If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

Main content

Getting exactly two heads (combinatorics)

A different way to think about the probability of getting 2 heads in 4 flips. Created by Sal Khan.

Want to join the conversation?

  • blobby green style avatar for user Gina
    To solve the second question of finding the probability of flipping 2 heads, is there an equation of formula to quickly solve the question?
    (32 votes)
    Default Khan Academy avatar avatar for user
  • piceratops seed style avatar for user ShehaЯyaЯ Khan
    there was a question on this website "A fair coin is flipped four times. What is the probability of getting tails at least twice? Write your answer as a simplified fraction". I did it like the one said here...but it said the answer is 11/16(which I can get by counting) but my answer was 3/8 using this exact method, i was going along ???? Please tell me what i'm doing wrong?it took my red bar to blue bar...it was gonna make me cry..HELP
    (15 votes)
    Default Khan Academy avatar avatar for user
    • marcimus pink style avatar for user Gintė Petrulionytė
      In the question it says AT LEAST two heads, so you don't just count the posibilities of getting exactly two heads, you also have the possibilities of getting three or four heads. This method talks about getting Exactly two heads, not 2, 3 or 4.
      If you want to use combinatorics: it's 6/16 (as shown in the video) + 4/16(the possibility of getting three heads 4*3*2/3*2=4 of 16)+ 1/16 (the possibility of getting 4 heads)=11/16
      (50 votes)
  • male robot hal style avatar for user Olivier Robles
    hey im still not quite understanding the intuition behind why under 4 places, each with 2 outcomes i can get 16 possibilities. Sure if i were to write out each possible outcome I'd indeed arrive at 16. But how do I know by going 2x2x2x2
    (7 votes)
    Default Khan Academy avatar avatar for user
    • leaf blue style avatar for user Dr C
      I think the easiest way to understand this is to think of a tree diagram. At each point there are two branches coming off of the previous one. Using coin flips, after 1 flip we have 2 branches: heads and tails. At the second flip we have two branches off each of the original two branches, doubling the number of branches (4 total - HH, HT, TH, TT). At the third flip, each of these 4 branches has two new branches coming off of it for a total of 8. And so on, with every flip we are doubling the number of branches. Hence the total number of possible outcomes is 2^x, where x is the number of flips we make.

      Another way to think of it is that you have a particular sequence of flips. Let's say we have two flips, I'll call them F1 and F2. The sequence so far is then (F1, F2). Using your reasoning, you can figure out that there are four ways for this sequence of two flips to have occurred. Now let's add a third flip. There are two options: heads or tails. To count the number of possible outcomes for the entire sequence, we take the entire sequence (4 ways), and put an H on the end, or we can take the entire sequence and put a tails on the end. So we have 4 ways (F1,F2,H) plus 4 ways (F1, F2, T). So we've taken the total number of possible outcomes thus far, and doubled it to account for each of the potential outcomes of the next flip.
      (19 votes)
  • aqualine ultimate style avatar for user Sarah
    In the last example of exactly two heads, I'm not clear on how that's calculating exactly two rather than at least two. We calculate the likelihood of the two heads appearing in each of the places, and clearly that's correct because it gets the right answer compared to just counting the possibilities, but how is the equation constraining out the situations where you have three or four heads?
    (6 votes)
    Default Khan Academy avatar avatar for user
    • duskpin ultimate style avatar for user Anwar
      Here's another way to think about it. We want exactly 2 heads and as a result 2 tails in 4 flips. Let's denote them as Ha, Hb, Ta, and Tb.

      How many different ways we can arrange these 4 items? Well, that would be 4! or 4 * 3 * 2 * 1. We don't need to differentiate between Ha and Hb and double count them. For example, we want HaHbTaTb and HbHaTaTb counted as one instead of two. So, we divide by 2! to account for double counting. Same with Ta and Tb. So, we divide by another 2! to cancel out double counting of two T's.

      4! / (2! * 2!) = 6

      Finally, if we divide all 6 different ways of getting exactly 2 heads (and 2 tails) in 4 flips by all possible outcomes 2 * 2 * 2 * 2 = 16 we would get the probability of exactly 2 heads in 4 flips.

      6 / 16 = 3 / 8
      (7 votes)
  • duskpin ultimate style avatar for user newtodisworld
    Then, what do we say when there are more than one heads? Two headss?
    (3 votes)
    Default Khan Academy avatar avatar for user
    • aqualine ultimate style avatar for user wchargin
      Yes, you would say two headses. This is because "heads" is a single state, so you're really saying two "heads"es. Although, most people would probably say two heads. :)

      You could also call them "successes" and "failures," which would eliminate the need for this odd terminology.
      (7 votes)
  • aqualine seed style avatar for user Ben Miller
    The majority of operations involved with probability involve multiplication, even after watching intro videos and working through the modules I'm having a hard time in general intuitively understanding why. Specifically at , why use multiplication? Couldn't you break it down into complicated addition at some point?
    (5 votes)
    Default Khan Academy avatar avatar for user
  • spunky sam blue style avatar for user surabhi.shivmath
    Can we solve this the same way we solve word problems in combinations? Eg: How many different ways are can we write the word "THAT" uniquely? The answer being --> 4!/2. Similarly we can apply for HHTT,but by selectively letting T repeat--> 4!/2. Am i going wrong somewhere?? Does this have any loop hole?
    (3 votes)
    Default Khan Academy avatar avatar for user
    • aqualine ultimate style avatar for user Saranyaraj Rajendran
      If you look at the video and solution, and reason through it, you'll find that's exactly what being done here.
      But before that, the number of different ways we can write (or re-arrange) the word "THAT" is 4!/2! not 4!/2 (2! for the different ways the two T's would have permuted if they were different, which they are not, so we are dividing that value from the whole).
      And going by that, the ways we can re-arrange HHTT (exactly two heads) will be 4!/(2!*2!) = 6. The first 2! in the denominator for the same two H and the next 2! in the denominator for the same two TT. This is if you look at it a combination given by 4C2 = 4!/[(4-2)! * 2!] = 6. Hope this make stuff a bit clearer for you.
      (4 votes)
  • duskpin ultimate style avatar for user Pixie
    Please help me understand and have intuition as to why 4 has to be multiplied by 3 and why there are 12 places. I've watched the permutations video and understood that by drawing diagrams, however, I can't quite seem to understand this.
    (4 votes)
    Default Khan Academy avatar avatar for user
    • leaf blue style avatar for user stems123
      We are flipping the coin four times.
      As Sal explained in the video, with Ha (Heads A), there are four potential "spots" or flips for us to get a heads. This is because we are flipping for the first time, meaning there are still four outcomes.
      Next, we have Hb (Heads B). Since we have already flipped a heads (Heads A), there are now only 3 spots/flips for an outcome of heads (or any outcome).
      For every spot/flip that we can get a heads the first flip (4 different spots for heads A), there are 3 other flips that we could get a heads for (heads B).
      You can create a tree diagram by having Heads branch out into 4 (since we are flipping the coin 4 times), and for each of those Heads, you would branch out 3 times since you have already flipped the coin once (4-1=3), and you would have three more flips afterwards to get another heads.
      We would then multiply: 4 * 3 = 12
      because for every possible heads (4 possibilities: first flip, second flip, third flip, or fourth flip), there would be three others.
      Hope this makes sense! I strongly recommend drawing a tree diagram that I mentioned above to help comprehend the reasoning behind this.
      (2 votes)
  • female robot grace style avatar for user kim16457
    There was this question on dependent probability, 'You have 6 coins in a bag. 3 of them are unfair in that they have a 55% chance of coming up heads when flipped (the rest are fair coins). You randomly choose one coin from the bag and flip it 3 times.

    What is the probability, written as a percentage, of getting 3 heads? Round your answer to the nearest hundredth of a percent.'
    and I had no idea how to solve it :(
    (4 votes)
    Default Khan Academy avatar avatar for user
  • blobby green style avatar for user shoshinsamurai
    For the second part of the question, I figured this below. I wanted to know if I might be right.
    The probability of getting any combination of 2 heads and 2 tails is (0.5)^2.(0.5)^2. = 1/16
    The number of ways to choose these two winning possibilities in 4!/(2!(4-2)!), which is 6.
    So the probability of getting the winning 2 heads will be 6*(1/16) or 3/8.
    Is this right?
    (2 votes)
    Default Khan Academy avatar avatar for user
    • old spice man green style avatar for user Bastian Widanski
      yes that is correct. small typo: 4!/(2!(4.2)!) should be 4!/(2!(4-2)!)

      another way to think about it:
      throwing a coin has two possible outcomes. So throwing it 4 times has 2*2*2*2 = 2^4 = 16 possible outcomes (with fixed order).
      and out of this 16 possibilities 4!/(2!(4-2)!) = 6 all the combinations of 2 one side with 2 of the other side. Or in other words all the possible positions (permutations) of the single outcomes.
      so overall you have 6 out of 16 possibilities -> the probability is 6/16 = 3/8 = 0,375 = 37,5%

      with such small examples you even could write down all possible outcomes. Just to see what it is that you are actually calculating:
      (H H H H)
      (T H H H) (H T H H) (H H T H) (H H H T)
      (T T H H) (T H T H) (T H H T) (H T T H) (H T H T) (H H T T)
      (T T T H) (T T H T) (T H T T) (H T T T)
      (T T T T)
      (3 votes)

Video transcript

I'm going to start with a fair coin, and I'm going to flip it four times. And the first question I want to ask is, what is the probability that I get exactly one head, or heads? This is one of those confusing things, when you're talking about what side of the coin. I know I've been not doing this consistently. I'm tempted to say, if you're saying one, it feels like you should do the singular, which would be head. But I've read up a little bit of it on the internet, and it seems like when you're talking about coins, you really should say one heads, which is a little bit, it seems a little bit difficult for me. But I'll try to go with that. So what is the probability of getting exactly one heads? And I put that in quotes to say, you know, really, we're just to talking about one head there. But it's called heads when you're dealing with coins. Anyway, I think you get what I'm talking about. And to think about this, let's think about how many different possible ways we can get four flips of a coin. So we're going to have one flip, then another flip, then another flip, then another flip. And this first flip has two possibilities. It could be heads or tails. The second flip has two possibilities. It could be heads or tails. The third flip has two possibilities. It could be heads or tails. And the fourth flip has two possibilities. It could be heads or tails. So you have 2 times 2 times 2 times 2, which is equal to 16 possibilities. 16 possible outcomes when you flip a coin four times. And any one of the possible outcomes would be 1 of 16. So if I wanted to say, so if I were to just say the probability, and I'm just going to not talk about this one heads, if I just take a, just maybe this thing that has three heads right here. This exact sequence of events. This is the first flip, second flip, third flip, fourth flip. Getting exactly this, this is exactly one out of a possible of 16 events. Now with that out of the way, let's think about how many possibilities, how many of those 16 possibilities, involve getting exactly one heads? Well, we could list them. You could get your heads. So this is equal to the probability of getting the heads in the first flip, plus the probability of getting the heads in the second flip, plus the probability of getting the heads in the third flip. Remember, exactly one heads. We're not saying at least one, exactly one heads. So the probability in the third flip, and then, or the possibility that you get heads in the fourth flip. Tails, heads, and tails. And we know already what the probability of each of these things are. There are 16 possible events, and each of these are one of those 16 possible events. So this is going to be 1 over 16, 1 over 16, 1 over 16, and 1 over 16. And so we're really saying the probability of getting exactly one heads is the same thing as the probability of getting heads in the first flip, or the probability of getting heads-- or I should say the probability of getting heads in the first flip, or heads in the second flip, or heads in the third flip, or heads in the fourth flip. And we can add the probabilities of these different things, because they are mutually exclusive. Any two of these things cannot happen at the same time. You have to pick one of these scenarios. And so we can add the probabilities. 1/16 plus 1/16 plus 1/16 plus 1/16. Did I say that four times? Well, assume that I did. And so you would get 4/16, which is equal to 1/4. Fair enough. Now let's ask a slightly more interesting question. Let's ask ourselves the probability of getting exactly two heads. And there's a couple of ways we can think about it. One is just in the traditional way. [? We ?] know the number of possibilities and of those equally likely possibilities. And we can only use this methodology because it's a fair coin. So, how many of the total possibilities have two heads of the total of equally likely possibilities? So we know there are 16 equally likely possibilities. How many of those have two heads? So I've actually, ahead of time so we save time, I've drawn all of the 16 equally likely possibilities. And how many of these involve two heads? Well, let's see. This one over here has two heads, this one over here has two heads, this one over here has two heads. Let's see, this one over here has two heads, and this one over here has two heads. And then this one over here has two heads, and I believe we are done after that. So if we count them, one, two, three, four, five, six of the possibilities have exactly two heads. So six of the 16 equally likely possibilities have two heads. So we have a-- what is this-- a 3/8 chance of getting exactly two heads. Now that's kind of what we've been doing in the past, but what I want to do is think about a way so we wouldn't have to write out all the possibilities. And the reason why that's useful is, we're only dealing with four flips now. But if we were dealing with 10 flips, there's no way that we could write out all the possibilities like this. So we really want a different way of thinking about it. And the different way of thinking about it is, if we're saying exactly two heads, you can imagine we're having the four flips. Flip one, flip two, flip three, flip four. So these are the flips, or you could say, the outcome of the flips. And if you're going to have exactly two heads, you could say, well, look, I'm going to have one head in one of these positions, and then one head in the other position. So how many, if I'm picking the first, so I have kind of a heads one, and I have a heads two. And I don't want you to think that these are somehow the heads in the first flip or the heads in the second flip. What I'm saying is we need two heads. We need a total of two heads in all of our flips. And I'm just giving one of the heads a name, and I'm giving the other head a name. And what we're going to see in a few seconds is that we actually don't want to double count, we don't want to count the situation, we don't want to double count this situation-- heads one, heads two, tails, tails. And heads two, heads one, tails, tails. For our purposes, these are the exact same outcomes. So we don't want to double count that. And we're going to have to account for that. But if we just think about it generally, how many different spots, how many different flips can that first head show up in? Well, there's four different flips that that first head could show up in. So there's four possibilities, four flips, or four places that it could show up in. Well, if that first head takes up one of these four places, let's just say that first head shows up on the third flip, then how many different places can that second head show up in? Well, if that first head is in one of the four places, then that second head can only be in three different places. So that second head can only be-- I'm picking a nice color here-- can only be in three different places. And so, you know, it could be in any one of these. It could maybe be right over there. Any one of those three places. And so, when you think about it in terms of the first, and I don't want to say the first head, head one. Actually, let me call it this way. Let me call it head A and head B. That way you won't think that I'm talking about the first flip or the second flip. So this is head A, and this right over there is head B. So if you had a particular, I mean, these heads are identical. These outcomes aren't different, but the way we talk about it right now, it looks like there's four places that we could get this head in, and there's three places where we could get this head in. And so if you were to multiply all of the different ways that you could get, all of the different scenarios where this is in four different places, and then this is in one of the three left over places, you get 12 different scenarios. But there would only be 12 different scenarios if you viewed this as being different than this. And let me rewrite it with our new-- So this is head A, this is head B, this is head B, this is head A. There would only be 12 different scenarios if you viewed these two things as fundamentally different. But we don't. We're actually double counting. Because we can always swap these two heads and have the exact same outcome. So what you want to do is actually divide it by two. So you want to divide it by all of the different ways that you can swap two different things. If we had three heads here, you would think about all of the different ways you could swap three different things. If you had four heads here, it would be all the different ways you could swap four different things. So there's 12 different scenarios if you couldn't swap them, but you want to divide it by all of the different ways that you can swap two things. So 12 divided by 2 is equal to 6. Six different scenarios, fundamentally different scenarios, considering that you can swap them. If you assume that head A and head B can be interchangeable. But it's a completely identical outcome for us, because they're really just heads. So there's six different scenarios, and we know that there's a total of 16 equally likely scenarios. So we could say that the probability of getting exactly two heads is 6 times, six scenarios and-- Or there's a couple of ways. You could say there are six scenarios that give us two heads, of a possible 16. Or you could say there are six possible scenarios, and the probability of each of those scenarios is 1/16. But either way, you'll get the same answer.