Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

8^8 – Find another you

8x8x8x8x8x8x8x8.com

Post reply on HN
Print view View original

71–80 of 166 posts

Re: 8^8 – Find another you

#71

Post by tajen »

DanBC wrote: post #57
tajen wrote: post #51

Earlier quoted context omitted.

> There seems to be an awful lot of gratuitous negativity going on here. There isn't much negativity above your comment, no-one above you has criticized the speed or the CAPTCHA. Were you referring to a specific comment without using the reply-to feature? Are you saying the HN community is "mean" about it? Most people on HN are very positive even if the project is slow, because everyone knows what it means to be on t…

Did you see this post and thread, asking HN to avoid gratuitous negativity? https://news.ycombinator.com/item?id=9317916

I'm not sure why my comment seems negative to you, while the comment above doesn't. The person above was generalizing negatively about "the people here", and it seemed gratuitous because no-one had criticized the speed or the CAPTCHA before.

In other words, please hellban me now, because I didn't get what was negative in my first comment.

Re: 8^8 – Find another you

#72

Post by gattis »

Nadya wrote: post #56
gattis wrote: post #49

Earlier quoted context omitted.

Even distribution is just important for doing it in as few questions as possible. If you want to do it in 8, and cover 16M distinct personalities, you're gonna need them to be very evenly distributed. Check out http://en.wikipedia.org/wiki/Entropy_%28information_theory%2... Information gain is maximized when the distribution is even. Imagine a very extreme biased form of this where everyone answers the same way to a…

There have been 0 matches with over 12,000 tests. Assuming their claim of 1/1,000~ is anywhere near remotely true - then you may have increased your chances of finding a 'good match' tenfold already with the assumption that matches will be good (there is of course the chance they won't be) The Myers-Briggs places weight on questions to give one of the 16 personality types made of 8 traits. For example, one answer mig…

the MBTI example and use of weighting only makes it more clear that this "case study" is dead from the start, because it shows it takes many questions just to ascertain someone's position on each eigenvector. For example, the most important explanatory personality trait is masculinity vs femininity. How do you ascertain that with one question and 8 answers? For example a question asking if you are male or female only tells you so much and not the whole story. none of the real underlying important explanatory variables can be directly measured with a question, or 8 questions.

And I highly doubt they have no matches in 10k samples, especially given the homogeneity of their audience. And if so then they are just measuring traits of personality that, while diversified, nobody cares about. Check out the birthday paradox. Even if there are 16M possibilities, the chance that two random people in 10,000 matched is much higher than you think.

Re: 8^8 – Find another you

#73

Post by yellowapple »

ordinary wrote: post #32
braaap wrote: post #10

Ummm, I like other people precisely because they aren't me. My soulmate is not a duplicate of me. They are a compliment. At best, 8^8 might expose someone who ends up making a lot of the same decisions as me. Boring!

Nitpick: you mean complement. I'm going to go ahead and assume your soulmate isn't "your hair looks nice today!" :)

What if the parent commenter is narcissistic?

Re: 8^8 – Find another you

#74

Post by yellowapple »

insickness wrote: post #47

This whole thing has been done before and in a way that allowed the questions to be crowdsourced, not just one person coming up with what he/she thinks are meaningful. On OKCupid, There are thousands of questions, many of them penned by users. You choose which questions to answer, then mark how important this question is to you. You also choose the answer you want your partner to have. For example: Are you looking fo…

IIRC, OkCupid also has a "not at all important" option; that way, you could answer "yes" to the hypothetical question without caring about someone else's answer (while at the same time providing an answer for somebody else who does care about the answer).

Re: 8^8 – Find another you

#75

Post by Someone »

matheist wrote: post #6
ColinWright wrote: post #2

To prevent abuse, we need to ensure that everyone taking the 8^8 test is a human and not a bot. Click on all of the icons below that represent animals. And then the site is so slow it only shows me 5 of 7 images, the other two failing to load. Good one. Don't get me wrong, it's a cute idea, although I have to say that I'm not sure I'd get along with someone who thinks exactly as I do. Edit: I've passed the spam test…

And if there are always 3 animals (like on mine), then one out of every 35 (= 7 choose 3) bots will get through it anyway. Even if each picture is uniformly and independently an animal or non-animal, one out of every 128 bots will get through. If bots are a problem, maybe use a stronger captcha. It sounds like a neat idea but it's taking too long to load for me to check it out. I'll try again later.

That either assumes a spammer makes the effort to program his bot to handle this problem or a bot so intelligent that it can interpret what this "proof you're human" task entails.

The second, I think, can be ignored for now. The first is fairly unlikely, given the current popularity of the site and what a spammer stands to gain (hm, maybe there is quite somewhat to gain. One could register a few million different fake profiles, and use them to find gullible clients for a "I would like to come and visit you, but need money for the trip" scam. Edit: maybe, this whole site is part of such a scam)

Also, if I were to make my bot aware of this "I'm human" test, I also would program in knowledge of the URLs or the pixels of the pictures used, and get a 100% success rate.

Re: 8^8 – Find another you

#76

Post by dsyko »

david-given wrote: post #55
matheist wrote: post #6

Earlier quoted context omitted.

And if there are always 3 animals (like on mine), then one out of every 35 (= 7 choose 3) bots will get through it anyway. Even if each picture is uniformly and independently an animal or non-animal, one out of every 128 bots will get through. If bots are a problem, maybe use a stronger captcha. It sounds like a neat idea but it's taking too long to load for me to check it out. I'll try again later.

I got two animals and a bird. Whether I click on just the animals or all three of them is part of the test, right?

I think you may have mis-read 'animal' as 'mamaml' A bird is an animal.

Re: 8^8 – Find another you

#77

Post by Nadya »

gattis wrote: post #72
Nadya wrote: post #56

Earlier quoted context omitted.

There have been 0 matches with over 12,000 tests. Assuming their claim of 1/1,000~ is anywhere near remotely true - then you may have increased your chances of finding a 'good match' tenfold already with the assumption that matches will be good (there is of course the chance they won't be) The Myers-Briggs places weight on questions to give one of the 16 personality types made of 8 traits. For example, one answer mig…

the MBTI example and use of weighting only makes it more clear that this "case study" is dead from the start, because it shows it takes many questions just to ascertain someone's position on each eigenvector. For example, the most important explanatory personality trait is masculinity vs femininity. How do you ascertain that with one question and 8 answers? For example a question asking if you are male or female only…

If it were trying to match people based on their personalities and not their direct line of thinking - you would be correct.

Until the results speak for themselves (ie. accuracy of matches) you have no way of knowing if this is a complete failure or successful venture. Although you are free to speculate that it will be meaningless due to the methodology, you have no evidence to back that up yet.

Many dating sites work on a similar principle, although they all use weighted questions. Weighted questions for personality increase possible matches by narrowing down the preciseness necessary. The goal here is to decrease possible matches by requiring exact matches.

Dating sites want to match you with someone quickly who is "close enough" to be compatible based on a personality test and their user data of what answers pair well with other peoples' answers.

8x8 doesn't care how long it takes to find a match, they want an "exact match" to be compatible based on how you answer arbitrary questions.

8 questions was chosen because to have answers be "100% exact" are already rare. If they chose 50 questions - the chances of an exact match become increasingly rare as to possibly never exist in a persons' life time.

E:

The birthday paradox is 1/365, and I'm well aware of the statistics of someone out of a group of 20 random people sharing a birthday with another person of that same group.

Re: 8^8 – Find another you

#78

Post by LordHumungous »

It's cool, but it kind of reminds me of Ok Cupid back in the day when they would ask you 200 questions about your personality, and then it would match you with people who gave similar answers. I went on dates with a couple people who were 99% matches, expecting love at first sight, only to find that in real life we didn't click at all. It seems that there are many dimensions to human interaction that are not easily captured by these kinds of quizzes.

Re: 8^8 – Find another you

#79

Post by robocat »

temuze wrote: post #33
aidenn0 wrote: post #27

Earlier quoted context omitted.

Simple stupid example. A lot more than 1/5 of the people would choose A for the following question: How many people have you killed? a) 0 b) 1-9 c) 10-99 d) 100-999 e) 1000 or more

Or, for the second problem or correlations: Do you have a Y chromosome? a) Yes b) No Do you have two X chromosomes: a) Yes b) No Each are 50-50 questions, but the answer to A is correlated to the answer to B. It's hard to find questions whose answers are independent variables.

Perhaps not correlated in the way you are think, due to confounding factors:

A) how many people don't understand the questions

B) how many people understand the questions but don't choose to answer "correctly" e.g. transgender. E.g. smart arse Yes, Yes answers.

C) correlation with a third variable e.g. sampling bias (HN readers that answer internet surveys).

Re: 8^8 – Find another you

#80

Post by Arnor »

TazeTSchnitzel wrote: post #65
cbaker wrote: post #61

Why would you want to meet someone exactly like yourself? You already know you. I'd rather meet people different from me.

Truth be told, opposites don't usually attract. People prefer others who are similar to themselves.

There's a wide spectrum between "exactly like you" and "opposite"
Post reply on HN