How does this work?

[SezMe]

Try this (http://www.smalltime.com/dictator.html) questionaire.

I have tried it four times with both types of people and it has been 100% correct. Now, I did not count the number of questions it asks (about a dozen+) but is it reasonable to deduce that a people data base could discriminate between tens of thousands of people based on binary questions? I know, theoretically that 2^12 is a huge number, but I don't think it is fully relevant here.

Can you beat the genie?

[thaiboxerken]

Wow, I did Drew Carey, and it worked!

[canadarocks]

I beat the Genie with no problem. I used a Sit-com character from one of my favorite BBC shows. If you beat the "Genie", the program asks you to write a question that would identify the character. I had to go through a number (>25) of questions that I answered "No" to. I think that the program has a large database of Y/N questions and it gets added to (improved) everytime you beat the "Genie".

It's fun though!!

Cheers

[Dr Adequate]

Originally posted by RSSchlueter
Try this (http://www.smalltime.com/dictator.html) questionaire.

I have tried it four times with both types of people and it has been 100% correct. Now, I did not count the number of questions it asks (about a dozen+) but is it reasonable to deduce that a people data base could discriminate between tens of thousands of people based on binary questions? I know, theoretically that 2^12 is a huge number, but I don't think it is fully relevant here.

Can you beat the genie?
How does it work? This is the Pangolins algorithm. When it offers a guess as to who you are, if it gets it wrong, it asks you a question of the form:

"Tell me a good general question that guesses you with a 'yes' answer and (the person it guessed you were) with a 'no' answer."

So the users expand the database.

[geni]

it didn't get Ayatollah Ali Khamenei or Nong Duc Manh

[Dr Adequate]

Originally posted by geni
it didn't get Ayatollah Ali Khamenei
That's because he's not a dictator as such, but simply the top Islamic jurisprudent in Iran. Iran is a democracy, but with the Qur'aan as the constitution and Ayatollah Ali Khamenei as supreme court. Of course, if you answered the last question, the program now thinks that he is a dictator, and there's no way to get your data out of its files. Ooops.

[Gastric ReFlux]

Originally posted by Dr Adequate
That's because he's not a dictator as such, but simply the top Islamic jurisprudent in Iran. Iran is a democracy, but with the Qur'aan as the constitution and Ayatollah Ali Khamenei as supreme court. Of course, if you answered the last question, the program now thinks that he is a dictator, and there's no way to get your data out of its files. Ooops.

We could sabotage it.

We could teach it that Edith Bunker was a dictator, and that Gilligan was a transvestite.

[geni]

Originally posted by Dr Adequate
That's because he's not a dictator as such, but simply the top Islamic jurisprudent in Iran. Iran is a democracy, but with the Qur'aan as the constitution and Ayatollah Ali Khamenei as supreme court. Of course, if you answered the last question, the program now thinks that he is a dictator, and there's no way to get your data out of its files. Ooops.

He has pretty much total power. He may choose to share some of it but that doesn't mean much.

[thaiboxerken]

It got GW Bush correct.

[Dr Adequate]

Originally posted by thaiboxerken
It got GW Bush correct.
As a dictator or a comic character? There's some ambiguity there... HE'S BOTH. Arrrgh!

[phildonnia]

Stumped it with His Excellency Saparmurat Niyazov Turkmenbashi.

[SezMe]

Originally posted by Dr Adequate
How does it work? This is the Pangolins algorithm. When it offers a guess as to who you are, if it gets it wrong, it asks you a question of the form:

"Tell me a good general question that guesses you with a 'yes' answer and (the person it guessed you were) with a 'no' answer."

So the users expand the database.
That is certainly a reasonable way to expand the database. I wonder how many people are in it and how many it takes to get it to the point where is sufficiently robust to interest users.

canadarocks had a good approach - a non-US TV character. I'll bet the Aussies can fool it too.

Edited to add: Googling does not yield anything useful about the Pangolin algorithm. Can you provide a link?

[SezMe]

Originally posted by phildonnia
Stumped it with His Excellency Saparmurat Niyazov Turkmenbashi.

Not sure this counts. This (http://www.turkmenistanembassy.org/turkmen/gov/presbio.html) site describes him as "His Excellency Saparmurat Niyazov "Turkmenbashi" President of Turkmenistan and Chairman of the Cabinet of Ministers. Since it got Dubya right, I think the title of the quiz is no longer correct. It is more than just sitcom characters and dictators. Thanks, in part, to us it is filling up with other shady characters.

[geni]

Originally posted by RSSchlueter
Not sure this counts. This (http://www.turkmenistanembassy.org/turkmen/gov/presbio.html) site describes him as "His Excellency Saparmurat Niyazov "Turkmenbashi" President of Turkmenistan and Chairman of the Cabinet of Ministers. Since it got Dubya right, I think the title of the quiz is no longer correct. It is more than just sitcom characters and dictators. Thanks, in part, to us it is filling up with other shady characters.

he had himself declared "President for Life". I don't think there can be any real question over his credentials

[apoger]

Stumped it with The Chief from Get Smart.

On the other hand it got a bunch of others.

I am incredibly amused. :)

[phildonnia]

On the other hand, it had no problem with "Chrissy Snow" or "Rachel Green".

[Zep]

Originally posted by RSSchlueter
canadarocks had a good approach - a non-US TV character. I'll bet the Aussies can fool it too. Yep, we can. I used Norman Gunston, the "Little Aussie Bleeder", and stumped it first time. Who's Norman Gunston, you ask? Heh heh! Have a look at AUP's avatar...

[EdipisReks]

this is kinda cheap. i chose Niles Crane, and the questions "is your brother a talk show host" and "do you live in seattle" obviously clinched it.

[Batman Jr.]

I tried Simon Pegg's character from "Spaced" and it got it right!

[RebeccaBradley]

That was the most fun I've had all week. I think the database is catching up on the BBC shows, though - it got me on characters from The Office, Royle Family, Black Books, Blackadder, Some Mothers Do 'Ave 'Em, AbFab, Red Dwarf, Jeeves and Wooster... the only time I won was with Honoria Glossop from Jeeves and Wooster, and maybe she's not considered a major character. Okay, so now I'll go back to having a life.

[canadarocks]

They are catching up on the BBC shows. It got "Susan" from Coupling, but not Steve.

[Dr Adequate]

Originally posted by RSSchlueter
Edited to add: Googling does not yield anything useful about the Pangolin algorithm. Can you provide a link?
Er, not particularly. But I've told you how it works. "Pangolin" --- it's a sort of South American anteater. The canonical example involves guessing what animal you're thinking of...

[BPSCG]

It guessed - wrongly - Fox Mulder from The X Files for me.

Didn't know that was a sitcom. Oh well.

Anyway, I can see how it got there after I answered "yes" to "Are you in law enforcement?" (or police - I forget which) and then followed up with "Do you believe in aliens?".

Problem is, those answers apply equally to Arthur Dietrich, from Barney Miller. There's actually some evidence in the show's scripts to suggest that Dietrich was himself an alien...

[rppa]

Originally posted by RSSchlueter
Try this (http://www.smalltime.com/dictator.html) questionaire.

I have tried it four times with both types of people and it has been 100% correct. Now, I did not count the number of questions it asks (about a dozen+) but is it reasonable to deduce that a people data base could discriminate between tens of thousands of people based on binary questions? I know, theoretically that 2^12 is a huge number, but I don't think it is fully relevant here.

Can you beat the genie?

Beat it first time out with "Herman Munster" from the Addams Family ripoff "The Munsters". It thought I was "Maxwell Smart".

Did you say 2^12 because you got asked 12 questions? I was asked far more than that, most of which were "no". I think it has a bunch of questions ("Do you live on an island?") which are meant to be unique discriminators.

But say it's 12. That is indeed relevant. It provides a binary number which is unique to each person in the database. And every time it gets stumped, it asks you for one more question which will let it now discriminate the person you just added. So every time you add a new person to the database, you are asked to create a unique signature for that person.

It's probably also counting on characters uppermost in people's minds, either recent, headline makers, or popular shows. Think I'll try again with "Idi Amin" and "Mr. Ed".

I'm not going to update it with a Munster question because I didn't know the show well enough. I might have answered a question wrong.

[Ashles]

It got Lord Melchett from Blackadder 2. Wow.

Although basically everyone is help writing it. If you catch it out you ensure it won't be cought out again.

[Ashles]

Damn, now it got Eddie the Shipboard computer from The Hitchhiker's Guide to the Galaxy!

This thing is Gooood.

[rppa]

Think I'll try again with "Idi Amin" and "Mr. Ed".

Mr. Ed:
Once it got past "Are you human? No" the questions got pretty obvious. Are you a dog? Are you a beaver? Do you have nine lives? About six or seven of those (possibly even "Are you a horse?") before it got to "Is your friend's name Wilbur?"

Ever play the parlor game Twenty Questions? It's amazing how far you can get with just 20 questions, and you don't even have to limit yourself to TV characters or dictators.

[TragicMonkey]

I was the 136th person to try Julius Caesar. It took a lot more questions to light on him than it did for Herman Munster, Murray Slaughter, and Mrs Slocombe. I did stump it with Jerri Blank, but it's debatable whether Strangers With Candy meets the technical definition of a sitcom.

[TragicMonkey]

I am impressed. It got me with the Empress Tzu Hsi of China!

[McFunley]

I rocked it with Julius Caesar. I don't think it's chance that the database has a lot more sitcom characters than important historical figures.

[rppa]

[B]Think I'll try again with "Idi Amin" and "Mr. Ed".[b]

Idi Amin:
Got through at least 20 "no" answers before it asked me flat out "Are you a dictator?" Seems like cheating to me. Went quickly after that:

Are you from South America? No.
Are you a cannibal? Yes.

I guess you are Idi Amin.

[TragicMonkey]

Hmm. It got Jessica Tate, but not her daughter Corinne. This is challenging, and by playing we're just making it harder!

[Dr Adequate]

I decided to see what would happen if I kept saying "No" all the time. The results are hilarious, and moreover, after about 100 times it came up with:

"Are you a Jack in the Box employee who is really lame and keeps saying no?"

Spookily accurate. Now I know what you're thinking --- comic character or dictator?

World Leader Pretend.

[TragicMonkey]

Marilyn from the Munsters gave it fits, but it got it eventually.

[Dr Adequate]

Originally posted by Ashles
It got Lord Melchett from Blackadder 2. Wow.
"Lord Melchett, Lord Melchett, intelligent and deep
Lord Melchett, Lord Melchett... a shame about the sheep".

Is there any reason besides Rowan Atkinson why you choose to live in the 80s? I mean stop me if I'm wrong, but ---

Good things about the 1980s:

(a) Blackadder
(b) The Smiths

Bad things about the 1980s:

(a) The 1980s

[Ashles]

Although the 80s is my home, even I feel you put (b) in the wrong list.

Unless you literally meant the Smiths i.e. their music and performance was excellent, but it was slightly ruined by a whiny annoying git singing all over it.

[Dr Adequate]

I feel that a fey Elvis impersonator flagellating himself with daffodils was was the world needed at that point. BTW, while you're there, could you shoot Mrs Thatcher. Don't worry about temporal paradox --- we'll cope.

[Ashles]

What are you talking about? Mrs Thatcher will be in power for years - I can't see anything or anyone that would end her reign of power.

Who would take her on? John Major? Don't make me laugh. Next you'll be telling me he's the type to have a sex scandal.

[Dr Adequate]

Originally posted by Ashles
What are you talking about? Mrs Thatcher will be in power for years
Not if you shoot her. Now, I realise it's a terrible thing to take a human life... but how do you feel about evil soulless killer robots?

[Ashles]

I'm too busy listening to A-ha and waiting for Michael Aspel to host his next telethon.

Life is good.

Although I would quite like Labour to get in in some ways - they would deal with all the fat-cat business types, they'd improve public transport, pour millions into the NHS and encourage non-violent solutions to international conflicts.

[SezMe]

Originally posted by rppa
...snip...

Did you say 2^12 because you got asked 12 questions? I was asked far more than that, most of which were "no". I think it has a bunch of questions ("Do you live on an island?") which are meant to be unique discriminators.

But say it's 12. That is indeed relevant. It provides a binary number which is unique to each person in the database. And every time it gets stumped, it asks you for one more question which will let it now discriminate the person you just added. So every time you add a new person to the database, you are asked to create a unique signature for that person.

Yes, I assumed 12 questions, just as a guess. I questioned the relevance because so many answers are potentially "no" when dealing with distant historical figures or obscure characters because that is how to answer if you don't know.

The fallacy of my reasoning is that the number of questions is fixed. Apparently it is not. And now that I know what happens when you do beat the genie, it makes even more sense.

[Ratman_tf]

I win again! You are player number 9 to have chosen Megatron from Transformers. I knew you were Megatron from Transformers from the start, but I strung you along for a while to make it seem more sporting. I hope that one day you will overcome the powerful sense of humiliation that you now feel. Until then, good luck.

****in... *Grumble grumble*

*Edit*

And stumped it with Baron Karza from The Micronauts! Ha!

[Blondin]

You won this time, puny mortal! And I fail again. I guess my parents were right, I'll always be a failure. I guess it's time to start looking for a new job.

Who'd a thunk it wouldn't know about Oscar Madison?









(The slob from The Odd Couple)

[RebeccaBradley]

All riiiight! The Canuck shows are in - somebody beat me to Richard and Veronica from "Made in Canada". (Great show.)

[epepke]

Originally posted by RSSchlueter
Edited to add: Googling does not yield anything useful about the Pangolin algorithm. Can you provide a link?

It's very simple. It's a binary decision tree. Each node is either a guess or a question. The algorithm walks the tree from the root based on the answers to the questions. If the guess is wrong, it replaces the guess node with a question node and a pair of guess nodes, one for the existing guess and one for your answer.

http://www.csc.liv.ac.uk/~frans/COMP101/AdditionalStuff/javaDecTree.html appears to give an implementation.