Wednesday assorted links


Because the power of marginalrevolution will likely crash my website, here is a cached link to the article "which NBA city has the greatest number of "fair weather" fans.

Is wins-attendance correlation the right statistic to use? That could flag consistent but small changes in attendance; covariance might be a better measure. I'd also think you'd want to lag attendance by up to a year. This year's seasons tickets sales depend much more on last year's performance than this year's.

Thunder history in Oklahoma City is relatively short and fairly successful but seeing Seattle among the top 5 “band wagon” fan spots, changing cities looks like an “inspired” decision.


The article is incorrect, that question is actually from the Singapore Math Olympiad (an exam designed to be extremely difficult to filter elite talent for the International Math Olympiad).

Thank God. I'm stuck.

A: "I don't know when Cheryl's birthday is" --> this gives me no information,

A: "but I know B does not know either." --> this gives me information.

B: "At first I didn't know" --> this gives us no information that the previous statement

B: "but I know now." --> I can't figure out this out.

I'm still puzzling over it.

The fact that B doesn't know is information and allows you to start eliminating dates.

The eliminations I make from the third step are the same as the eliminations I made from the second step.

I.e., when B says he didn't know at first, that means the date must be 14, 15, 16, or 17.

I did make a mistake in my statement 4. The fact that B knows after getting the hint from A lets me narrow it down to three dates. I can see how A could narrow it down to those same three.

[puzzle spoilers, obviously]

When B says he didn't know at first, that means he now knows that it he didn't have a unique day, as you said. But the fact that A knows that B doesn't know means that B can eliminate May and June. He now knows because his date is one of the unique dates in July or August. A then says he knows given that knowledge, which tells us that it must be July 16, because if A was told "August", he couldn't eliminate the 15th or 17th of august (the other two unique day numbers of July and August).

Okay, gonna put all my thinking here.

A was told May, June, July, or August. B was told 14, 15, 16, 17, 18, or 19.

A: "I don't know when Cheryl's birthday is". This gives us no information now.

A: "But I know that B doesn't know." This means that the month cannot be May or June, since each of those have a unique date that could potentially let B know.

B: "I didn't know at first." This tells us that the number cannot be 18 or 19, which would uniquely identify the date. However, we got the same information from the prior statement. This means that B's number was 14, 15, 16, or 17.

B: "But now I know." Of the 5 possible dates in July or August, 2 share a date. If B's number was "14" B still wouldn't know. But with 15, 16, or 17, B can uniquely identify, so B's number must be one of those three.

A: . . .

oh I get it now. Not posting so I can think about it a bit more and to not spoil it.

Assuming, that is, that B knows that A has been given the month, and that A knows that B has been given the date-- something the puzzle doesn't actually tell you, but which you lose nothing by assuming, as it's insoluble otherwise.

Furthermore, assuming that we know they are eavesdropping on each other. Since A refers to B by name, you'd expect he is not talking to B.

That B says "well, I know now" it implies that he got some new information, and what happened was A speaking, so it's a pretty minor leap to say that B knew from listening to A.

Also, assuming that they speak the same language, recognize each other as rational individuals, and are not bothered by current bouts of IBS.

@ DW

It's sloppily formulated, which creates unnecessary confusion. You already know A didn't know at first, and B doesn't need to confirm he didn't know at first either, unless we are expected to account for the possibility of one or both (or all three) of them lying. Try this in plain English with the required minimum of info.

A: I know you don't know the date.
B: In that case I do know.
A: In that case so do I.

And the fact that A knows that B doesn't know gives B more information, allowing more dates to be eliminated, etc.

A tells you it can't be May, or he wouldn't know that Bernard hadn't been told the 19th and already knew.

Going to give this a little more thought

Yeah, not sure how to get it down from July 16th and August 15th to one possibility

You need to think about what "know" means at each stage. For Bernard to be sure, there's only one possibility.

It's a nice little economical little problem.

Besides July 16 and Aug. 15, Aug. 17 was a possibility at that stage (step two only eliminated July 14 and Aug. 14.) Bernard knows the month, and that allows him to identify the date, because unlike us he knows the month; and the fact that he can do so in turn allows us, and Albert, to identify the date.

The first quote by A (broken into two statements in my retelling) tells you it must be July or August. If the month were June, it would be possible for B to know the exact date.

B knows the date and since he knows the month must be July or August, he says he knows the date and month, meaning the date must be 15, 16, or 17.

1. Albert knowing that Bernard doesn't know rules out days "18" and "19", and if Albert still doesn't know that also rules out June 17.
2. With those ruled out, Bernard claims to know. The last remaining unique day is 17, so August 17.

That's what I got. The only answer which satisfies all those conditions.

The first clue says (parsed): It isn't possible for the date to be unique in the total set (selects only July and August) --> subset 1 including only those two months.
The second clue says: The date is unique within those two months (eliminating the 14th from both), AND it is the only date in its row (subset 2) --> July 16th

Yeah, the wording made it, to my ear, insoluble.

(I believe, from seeing the "answer key", that the first "when" means, in Singapore English, "the date of", which does make it soluble.

But to my very-precise American English parser, it's non-informative.)

(In retrospect, though, as stated above - the fact that B was able to infer the solution from A's statement reveals which was told which set of numbers and one *can* infer the rest.)

I didn't understand the wording of the problem. I didn't quite get that Albert was saying, "Based on what I know, I can say for sure that Bernard doesn't know such-and-such." I thought he was just asserting stuff and I didn't really get the significance of it. Once that was made clear I thought it was not too bad.

A & B are assumed to be perfect logicians. This may seem to come from left-field for many readers, but if you are participating in a Math Olympiad you take it for granted.

The myth of homo economicus lives on. I'd prefer to see some heterodox logic puzzles.

Three logicians walk into a bar. The bartender asks, hello fellows, are you all feeling perfectly logical today?

The first answers: maybe.

The second answers: maybe.

The third thinks for a second, and answers: yes.

The bartender brings out a glass for each one, and starts: okay folks, how about we fix that...

A's statement contains information which combined with B's information yields a certain conclusion. Then the next statement. For an example of this type of problem, which might help you get the logic without spoiling the answer, see:

My deduction: July 16th.

1) Albert is told the month
If it was May or June, it could have been unique (the unique days were the 18th and 19th) and Bernard could potentially know the exact date. Since he could state with certainty that Bernard did not it must be July or August
Ergo, it was July or August
2) Bernard didn't know at first, but when he knew from Bernard's statement that it had to be July or August he could isolate it, and after isolating he knows the correct answer.
You can rule out the 14th because that's common to both July or August. So, Bernard was told a date of the 15th, 16th or 17th. And now since he knows it can't be in July or August he can correctly guess the full date.
3) Albert already knows the correct month. But now Bernard says he knows the correct date. If the correct month were August, it could be one of two choices either August 15th or August 17th.
Ergo it must be July 16th.

The enforced formatting killed the readability of the deduction, but to simply it: __

Statement A narrows the correct answer down to July or August. __

Statement B narrows the correct answer down to July 16th, August 15th or August 17th. __

Statement C narrows it down to July 16th.

Best straight explanation for folks like me.

Aparently they use it to filter students from the top half (well, top two-fifths) in order to find the best talent ("Being Q24 out of 25 questions, this is a difficult question meant to sift out the better students. SASMO contests target the top 40% of the student population and the standards of most questions are just high enough to stretch the students").
And I don't want to brag-because I am humblest man alive- but I got the right answer. It is comforting knowing I MAY be as smart as a member from the smarter half from the 14 year-old students cohort. My possibilities are limitless. Mounsieur Coué was right, "every day, in every way, I'm getting better and better"

The question is poorly worded. If they made it crystal clear what each one was told (and not told) I think more people could solve it quicker.

Probably. The context makes it work, I think, but a better-worded question would yield more (and quicker) right answer, yes.

Probably. The context makes it work, I think, but a better-worded question would yield more (and quicker) right answers, yes.

But the puzzle was passed on to the facebook poster via a P5 student, i.e. the teacher of the P5 class expected some of his/her students might be able to solved it.

Also note that Singapore is not in the "notable (team) achievement" list but a single entrant from there had managed "to win five medals with at least three of them gold".

Note Terence Tao (Australia) won a gold medal when he just turned thirteen. Thus it is possible that some P5 student might be able to solve it.

"View dining through the lens of recreation" but also be sure to avoid appetizers, a nice location, and alcohol.

Joyless economists are never going to be able to reconcile restaurant dining by coming up with a correct theory of how to choose one and how to behave when you get there. Once you frame the interaction as monetary warfare between the owner and the patron, you might as well stay home with a sack of value menu items.

Appetizers are silly in my opinion...location not really (unless you're alone).

Alcohol especially in a group does contribute to entertainment.

Remember Cowen's three laws...

Appetizers are great if you are eating at a place that doesn't put a big mark-up on them because they allow you to try more things. They are often just as good as the entrees.

Your assumption that Tyler is not enjoying his style of dining seems quite incorrect.

I think a large portion of personal philosophy ends up being about mastering the skill of being able to optimize without being self-conscious of the fact that you are optimizing.

3) So how does one apply a mind-set intervention at home?

Shock collars.

Like Stuart Smalley did.

#1: At face value, 40% of a "substantial" increase sounds really good, when considering how inflexible life satisfaction tends to be.


It's also possible that self-reported overall wellbeing does not map closely on average moment-to-moment experience.

That was my thought exactly. That 40% of a “substantial” improvement *remains after 8 months* looks to me like evidence that the effect of hedonic adaptation is quite limited. I'm somebody who substantially upgraded housing 15 months ago, and I still feel very strongly the subjective improvement in my life satisfaction.

It could still be a relative status effect, though. Some people in the communities got the housing, others didn't. If everybody had equally good housing, the effect might be reduced further.

4 - took longer to solve than it should have taken - about 4 minutes. Really trivial; embarrassed that I didn't see it sooner.

Yeah, I had the same thought, but I think it took me even longer to get there. A part of my problem is that I initially misinterpreted the problem due to the strange wording.

Obviously I was being a little self mocking in my post, but it IS true that it took me about 4 minutes, and once I figured it out I was annoyed that it took me that long. Oddly enough, while I did not initially misinterpret the problem, it was the first clue that I was stuck on for about 3 minutes - brain freeze - after that it was obvious.

Everyone who teaches price theory should read Adam Minter's book.

Is this a Singapore puzzle question?

Well, I'm not really sure whether or not it's a Singapore puzzle question, and apparently you can't figure it out either. But now that I've told you that, you know, and so do I.

Sure, but are you better looking than a filipina 14-year-old?

Not really sure what mind-set intervention is but it sounds like it would be useful. Lots of people don't have an improvement mindset. Let them know how to make the most of their time in the cage.

The question isn't particularly difficult. It just requires you to read v carefully. Many will zero in on three dates only to get stuck, because they havent read the third clue carefully.

Questions of about the same level abound in CAT - the premier Indian MBA entrance exam. It's fairly routine, by Indian standards.

So the return on investment for buying a politician is higher than investing in a new product or service.

#3: The analysis in that paper seems pretty dubious. They have a big sample, but the effects are barely significant at p<0.05. They don't report the effects across the whole sample, but rather compare selected subsamples. They compare grades only in "core subjects", even though they have the data for "noncore subjects", too. They report all their results adjusting for school (?), race, gender, and prestudy GPA, so there's no way to tell if their intervention has any effect across the whole sample.

6. The end of imperfect information /s

Wasn't able to solve the puzzle because I'm good at math and logic, solved it because I know how tests are written. I wonder if that's common, and if it says anything about Singapore education.

Comments for this post are closed