These are much easier.

(1) It takes 4 workers 4 days to produce 4 units. How many workers are required to produce 100 units in 100 days?

(2) One person enters an empty concert venue 1 minute after the doors open, and every successive minute, the number of people in the venue doubles. After 12 minutes, the venue is full. After how many minutes is the venue half-full?

Thankfully, I got both of those right, so I kind of redeemed myself. Like a coupon.
link to original post




1. 4
2. 11
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
Ace2
Ace2
  • Threads: 32
  • Posts: 2792
Joined: Oct 2, 2017
August 13th, 2025 at 9:12:58 PM permalink
Quote: SkinnyTony

Here's one that I got from a YouTube video.

We are going to roll a single 6-sided die repeatedly until a 6 is rolled, and keep track of how many rolls it takes.

So for example, if you rolled a 2, 3, 1, 6, that would be 4 rolls. If you rolled a 6 on the first roll, that would be 1 roll.

What is the average number of rolls needed, conditioned on the fact that all rolls are an even number? In other words, if we throw out all sequences that contain an odd number, what's the average length of the remaining sequences?

So for example, 2, 2, 6 would be 3 rolls. 2, 2, 3, 6 would not count since it includes an odd number (3). Similarly 2, 1, 3, 6 would not count.
link to original post

Answer: The average rolls to get a six or odd is 6/(1+3) = 1.5. Call this a sequence

1/(1+3) chance the six came first and the string ends there. If odd came first (3/4 chance) then an average of six more rolls will be needed to get a six. So 1.5 average rolls to get a six when the six comes first and 7.5 average rolls to get a six when an odd comes first

Check: 1/4 * 1.5 + 3/4 * 7.5 = 6 average rolls to get a six irrespective of even/odd

One would initially assume that you can just consider 2,4,6 which would imply an answer of 3 average rolls to get a six. Assuming this, there is a 1/3 chance of ending a sequence with each roll. In actuality, there is a 4/6 chance of ending a sequence with each roll. The reciprocal (6/4) being the answer
Last edited by: Ace2 on Aug 13, 2025
It’s all about making that GTA
MichaelBluejay
MichaelBluejay
  • Threads: 93
  • Posts: 1820
Joined: Sep 17, 2010
August 13th, 2025 at 9:51:39 PM permalink
Quote: Wizard

Quote: MichaelBluejay

Here are two more from the Japanese puzzles book. These are much easier.

(1) It takes 4 workers 4 days to produce 4 units. How many workers are required to produce 100 units in 100 days?

(2) One person enters an empty concert venue 1 minute after the doors open, and every successive minute, the number of people in the venue doubles. After 12 minutes, the venue is full. After how many minutes is the venue half-full?

Thankfully, I got both of those right, so I kind of redeemed myself. Like a coupon.
link to original post




1. 4
2. 11

link to original post

Well, if you can solve the package/key problem, then obviously these two were trivial for you. I mean, if *I* can get them right...

(1) Since the Wiz didn't show his work, and non-math people tend to incorrectly answer "100", here's how it works:

Four workers working four days is (4x4=) 16 person-days. They produce 4 units, so it takes 16 person-days ÷ 4 units = 4 person days to make 1 unit.

So 100 units should take 400 person-days.
We know we have 100 days, so 400 person-days ÷ 100 days = 4 persons.

To me this is a units-of-measurement problem. "Man-hours" is a related unit, how many hours are worked by all workers in a given situation. Here I just changed the work time from "-hours" to "-days", because that's what the problem called for.

The most useful thing I learned in college was in freshman Chemistry, where we covered unit conversion. I like units of measurement so much I made one.

(2) The venue question is basically a concert Martingale.
I run Easy Vegas ( https://easy.vegas )
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 14th, 2025 at 3:09:12 PM permalink
Fair warning this one is a bit tough.

An evil wizard puts you in the middle of a desert. He informs you he hid a treasure directly north of you. Yes, you have a compass. To escape, you must find the treasure and bring it back to the starting point. You have absolutely no idea how far north the treasure is. There is a condition that you must declare how far north you will go each trip before leaving. If you don't find the treasure, you must turn back at that point. If you do find it, you must still travel that distance before returning. You may make as many trips as you wish.

What is the most efficient strategy to find the treasure to find it with the least total distance traveled?

Let me know if you have any questions or find the rules unclear.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
ThatDonGuy
ThatDonGuy 
  • Threads: 128
  • Posts: 7270
Joined: Jun 22, 2011
August 14th, 2025 at 3:53:15 PM permalink
Let me see if I understand the problem correctly.
I say, "I will walk north distance X, and then return," then I walk north distance X and return.
The treasure is hidden in such a way that I am guaranteed to find it if I walk to, or past it. For example, if the treasure is 16.8 km north, and I declare that I am walking any distance of 16.8 km or longer, then I will find it.
Note that if I do find the treasure before reaching distance X, then I must continue north until I am distance X from my starting point, then return.
If I do not find the treasure (because it is farther away than X), then I choose a new X and repeat the process.
Last edited by: ThatDonGuy on Aug 14, 2025
Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 14th, 2025 at 4:13:11 PM permalink
Quote: ThatDonGuy

Let me see if I understand the problem correctly.
I say, "I will walk north distance X, and then return," then I walk north distance X and return.



Yes!

Quote:

The treasure is hidden in such a way that I am guaranteed to find it if I walk to, or past it. For example, if the treasure is 16.8 km north, and I declare that I am walking any distance of 16.8 km or longer, then I will find it.



Yes!

Quote:


Note that if I do find the treasure before reaching distance X, then I must continue north until I am distance X from my starting point, then return.
If I do not find the treasure (because it is farther away than X), then I choose a new X and repeat the process.
link to original post



Yes, exactly. Every time you return you may set a new distance to anything you wish.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
ThatDonGuy
ThatDonGuy 
  • Threads: 128
  • Posts: 7270
Joined: Jun 22, 2011
August 14th, 2025 at 4:15:02 PM permalink

Since the treasure is "due north", it is between the starting point and the North Pole.
Let D be the distance to the North Pole.
Choose a positive integer N, and declare distances of D/N, 2D/N, 3D/N, and so on.
Assuming the treasure's location is uniformly possible anywhere between the starting point and the north pole, the expected distance traveled is:
1/N x 2 (D / N) + 1/N x 2 (2D / N) + 1/N x 2 (3D/N) + ... + 1/N x 2(ND/N)
= 2D / N^2 (1 + 2 + 3 + ... + N)
= 2D N (N+1) / (2 N^2)
= D (N+1) / N
This decreases as N approaches positive infinity, so choose as large an N as you feel you can handle.
In other words, choose as small of a distance as you can as your first distance, then after each unsuccessful trip, increase the distance by the initial distance.

Wizard
Administrator
Wizard
  • Threads: 1533
  • Posts: 27683
Joined: Oct 14, 2009
August 14th, 2025 at 4:41:31 PM permalink
Per Don's answer, assume that the Wizard transports you to another planet of unknown size. You may assume he chooses a location randomly between you and the north pole of this planet, with every spot equally likely.

Fair warning that as I double check my work, I find there are an infinite number of solutions that are all equally good, which I find surprising.

Never mind crossed-out statement above. I was wrong, I think.
Last edited by: Wizard on Aug 14, 2025
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
  • Jump to: