New Website Offers Provably Fair Solutions To Everyday Problems

An anonymous reader writes Carnegie Mellon researchers have just launched Spliddit, a website that offers methods for helping people split rent, divide goods, and share credit. The novelty is that these methods are all "provably fair": there are mathematical proofs showing that each algorithm on the site provides rigorous fairness guarantees. For example, the method for splitting rent is guaranteed to be envy free: the assignment of rooms and division of rent is such that a housemate would never want to swap places with another housemate. All it takes is a pair of siblings to prove that there's no such thing as "provably fair," non-mathematically.

Comments

It doesn't work

[wcrowe]

I keep telling my dogs that the way that Spliddit divvies up the expenses is PROVABLY FAIR, but all they do is look at me blankly, cocking their heads to one side. Damn dogs. I never see a dime out of either of them.

Re:It doesn't work

[OglinTatas]

In addition to dogs children won't get it. If you give two children 4 crayons each, who has more? The answer is always the other does.

Re:It doesn't work

[Quirkz]

Hell, if you give two children identical copies of the same color crayon, one will still want what the other has.

Re:It doesn't work

[chronoglass]

http://www.spliddit.org/apps/g...

I only got half of a cat!

Re:It doesn't work

[penguinoid]

It in fact doesn't work:

This property is guaranteed, assuming that each housemate wishes to maximize the difference between what she thinks her room is worth and how much rent she has to pay.

That's not the objective of most people -- they will value rooms differently depending on which other rooms they own.

And I suspect someone could gain an advantage by dishonestly reporting how much they value each room. Specifically, if you express a value for the rooms as almost the same as your roommate, but slightly lower for the rooms he wants the most and slightly higher for the rooms he likes the least, I suspect their mechanism for overdemanded and underdemanded rooms will give you the slightly inferior rooms at a greatly reduced price.

Re:It doesn't work

[ozmanjusri]

In addition to dogs children won't get it. If you give two children 4 crayons each, who has more? The answer is always the other does.

As usual, the bible has the answer.

First, "divide the living child in two" (1 Kings 3:25). This will give you four half-children, each with two crayons, and a blissfully quiet household.

Re:It doesn't work

[davester666]

paws or head?

Re:It doesn't work

[RockDoctor]

Isn't there an implication that the algorithm is provably fair to a rational being. The last time I interacted with them, human children did not appear (mostly) to be rational beings.

sibling fairness

[orgelspieler]

The best example of a fairness algorithm is an old one used with siblings. Tell one kid to divvy up the candy/cake/toys whatever, and let the other kid choose which half they want.

Greece used something similar when trying to get people to honestly report the value of their antiquities. If they listed it for a price the government thought was too low, the agency could buy it for that price. I wish local governments would do something like that with home values. If they want to tax me for a house worth more than what I can get for it on the open market, then I should have the right to sell it to them at that price.

Re:sibling fairness

[Ichijo]

If they want to tax me for a house worth more than what I can get for it on the open market, then I should have the right to sell it to them at that price.

Better yet, make property taxes reflect the property's burden on the government. What we have now is an unscrupulous diner's dilemma situation where people who maximize their street frontage in order to give themselves more places to park pay only a tiny fraction of the incremental cost (land, maintenance, lighting, emergency response, etc.), and everyone else does the same in return, driving up all our taxes. It's madness.

Re:sibling fairness

[Anonymous Coward]

And if you tell them it's worth $5 and you sell it for $500k, do they get to charge you with fraud for ripping off the other person by $499995

Not like that. If you claim the property is only worth $5, the state should have the option to purchase the property from you at your "asking" price.

Tit-for-tat.

If the state sets the 'appraisal' value, then the owner should be able to force the state to purchase at the state's appraised price.
If the owner sets the 'appraisal' value, then the state should be able to force the owner to sell at the owner's appraised price.

Re:sibling fairness

[stephanruby]

The best example of a fairness algorithm is an old one used with siblings. Tell one kid to divvy up the candy/cake/toys whatever, and let the other kid choose which half they want.

That's essentially how their web site works, except they're asking every roommate to divide up and assign a value to each space, before asking them which space they might prefer. In a way, I like their site better, because it seems to work more like a blind auction. By keeping the bids of other roommates initially secret, then they're essentially preventing them from changing their mind halfway before the process is finished. It makes the process much cleaner that way and much more haggle-free.

Re:sibling fairness

[Obfuscant]

What we have now is an unscrupulous diner's dilemma situation where people who maximize their street frontage in order to give themselves more places to park

It doesn't matter how much frontage you have, the street is still the same length and there are still the same number of parking spaces. In fact, the fewer the driveways, the MORE spaces to park there are. How much space you have to park depends on the area of your property, not the frontage. A dual driveway and a two car garage gives four spaces no matter how much frontage that property has.

You're confusing "public on-street parking" with "private parking". Those spaces on the street are intended for transients, not residential parking. That's why it is called "public on-street parking."

Re:sibling fairness

[disambiguated]

Hmm.... Maybe they should do this with intellectual property.

The owner of intellectual property would be taxed based on a value the owner specifies annually. The government would have the option to purchase it for that amount, perhaps with an added premium. If the property is sold, the purchaser is taxed initially based on what they paid. Licences & royalties would be limited based on the taxed value.

I was about to say maybe this would be too complicated to get right... then I remembered our tax code.

Re:sibling fairness

[Kvasio]

the algorith is old one, I remember it from Hugo Steinhaus's math book.
It works for any number of parties and goods.

Say we have 4 brothers who have to divide the heritage: home, car and bicycle

Step 1: each brother provides his valuation, e.g.
Adam home $200,000 car $10,000 bike $100 - total value $210,100 thus his "fair part" is $52,525
Brad home $150,000 car $3,000 bike $120 - total value $153,120 thus his "fair part" is $38,280
Caleb home $180,000 car $11,000 bike $80 - total value $191,080 thus his "fair part" is $47,770
Damon home $50,000 car $3,000 bike $60 - total value $53,060 thus his "fair part" is $13,265

Step 2
whoever "bid" the highest for given good, gets it, at his own valuation.
Adam gets home (valued by him $200,000), which is $147,475 more, than his "fair part", so has to pay $147,475 to the pool
Brad gets bike (valued by him at $120), which is $38,160 less, than he believes he should get
Caleb gets cat (valued by him at $11,000) which is $36,770 less than his definition of a fair part.
Damon gets no item, which is $13,265 lower, than he had hoped to get

Step 3
Adam should pay $147,475 to the pool.
Brad gets $38,160 from the pool
Caleb gets $36,770 from the pool
Damon gets $13,265 from the pool

Now every brother got exactly what he valued as a 1/4th of total items value.

And we've still got $59,280 in pool to share. Which may:
- be split equally - each brother gets $14,820 "bonus"
- be split proportionaly - each brother gets part of that $59k split by weights of their total valuation sum (in our example - each would bet 39.04% more than he expected)
- be stolen by the court/the man splitting goods ;-)

Hugo Steinhaus also mentioned that this procedure may be altered to minimise cash flows (items go to person with lowerst valuation, but results in everyone getting less than expected) or to consider not equal shares in total goods.

Re:sibling fairness

[radtea]

the algorith is old one, I remember it from Hugo Steinhaus's math book.

That's a really nice description. I wish this was better known. But...

The algorithm only works (in the sense of leaving the parties psychologically satisfied) if their preferences are transitive (that is, if they are not insane).

In reality, even sane people's preferences change in pseudo-non-transitive ways as possibilities become actualities. So when Caleb gets the car, Adam is going to wish he'd valued it more highly, and so on. Our inner monkey won't be happy until it gets more than everyone else.

There is also a considerable body of data showing that our ability to judge the value of stuff is very poor. Happiness research has been big on this, showing that most of what people think will make them happy is radically inferior to easily predictable things that will actually make them happy.

So while the algorithm is beautiful and general and ought to be used wherever appropriate, it is not going to satisfy people, and it will then fall out of use because no one is going to say, "I am broken" when they can say "The algorithm is broken" instead.

Re:sibling fairness

[Kvasio]

still, there's nothing better ... or at least pareto-better.

Re:sibling fairness

[UnderCoverPenguin]

If employers want to claim it costs $17,000 to provide an employee a healthcare benefit, then the employer should have to pay that amount to the employee when employees decline that benefit.

Employers are not claiming that is their cost, they are reporting what the insurance company claims would be the "retail price" of an equivalent individual policy. Employers also report the actual cost as part of their business expenses. The IRS has a set of formulae for deciding if the employee owes more tax.

Re: sibling fairness

[MickLinux]

I'm putting a high value on the bathroom and the foyer. when I get those two rooms, I put a coin-op pay slot on the bathroom ... and sleep in the foyer.

Re:sibling fairness

[Kjella]

Except you assume that all the players are actually honest about every option and valuation and not just playing the algorithm. Imagine a well off and not so well off student decide to share an apartment which consists of two rooms, one small and one big and a common area. They can't agree how much the extra space is worth, now the poor student he doesn't want the big room. He just wants the other guy to pay this "fair" share extra so he can pay as little as possible. Now the rich student he knows the poor student is trying to cost optimize, he doesn't want to pay extra for extra space. So he doesn't have to bid a "fair" value, he just has to outbid the poor student.

If the poor student tries to push up the price of the big room or the rich student tries to penny pinch low-balling the price and they end up swapping rooms neither is going to be happy, the poor student didn't really want to pay extra and the rich student didn't really want the cramped room. They were just trying to use their own knowledge of the other person's preferences to maximize their own value. No algorithm can reasonably cope with that because the total utility is not fixed, it's not like a cake where the sum always adds up to 100%.

You can see this for example in divorce proceeding where something has different sentimental values. You want that family heirloom? Well I don't really want it, I'm just trying to gouge you for as much as I can. If you gouge too much the other person may give it up and say "fuck you too". The threshold is an essential secret which limits the gouging, if you knew that $100 item has a $1000 sentimental value you might bid $999 even though you don't want it for $101, you want to lose but in the most rewarding way.

It doesn't even have to be the big stuff it might be chocolates in a bowl, if you know your buddy is big on toffee you might grab one even though you don't like it much because you know your buddy will offer a good swap. Until he catches on and lets you simmer with your toffee because he knows that's a lose-lose situation you want to get out of too. It's like trying to stand still at a game of rock-paper-scissors, in theory they all have a 1/3 chance of winning but if you pick rock every time you'll soon lose every time. The way to win is to recognize your opponents patterns and exploit them for a better than 1/3rd chance.

Re:sibling fairness

[JasonGoatcher]

The best example of a fairness algorithm is an old one used with siblings. Tell one kid to divvy up the candy/cake/toys whatever, and let the other kid choose which half they want.

Greece used something similar when trying to get people to honestly report the value of their antiquities. If they listed it for a price the government thought was too low, the agency could buy it for that price. I wish local governments would do something like that with home values. If they want to tax me for a house worth more than what I can get for it on the open market, then I should have the right to sell it to them at that price.

Have you been reading old Heinlein novels? That's where I first encountered both of those concepts. Number of the Beast, right?

As for letting one sibling divvy stuff up and letting the other sibling pick, that's why democracy doesn't work in practice. Because however many people are involved, you need to account for each individually, so it gets out of hand after only a few people are added to the society.

Re:sibling fairness

[JasonGoatcher]

Well, if the person paying $500k finds out about the undervaluation, they should then sue for the difference.

Re:sibling fairness

[JasonGoatcher]

The algorithm only works (in the sense of leaving the parties psychologically satisfied) if their preferences are transitive (that is, if they are not insane).

In reality, even sane people's preferences change in pseudo-non-transitive ways as possibilities become actualities. So when Caleb gets the car, Adam is going to wish he'd valued it more highly, and so on. Our inner monkey won't be happy until it gets more than everyone else.

You could also pair people up with others who have different, but somehow compatible values. For instance, I want internet and a good way to consume it, a bedroom, and food and shelter. I'm willing to value that higher so I don't need to worry about the other stuff. So I could pair myself with someone who wants lots of the other stuff and is willing to take responsibility for it. So we're both happy.

There is also a considerable body of data showing that our ability to judge the value of stuff is very poor. Happiness research has been big on this, showing that most of what people think will make them happy is radically inferior to easily predictable things that will actually make them happy.

So while the algorithm is beautiful and general and ought to be used wherever appropriate, it is not going to satisfy people, and it will then fall out of use because no one is going to say, "I am broken" when they can say "The algorithm is broken" instead.

Good point, I've made so many tech purchases that I've regretted later, it's not even funny.

Re:sibling fairness

[russotto]

If the state sets the 'appraisal' value, then the owner should be able to force the state to purchase at the state's appraised price.

Doesn't work. The state just deliberately under-assesses everything, then jacks up the millage to compensate.

Re:sibling fairness

[Moral Judgement]

I know one of the reasons you had the valuations be different is to show how the system works and produces differences, but am I the only one thinking that Damon is getting screwed on step 3? Brad and Caleb get something and three times as much money as Damon (Adam's position is unclear without the market value of the house*). I know the main reason for this is his evaluation for the house is really low compared to the other three, and a rational actor wouldn't value the house so lowly because they can always sell it, but I feel if Damon was just not very clever he could do just such a thing (he could also do such a thing if he really just didn't want any of the inheritance, but that seems a lot less unfair). If we were talking about goods of less obvious value, it seems weird to me that Damon should get less because he lacks insight into their market value, keep in mind we're talking about an inheritance. I also understand this could be ameliorated by dividing the remaining $59,280 equally among the brothers, but in the example given Damon seems to be really losing out because of this low evaluation of the house.

I guess my point is that if every actor is rationally evaluating the objects then the algorithm seems to produce fair results. It does a good job of punishing lying, if you artificially lower your evaluations your payout is less and if you artificially raise your evaluations then you have to pay more into the pool. But if you are just irrational then you equally get punished.

*One could argue that whichever evaluation won is the market value of the house, as that party has effectively bought the house for that evaluation.

Re:sibling fairness

[cant_get_a_good_nick]

I had heard this too, in a Probate Case. It really works well when you have things of asymmetrical value. Dividing money is easy. Figuring out the relative worth of baseball cards to a rug wasn't so much in the Probate case.

Re:sibling fairness

[devman]

If the divider rigs the piles, you just take the pile you know they want and then barter with them afterwards. By dividing it that way the divider loses leverage it actually works out in favor of the picker.

Commercialism.

[ledow]

I think the requirement to have all your roommate's email addresses is the actual point of the website.

People can work out fair regimes. They can't spam their roommates easily enough, apparently.

Re:Commercialism.

[gurps_npc]

They ask for all the emails, but only require one person to actually give the email.

That is probably where they make their money.

But the divorce settlement technique is actually pretty useful. Amazing how often people are just total SOB's and DOB's over silly things.

The real problem there, however, is the desire to screw over, get revenge/compensation for other issues.

Re:Commercialism.

[taustin]

It is a not-for-profit website (http://www.spliddit.org/about).

Some of the most relentless spammers I've ever been victim of have been non-profits.

Re:Commercialism.

[Kvasio]

what's the problem here? Just use disposable email addresses. Eg a browser plugin, Donottrackme.

Re:View

[ArcadeMan]

Or the (lack of) value of a room adjacent to a noisy neighbour?

it's not so hard

[Anonymous Coward]

with 2 roomates I've done things this way: have one person set the prices of each room and the other person can then pick which room they want. this has worked well. the main practical complication I've had in real life is when one person makes much more money than the other person. the well off person would likely be willing to spend far more than the "fair" price in this situation for the better room. I don't think there's any clear cut solution that works in every case.

Re:it's not so hard

[guruevi]

The market tells you what is fair and what is not.

Let's say:
If I have 1 really large attic-room and 2 small bedrooms for rent, I could (locally) ask probably $350-450 per small room including utilities. For the large room even though it's 500sqft vs 150 sqft for the small bedroom, I can't ask >$600, because once you get closer to 800-1000, they may as well rent an entire house and sublet the bedrooms (although they would have to pay utilities).

The poorer person may want the bigger room because they also have dogs and work from home, the richer person may only need a place to crash while going to college. Both of them think they got a fair deal although per square footage, $400 is a heck of a lot of money for such a small room and $600 is very cheap for what amounts to a studio apartment (with a shared kitchen and bathroom). As the homeowner I pay $1000 between mortgage and utilities and thus it is objectively not "fair" for me to collect more than the amount I spend while living in the same house.

Yet that is what the market offers for the price points selected.

Re:it's not so hard

[Lehk228]

there is nothing objectively unfair for you, the owner, to make a profit on renting out a space. if it was wrong to do so how would landlords exist at all?

Re:it's not so hard

[devman]

If the well off person values the better room at more than "fair" price (fair being defined here as what the other roommate valued it at) that means the other roommate gets the other room cheaper than he would have valued it. That's win-win is it not?

The ice cream method

[drooling-dog]

To divide ice cream equally between two kids, have one dish it out and the other choose. My parents did this with my brother and me, and there was never anything to fight about afterwards. You'll never see more precise measurement in your life, though.

Re:The ice cream method

[Anonymous Coward]

I tried this once. One kid divided it up unfairly and stuck his finger in the bigger portion. The other kid would rather have the smaller portion than the tainted one.

The second kid didn't get as much as was fair, but the trash can got the first kid's share...

Re:The ice cream method

[grcumb]

But what if there are three people?

Then you vote on who cuts, and you can't vote for yourself. You clearly have never had a tabletop covered in dope in front of you.

Nor have I, of course, but I'm told that those who do, would do it that way.

In theory.

Re:The ice cream method

[swillden]

But what if there are three people?

http://en.wikipedia.org/wiki/Selfridge%E2%80%93Conway_discrete_procedure

Re:The ice cream method

[rahenri]

What if you want to share it between three kids?

Re:The ice cream method

[Kvasio]

look in this discussion for my post on Hugo Steinhaus. There is a good algorithm for splitting number of various things between any number of people.

Re:The ice cream method

[Anonymous Coward]

I did this with my little brother, and he always answered that he wanted the bigger one. I had to explain every time that he had to choose one of the bowls. Then, when he was older, he realized he could just spit on both...

Re:The ice cream method

[Vintermann]

That's the kind of ad-hoc solution that doesn't work, and people have actually been using math to find better ones. Go tell it to your dope fiend buddies.

Re:The ice cream method

[JimFive]

Person 1 scoops 3 bowls, person 2 can either choose a bowl or scoop into 2 bowls, person 3 likewise can choose or scoop into 2 bowls. Persons 1-3 pick bowls if they haven't already.

--
JimFive

Toilet etiquette

[agm]

Can we finally solve the age old question as to whether the seat should be left up or down? This is a function based on how many males vs females there are, and how often a male needs to, er, sit.

Re:Toilet etiquette

[Anonymous Coward]

Always down -- including the lid.

Re:Toilet etiquette

[Some_Llama]

it should ALWAYS be that the person sitting is in charge of making sure it is down.

Men sit down to poop, i don't hear them ever argue about who left the seat up, they just check before sitting.

common sense?

Re:Toilet etiquette

[orgelspieler]

You would be surprised at how popular a topic this is amongst economists

Re:Toilet etiquette

[OglinTatas]

Would she rather he lift the seat and leave it up or pee on it when it is left down? They both need to reach a compromise--one can't unilaterally impose a toilet seat rule.

Re:Toilet etiquette

[Anonymous Coward]

your argument doesn't work even though it is logical

most men are rational about everything except relationships

most women are the opposite

It is impossible to discuss anything rationally with a woman because she believes that her "feelings" have a value, and that value is infinite. Therefore the only thing of importance is that she gets the result that makes her "feel" good (usually involving the man making a lot of extra effort/resource expenditure)

It doesn't "feel" good for her to have to make the effort to check the seat and possibly lower it. Therefore its the man's job to expend the effort to prevent her from "not feeling good".

Once you realize that logic and reason are not useful tools when it comes to interacting with women then everything becomes much easier.

Re:Toilet etiquette

[SimonInOz]

Yeah - how come nobody ever asks "Why can't women put the damned seat up after them? They are so inconsiderate."

Maybe someone could hack one of those dopey toilet seat closers?

Re:Toilet etiquette

[nuckfuts]

Men sit down to poop, i don't hear them ever argue about who left the seat up, they just check before sitting.

Of course you don't "hear them ever argue". Who do you suppose the man might argue with, given that he would be the one who left the seat up?

Re:Toilet etiquette

[sexconker]

Given one man, one woman, one toilet. Women put the seat up when they're done, men put (or leave) the seat down when they're done.

The next user will likely be the opposite person, and they will have the seat set for them unless:

1 - They are a man who needs to take a shit after a woman last used the toilet. He has to put the seat down before shitting, but he then does not have to take an additional action afterward. No effort wasted here.

2 - They were the one who used it last. They have to undo and repeat their action.

Re:Toilet etiquette

[tlhIngan]

Can we finally solve the age old question as to whether the seat should be left up or down? This is a function based on how many males vs females there are, and how often a male needs to, er, sit.

In a domestic (i.e., household) setting, the simple compromise is toilet LID down. This way EVERYONE has to lift something to do something. If you're a lady, you lift the lid and the seat is down. If you're man doing #1, then you lift the lid and seat together. for #2, you lift just the lid.

Then when you're done, you put the lid back down. If also keeps the bathroom more hygienic as the act of flushing creates a plume of toilet water. Keeping the lid closed keeps that plume within the toilet and not the entire bathroom.

It has the advantage that stuff doesn't accidentally fall into the toilet too.

Re:Toilet etiquette

[Anonymous Coward]

Let me spell it out for you since you can see past a fucking pebble. Men don't get bitchy when women leave the seat down.

Re:Toilet etiquette

[Solandri]

Least-effort solution (minimum number of seat position changes) is for the person who uses the toilet to move it into the configuration they need it in, then leave it.

Least-agony solution (minimum number of gross incidents) is to always lower the seat after use. However, the fact that "men won't follow" this solution is merely coincidence - this solution happens to coincide with the configuration women always use so they can never be guilty of transgressing it. When I was living alone I had a dog who liked to drink out of the toilet. Consequently, I always told guests to lower the lid of the toilet after using it. My female guests always left the lid up. About half the men would lower it (probably due to being scolded about it by women all too often).

Re:Toilet etiquette

[Anonymous Coward]

Both the seat and the lid get closed after use every time because the cats will otherwise play or fall into the toilet. That is all.

Re:Toilet etiquette

[SimonInOz]

>> This is a function based on how many males vs females there are, and how often a male needs to, er, sit.

No it isn't.
The logical thing to do is the minimum possible - ie leave the seat where it is when the operation is complete. When the next user comes along they may - or may not - have to adjust it.

Or are you suggesting that women reverse blindly into a bathroom and sit down?

Re:Toilet etiquette

[Lehk228]

always close the damned lid, to not do so is degenerate

Re:Toilet etiquette

[volmtech]

My 90lb first wife was raised in a mostly female household. We had only been married a few days when from the bathroom I hear, "eek, splash".

Re:Toilet etiquette

[LordWabbit2]

Tell that to me (ex) wife. I never believed that a couple could have such a petty argument (we're all rational adults right). Boy was I wrong.

Re:Toilet etiquette

[kooky45]

This is what I do at work, but it's because too many colleagues are either too lazy or too scared to touch the seat and they so wee all over it. With the lid down they either have to touch it and learn that a messy seat is a hazard, or choose another toilet.

Re:Toilet etiquette

[u38cg]

Actually, closing the lid traps an aerosol of toilet water which blows up in the face of the next user. It persists a surprisingly long time. Citation: some Reader's Digest magazine I read in a dentist's waiting room twenty years ago.

Re:Toilet etiquette

[dotancohen]

Can we finally solve the age old question as to whether the seat should be left up or down? This is a function based on how many males vs females there are, and how often a male needs to, er, sit.

Pee on the seat twice, then she'll start begging you to leave it up!

Re:Toilet etiquette

[jittles]

it should ALWAYS be that the person sitting is in charge of making sure it is down.

Men sit down to poop, i don't hear them ever argue about who left the seat up, they just check before sitting.

common sense?

Hmmm. I get pissed at girlfriends when they come over and leave the seat down but the lid up. I hate staring into the gaping maw of the toilet bowl every time I need to walk into the bathroom. Furthermore, at a previous home, someone put a cabinet above the toilet. Can't tell you how many times I accidentally dropped something into the bowl from the cabinet shelf. Now the seat and the lid are second nature to me.

Re:Toilet etiquette

[CmdrTamale]

It has the advantage that the cat doesn't fall into the toilet, too.
FTFY
--
"Gravity just affects how distances are measured." Tell that to a skydiver's next of kin.

So... BillPin?

[Wootery]

Sounds a lot like BillPin.

saw in a MAD magazine comic

[Some_Llama]

The best fairness concept i have ever seen involves two people trying to split one thing evenly, the way it works is one person does the splitting and the other gets to make the first choice of which half they want. this article reminds me of this concept.

Re:saw in a MAD magazine comic

[Some_Llama]

and in rush to post my brilliant info i faled to notice the 50 other people who know this already.. yay!

I can't see anything wrong with this

[Minwee]

Once you have an algorithm that is provably fair, all you need to do is prove that the algorithm itself was chosen in a provably fair manner.

By another algorithm which is provably fair. With values which were chosen in a provably fair way. And so on. And so on.

As long as that chain of logic doesn't end with "BECAUSE I SAID SO NOW SHUT UP AND PAY YOUR SHARE!" then everybody should be satisfied by how fair everything is.

Re:View

[orgelspieler]

Now that's a smart algorithm!

solved: Up proves that it was raised. Also, forcin

[raymorris]

The male wants to avoid unnecessary raising and lowering - conservation of energy.
The female wants TWO things - she wants the seat lifted before the gentleman urinates, AND she wants it lowered before she does.

If the gentleman leaves the seat up, the female (provably) got the first thing she wants - the seat was raised before he urinated. Each party ends up doing the same amount of effort - they either raise or lower the seat before using it. That's fair.

The other option, that the seat is left down, means that a) the man is expected to do 100% of the work, both raising and lowering, while the female does none. More importantly from her point of view, if the seat is down, she doesn't know whether or not it was down when he peed. She might be sitting on pee spots.

Fairness, and her own piece of mind, therefore dictate that he leave the seat up.

However, if she's clever, she can't gently force the seat to be lowered afterwards by placing a tray of soaps, potpourri, etc. on top of the toilet. The tray will prevent the seat from being raised all way to vertical and gravity will ensure it ends up down. The clever gentleman can respond to this forcing function by pointing out that it prevents her from knowing whether male guests ever raised the seat at all.

e-split? iSplit? splitX? splittodyne? splitrotron

[Thud457]

too mny fckng vwls, shld hv calld it it splittr

Real problem is emotions

[gurps_npc]

People would never have a problem with dividing things if they didn't get all emotional about it.

Divorce is the prime example - it's rarely about the 2nd home, the dog, the china, etc.

It's about the cheating, the 'stealing the best years of my life', the drug addiction, etc.

Nine times out of ten, people are not really looking for 'fair', they want 'JUSTICE' (in quotes).

A pity, because in reality, 'JUSTICE" is another word for spending all your time and money on lawyers to punish someone else.

If their algorithms won't let you spend all your time and money on punishing your opponent, it won't actually solve all the problems of sub-dividing property.

Re:Real problem is emotions

[david_thornley]

In my divorce, I decided I wasn't going to put up a fuss about owning things. I could replace them, and I wanted a divorce with as little fuss as possible.

It may have helped that, when my soon-to-be-ex-wife got the papers, she seemed to think I needed her to sign them, so she didn't file an objection to what I thought reasonably fair. (The lack of kids and real estate also helped a lot.)

Complaining about free room and board?

[raymorris]

the main practical complication I've had in real life is when one person makes much more money than the other person. the well off person would likely be willing to spend far more than the "fair" price in this situation for the better room. I don't think there's any clear cut solution that works in every case.

Suppose the rent on the two bedroom place is $1,000. The "rich" roommate over bids as you suggest and says he'll pay $800 for the room on the left. Fine, that leaves the other roommate to cover only $200 for the other room. One can hardly complain about getting half of a $1,000 apartment for $200.

Cake cutting

[aaaaaaargh!]

Cake-cutting algorithms, I love them. I've read parts of the book by Robertson and Webb and always wonder about unusual and new applications for them.

Re:won't work with my kids

[Mr D from 63]

Of course, for the splitting rent, the most fair approach is for the wealthiest roommate to pay more.

Re:won't work with my kids

[Anonymous Coward]

And to get the nicer room (since he is paying more).

Re:solved: Up proves that it was raised. Also, for

[Quirkz]

There's also the "put things away when you're done" and "leave things the way you found it" and principles, both of which call for a closed lid.

Plus the "it's completely disgusting that my pets drink out of the toilet so please close the lid before they not only get sick but track toilet water everywhere" variant that doesn't apply to everyone, but is critical when it does apply.

Re:solved: Up proves that it was raised. Also, for

[bn557]

This is why I always pee in the shower. Problem solved!

Here's a great method for splitting chores

[PapayaSF]

We did this back in college, and it worked great.

1. – Make a list of all chores that need to be done every week.
2. – Agree on a point value for each one, with more points for longer or less pleasant chores.
3. – Divide the total points by the number of roommates, so everyone has X points to do per week.

The real genius of the system then comes in: whoever does their chores first gets to pick which ones to do, and whoever puts it off until the end has to do whatever's left. So there's a built-in incentive to do chores early, and no squabbling, because everyone agreed to the point rankings ahead of time.

Re:Here's a great method for splitting chores

[david_thornley]

Another neat thing about this: when somebody picks which chores to do, he or she picks the least onerous chores for the point value. That takes advantage of any differences in how unpleasant people think chores are. I'd likely wash all the dishes, for example, since I don't mind it as much as some people do. (Also, many people don't wash up to my standards, so there's another incentive that is not accounted for.)

Re:How is this news for nerds

[sexconker]

Your ideas are intriguing to me and I wish to subscribe to your newsletter.

Re:solved: Up proves that it was raised. Also, for

[mjwx]

It sounds like you're applying a rational process to women... Stop it.

Hold up now

[fyngyrz]

I was about to say maybe this would be too complicated to get right... then I remembered our tax code.

Are you, in any way, implying that the government got the tax code right?

Re:Hold up now

[disambiguated]

Oops, I guess I did imply that. I just meant that their are good reasons why the tax code is complicated, and that shouldn't prevent the taxation of intellectual property. The whole system needs massive reform anyway; companies that own a lot of IP already put them in holding companies and then licence them back to themselves.

Re:solved: Up proves that it was raised. Also, for

[cascadingstylesheet]

The other option, that the seat is left down, means that a) the man is expected to do 100% of the work, both raising and lowering, while the female does none.

You're probably not serious, but just in case ...

1. Women can be hard to please
2. This is an easy way to please them

Just sharing what little wisdom I've picked up over the years ..

Etiquette my back; *rules*

[fyngyrz]

We had cats and kids at one point. Letting cats (or any other dependent) drink water from a toilet is an act of stupidity. Eventually, they're going to catch something, and it may be very nasty -- even nasty enough that you might get it.

So... I set an alarm up that would go off in the kid's bathroom (they weren't allowed in ours) if the seat was up and the bathroom door was open.

The rules were: (1) if the alarm went off, they lost 10% of their allowance. (2) If the alarm was found to be disabled in any way, they lost their allowances permanently.

Worked great. The one time there was a problem with the alarm -- magnetic reed relay had stuck -- I was told before the kid left the bathroom. That seat and lid stayed down and the cats remained healthy. To this day, those kids (now in their 30's) don't leave seats and/or covers up. They know why, and they're properly habituated so it's not an issue.

Re:Etiquette my back; *rules*

[Anonymous Coward]

Cool system, but...

Cat's are a million times tougher than you think.

Dogs are even more so. I've seen watched many dogs lap up shit (from a wide range of animals, carnivores and omnivores) and guzzle it down, doing so all their lives, and still remained perfectly healthy.

Re:Etiquette my back; *rules*

[fyngyrz]

That's all fine... until something gets them. Better not to count on tough immune systems and provide a clean environment instead.

I don't even let them drink out of a bowl. I use an IR sensor to drive a 12v valve, which runs a stream of water they lap out of the air from our RO system. My cats have consistently been living past 20 years. Pretty happy with that.

you'll just get the room you said you wanted

[raymorris]

-> Specifically, if you express a value for the rooms as almost the same as your roommate, but slightly lower for the rooms he wants the most and slightly higher for the rooms he likes the least, I suspect their mechanism for overdemanded and underdemanded rooms will give you the slightly inferior rooms at a greatly reduced price

You'll only screw yourself that way. Suppose he values the first bedroom at $600 and the second at $500, because the first bedroom is better in some way.

You suggest going lower on the room he wants, so $550, and slightly higher for the room he doesn't want, so again $550. So you've said that you don't care which room you get. Since you don't care and he does, he'll get the better room. You said it's worth $550, he said $600, so he'll end up paying about $575.

Most of these algorithms automatically account for lying in the same way - the try to give you what you say you want. If you lie and say you want to sleep in the dumpster, you'll get what you asked for. The only time you won't get what you asked for is if someone else asked for the same thing.

What the algorithms don't cover well is if preferences vastly differ between people, AND you know what the other person's true preferences are. Suppose a brother and sister are dividing up a Barbie, a GI Joe, and a slice of cake. Brother can gain an advantage by OVER valuing the Barbie and under valuing the GI Joe, but not by too much - he doesn't want to actually end up getting the Barbie. He wants to end up with the "undesireable" GI Joe AND the cake. However , if the sister does the same thing, undervaluing the Barbie and overvaluing the GI Joe, they'll likely end up with what neither of them wanted.

Did it wrong...

[pubwvj]

"All it takes is a pair of siblings to prove that there's no such thing as "provably fair," non-mathematically."

Obviously you did it wrong, you raised your kids wrong. They should be cooperating because they are fundamentally pack, team members and because they are mathematicians. You are illogical.

no seriously, pee all over one commode

[raymorris]

I wasn't serious. My actual strategy is to leave drips all over whichever restroom does not contain her makeup. She'll quickly stop using that restroom and only use "her" bathroom, the one with her makeup in it. Bonus points if her bathroom is also the guest bathroom.

Totally kidding- I want my woman to be happy when she has her pants down.

Congratulations to the researchers

[thisisauniqueid]

Congratulations to the researchers. They just re-invented PageRank.

Warmup to Voting Reform

[drewm19801927]

If enough people start using tools like this and trusting game theorists, perhaps we will ultimately be able to get non-academics on board to fix our broken voting system. The American voting system was designed by people who thought slavery was fair, and was ultimately a compromise between people who thought land ownership should grant rights, and people who thought every white male should have equal rights.

Re:Rubbish

[devman]

Easy, the lactose intolerant party takes the 16 cakes, trades 6 of them with cream to the other party for the 6 he has without cream and throws the remaining 5 cakes with cream in the garbage. Even if the other party doesn't trade the picker gets to eat 5 and the divider gets to eat 6, but none of them with cream.

Rigging piles always works to the advantage of the picker not the divider, that's why the system is fair.

Self-assessment of property value for tax purposes

[eric_harris_76]

Something like that was done for real estate assessment in fiction (some Robert Heinlein novel) and also (I think) in reality (Singapore, perhaps). As I recall, in the fictional instance, there was another wrinkle: anyone could buy it for your assessment of its value (and had to pay the difference of the taxes for the previous 3 years, to boot). The purchase price you pay to the owner; the increment on the tax you paid to the government.

Leibniz

[mcswell]

"The only way to rectify our reasonings is to make them as tangible as those of the Mathematicians, so that we can find our error at a glance, and when there are disputes among persons, we can simply say: Let us calculate, without further ado, to see who is right."