# Khan/khan-exercises

Switch branches/tags
Nothing to show
Fetching contributors…
Cannot retrieve contributors at this time
158 lines (155 sloc) 11.8 KB
 Combinations
randFromArray([ new Plural(function(num) { return i18n.ngettext("shirt", "shirts", num); }), new Plural(function(num) { return i18n.ngettext("tiger", "tigers", num); }), new Plural(function(num) { return i18n.ngettext("book", "books", num); }), new Plural(function(num) { return i18n.ngettext("Fabergé egg", "Fabergé eggs", num); }), new Plural(function(num) { return i18n.ngettext("sock", "socks", num); }), new Plural(function(num) { return i18n.ngettext("action figure", "action figures", num); }) ]) randRange(5,9) randRange(3,5) factorial(NUM_THINGS) / (factorial(NUM_THINGS-NUM_TAKEN) * factorial(NUM_TAKEN))

person(1) is packing his bags for his vacation. He has NUM_THINGS unique plural_form(THING, NUM_THINGS), but only NUM_TAKEN fit in his bag.

person(1) is packing her bags for her vacation. She has NUM_THINGS unique plural_form(THING, NUM_THINGS), but only NUM_TAKEN fit in her bag.

How many different groups of NUM_TAKEN plural_form(THING, NUM_TAKEN) can he take?

How many different groups of NUM_TAKEN plural_form(THING, NUM_TAKEN) can she take?

person(1) has NUM_TAKEN spaces for his plural_form(THING), so let's fill them one by one. At first, person(1) has NUM_THINGS choices for what to put in the first space.

person(1) has NUM_TAKEN spaces for her plural_form(THING), so let's fill them one by one. At first, person(1) has NUM_THINGS choices for what to put in the first space.

For the second space, he only has NUM_THINGS-1 plural_form(THING, NUM_THINGS-1) left, so there are only NUM_THINGS-1 choices of what to put in the second space. So far, it seems like there are NUM_THINGS \cdot NUM_THINGS-1 = NUM_THINGS * (NUM_THINGS-1) different unique choices person(1) could have made to fill the first two spaces in his bag. But that's not quite right.

For the second space, she only has NUM_THINGS-1 plural_form(THING, NUM_THINGS-1) left, so there are only NUM_THINGS-1 choices of what to put in the second space. So far, it seems like there are NUM_THINGS \cdot NUM_THINGS-1 = NUM_THINGS * (NUM_THINGS-1) different unique choices person(1) could have made to fill the first two spaces in her bag. But that's not quite right.

Why? Because if he picked THING number 3, then THING number 1, that's the same situation as picking number 1 and then number 3. They both end up in the same bag.

Why? Because if she picked THING number 3, then THING number 1, that's the same situation as picking number 1 and then number 3. They both end up in the same bag.

So, if person(1) keeps filling the spaces in his bag, making _.map(_.range(NUM_TAKEN), function(l){ return (NUM_THINGS - l);}).join("\\cdot") = \dfrac{NUM_THINGS!}{(NUM_THINGS-NUM_TAKEN)!} = factorial(NUM_THINGS)/factorial(NUM_THINGS-NUM_TAKEN) decisions altogether, we've overcounted a bunch of groups.

So, if person(1) keeps filling the spaces in her bag, making _.map(_.range(NUM_TAKEN), function(l){ return (NUM_THINGS - l);}).join("\\cdot") = \dfrac{NUM_THINGS!}{(NUM_THINGS-NUM_TAKEN)!} = factorial(NUM_THINGS)/factorial(NUM_THINGS-NUM_TAKEN) decisions altogether, we've overcounted a bunch of groups.

How much have we overcounted? Well, for every group of NUM_TAKEN, we've counted them as if the order we chose them in matters, when really it doesn't. So, the number of times we've overcounted each group is the number of ways to order NUM_TAKEN things.

There are NUM_TAKEN! = factorial(NUM_TAKEN) ways of ordering NUM_TAKEN things, so we've counted each group of NUM_TAKEN plural_form(THING, NUM_TAKEN) factorial(NUM_TAKEN) times.

So, we have to divide the number of ways we could have filled the bag in order by number of times we've overcounted our groups.

\dfrac{NUM_THINGS!}{(NUM_THINGS - NUM_TAKEN)!} \cdot \dfrac{1}{NUM_TAKEN!} is the number of groups of plural_form(THING) person(1) can bring.
Another way to write this is \binom{NUM_THINGS}{NUM_TAKEN} , or NUM_THINGS choose NUM_TAKEN, which is ANSWER.

randRange(5,6) randRange(2,FRIENDS-2) factorial(FRIENDS) / (factorial(FRIENDS-SLOTS) * factorial(SLOTS))

You just got a free ticket for a boat ride, and you can bring along SLOTS friends! Unfortunately, you have FRIENDS friends who want to come along.

How many different groups of friends could you take with you?

There are SLOTS places for your friends on the boat, so let's fill those slots one by one. For the first slot, we have FRIENDS different choices we can make (because FRIENDS different friends could be in that slot).

Once we've filled the first slot, there are FRIENDS-1 friends who could fill the second. So far, if we've filled the first two slots, and it seems like there are FRIENDS \cdot FRIENDS-1 = FRIENDS * ( FRIENDS-1) different choices we could have made. But that's not quite true.

Why? Because if we picked person(1), then person(2), that's the same thing as picking person(2), then person(1). They both get to be on the same boat.

So, if we continue filling the slots on our boat, making _.map(_.range(SLOTS), function(l){ return (FRIENDS - l);}).join("\\cdot") = \dfrac{FRIENDS!}{FRIENDS-SLOTS!} = factorial(FRIENDS)/factorial(FRIENDS-SLOTS) decisions altogether, we've overcounted a bunch of groups.

How much have we overcounted? Well, for every group of SLOTS, we've counted them as if the order we chose them in matters, when really it doesn't. So, the number of times we've overcounted each group is the number of ways to order SLOTS things.

There are SLOTS! = factorial(SLOTS) ways of ordering SLOTS things, so we've counted each group of SLOTS friends factorial(SLOTS) times.

So, we have to divide the number of ways we could have filled the boat in order by number of times we've overcounted our groups:
\dfrac{FRIENDS!}{FRIENDS-SLOTS!} \cdot \dfrac{1}{SLOTS!} is the number of groups we can bring on the boat trip.
Another way to write this is \binom{FRIENDS}{SLOTS} , or FRIENDS choose SLOTS, which is ANSWER.