double factorial vs. regular factorial

Поділитися
Вставка
  • Опубліковано 3 січ 2025

КОМЕНТАРІ • 496

  • @ethanpfeiffer7403
    @ethanpfeiffer7403 6 років тому +489

    Logic: A double factorial gives you a smaller number than a single factorial.

    • @BigDBrian
      @BigDBrian 6 років тому +29

      Ethanol 314 maybe it should be called a fractional factorial, notated with n factorial to the power of one half

    • @wompastompa3692
      @wompastompa3692 6 років тому +45

      n! = n!! × (n-1)!!

    • @dagatho
      @dagatho 6 років тому +39

      Logic: (1/2)^2 < (1/2)^1

    • @earthbjornnahkaimurrao9542
      @earthbjornnahkaimurrao9542 6 років тому +14

      yeah this should be n(!^(1/2)) imo.
      this would be consistent with defining n(!^0) to just equal n
      and higher powers should get bigger faster while fractional powers should approach n as the fraction approaches 0.

    • @kenhaley4
      @kenhaley4 6 років тому +27

      How about calling it a fractorial?

  • @AbhimanyuKumar_23
    @AbhimanyuKumar_23 6 років тому +199

    The given problem transforms to a limit of (2k)!/(2^k*k!)^2 where k approaches infinity. Using Ramanujan's or maybe Stirling's approximation on factorials, the limit reduces to 1/(pi*k)^0.5 where k approaches infinity so we get the answer as zero.

    • @AbhimanyuKumar_23
      @AbhimanyuKumar_23 6 років тому +7

      Thank You blackpenredpen

    • @yamangoyal5767
      @yamangoyal5767 6 років тому +3

      bprp Please pin this comment.

    • @clockfixer5049
      @clockfixer5049 5 років тому +1

      Isn't it just (2k+1)!!/(2k)!! ?

    • @samuelgarcia8571
      @samuelgarcia8571 5 років тому +1

      @@clockfixer5049 Almost... actually it's the limit of (2k-1)!!/(2k)!! but how do you calculate that ? You have to develop it as Abhimanyu said.

    • @MarceloSilva-lh9mh
      @MarceloSilva-lh9mh 5 років тому +5

      Great job! I wonder if there is a way to prove it using logic or induction.

  • @Lucroq
    @Lucroq 6 років тому +28

    Double factorial looks really excited, like you should shout it. SIX!!

    • @blackpenredpen
      @blackpenredpen  6 років тому +6

      Lucroq lol yea. There also the triple one haha

  • @jzanimates2352
    @jzanimates2352 6 років тому +13

    Hey. I really love your videos. I’ve always loved experimenting with maths and numbers, not just learning simple things in school. Your content is really good so keep up the good work!

  • @beng2995
    @beng2995 6 років тому +68

    Here's something interesting. Substitute k=0 into the odd case to find that (-1)!!=1.

    • @blackpenredpen
      @blackpenredpen  6 років тому +14

      Ben G lol nice!

    • @beng2995
      @beng2995 6 років тому +2

      Jinger McBlabbersnitch you're right. In that case, you could define n!! for any negative integer and 0

    • @EspGameplayer
      @EspGameplayer 4 роки тому +3

      @@blackpenredpen But, as n has to be positive, 2k and 2k-1 have to be positive too

    • @monicadickson4842
      @monicadickson4842 4 роки тому

      @@beng2995 nice one

    • @Vordikk
      @Vordikk 2 роки тому

      It has been said that k should be positive integer (natural number)

  • @weerman44
    @weerman44 6 років тому +66

    I love to see that my "FactOREO" is still being used :D

    • @blackpenredpen
      @blackpenredpen  6 років тому +13

      Yay!!! That was a great one! I loved it!

  • @TheReligiousCrap
    @TheReligiousCrap 6 років тому +25

    Never thought about this. Fascinating!

  • @Lucaazade
    @Lucaazade 6 років тому +5

    Spotted this straight away when you wrote 6!! = 6×4×2 and 5!! = 5×3×1
    Recursive/inductive definition: n!! = n! / (n-1)!!

  • @15schaa
    @15schaa 6 років тому +111

    See what you did in the title there.

    • @lukekennedy4970
      @lukekennedy4970 6 років тому +2

      Duncan Schaafsma 2!!!

    • @15schaa
      @15schaa 6 років тому +1

      That's a triple factorial.

    • @lukekennedy4970
      @lukekennedy4970 6 років тому +1

      I was wrong the title is actually”(!!)!!”

    • @15schaa
      @15schaa 6 років тому

      6!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

  • @callumvlex7059
    @callumvlex7059 6 років тому +3

    How I think about whether you should use 2k-1 or 2k+1 is whether you want to start counting at 1 or 0, it's how you pair up the odds and evens with their "internal" integer.

  • @pawepajak7142
    @pawepajak7142 3 роки тому +3

    WP to this Sion he played safe during laning phase, minimized deaths and yea he let some cs go but now hes scaled and is working with his team to win the game! Well played to this Sion!

  • @calyodelphi124
    @calyodelphi124 6 років тому +2

    Another observation that can be made: n! = n!!*(n-1)!! for all n in N+
    Also, if n is even, you can also say n!! = n!/(n-1)!! for all n in N+, but that probably needlessly complicates matters a bit. It's still an observation of an equivalent expression at the least.

  • @colorfulcalculus4526
    @colorfulcalculus4526 6 років тому +2

    First, we can express the denominator as 2^n *(n)! where n -> oo. Then, multiply both numerator and denominator by 2*4*6*.... Thus, we have: n! / [(2^n * n!) ^ 2]. After further simplication, we obtain: 1/ [2^(2n) * n!], which approaches 0 as n -> oo. Therefore, the answer is 0.

  • @AndDiracisHisProphet
    @AndDiracisHisProphet 6 років тому +84

    now n!!!

    • @blackpenredpen
      @blackpenredpen  6 років тому +30

      AndDiracisHisProphet N!!!!!!!!
      :)

    • @AndDiracisHisProphet
      @AndDiracisHisProphet 6 років тому +7

      That would have been my nect suggestion
      n!^m where !^m is m factorials

    • @MagicGonads
      @MagicGonads 6 років тому +15

      It's called the "mth" step falling factorial and it has a general expression in gamma even for real-step :)

    • @AndDiracisHisProphet
      @AndDiracisHisProphet 6 років тому +2

      cool

    • @dlevi67
      @dlevi67 6 років тому +2

      Shadows of fractional operators á la Dr. Peyam!

  • @akshat9282
    @akshat9282 6 років тому +40

    The given product can be written as (using a bit of algebra)
    [(n!)/(((2^(n/2))((n/2)!))^2)]
    Since this was inf/inf, I used L'H with the gamma expansion of the factorial and after applying limits, you'll get zero. So yea the product is 0
    Edit: I'm not sure if bprp intended on using the Ramanujan or Stirling expansion because honestly, I have no idea what it is(or maybe I'm just missing out on something here). I saw the pinned comment and well I was a bit late but at least my solution isn't as complicated xD

    • @VibratorDefibrilator
      @VibratorDefibrilator 6 років тому +1

      The limit is equal to 0 indeed, but for simpler reason: in the fraction the numerator, n!/((n/2)!(n/2)!) is in fact the biggest binomial coefficient C(n;n/2) and the divisor is 2^n. But we know that the sum of all binomial coefficients at the n-th line of the Pascal's triangle is exactly 2^n. Thus, our fraction is less than 1 (obviously), but why the limit=0? There is connection with harmonic series 1/2 + 1/3 + 1/4 + 1/5 + ... + 1/n + ... which diverge with n -> inf.

    • @akshat9282
      @akshat9282 6 років тому

      VibratorDefibrilator oh yes I get your point. I couldn't really think about that.
      ++Love your username, man

    • @VibratorDefibrilator
      @VibratorDefibrilator 6 років тому +2

      Thank you for the swift reply, Akshat. Nevertheless, I think that this problem needs a bit more detailed explanation. Once we've spotted that binomial coefficient C(n;n/2), we might conjecture that its part of the sum 2^n will decrease constantly until it reach 0 with n->inf. If we compare two consecutive values: C(n;n/2) and C(n+2;(n+2)/2) indeed we'll have C(n+2;(n+2)/2) / C(n;n/2) = (2n+2)*(2n+1) / (n+1)^2 = 4*(1 - 1/(n+1)) when at the same time the corresponding sum will be quadrupled - 2^(n+2) / 2^n = 2^2 = 4. So, the temp of decreasing of that binomial coefficient will be (1 - 1/(n+1)). But if we multiply all these, we'll have our infinite product 1/2*3/4*5/6*7/8*... It turns out that our argument will be circular and finally we have to try the approach with harmonic series. At least now we know how to do it:
      We are taking reciprocal fraction of every term of the infinite product: 1/2 -> 2/1; 3/4 ->4/3; 5/6 ->6/5;...
      1/2*3/4*5/6*7/8*... = ((2/1)*(4/3)*(6/5)*(8/7)*...)^(-1) = ((1+1/1)*(1+1/3)*(1+1/5)*(1+1/7)*...)^(-1)
      Now, this infinite product contains (after uncovering the brackets) a sum 1/1 + 1/3 + 1/5 + 1/7 + 1/9 + 1/11 + ... and we can easily prove that it diverges. It is part of the harmonic series 1+1/2+1/3+1/4+1/5+1/6+1/7+1/8+...
      If we pair every two consecutive terms of that sum we'll get:
      1/3 + 1/5 = 8/15 > 1/2
      1/7 + 1/9 = 16/63 > 1/4
      1/11 + 1/13 = 24/143 > 1/6
      1/(2*k-1) + 1/(2*k+1) = 4*k/(4*k*k - 1) > 4*k/(4*k*k) =1/k, k is even
      1/3 + 1/5 + 1/7 + 1/9 + 1/11 + ... > 1/2 + 1/4 + 1/6 + 1/8 + ... = (1/2)*(1 + 1/2 + 1/3 + 1/4 + 1/5 + 1/6 + ...)
      Thus, we have P = 1/2*3/4*5/6*7/8*... = 1/(1+(1/2)*(1 + 1/2 + 1/3 + 1/4 + 1/5 + 1/6 + ...))=0
      Frankly, I don't see what the double factorial has to do with all this proof, but, on the other hand, it is good exercise.
      P.S.: My username came up after a hilarious conversation with a fellow female doctor and somehow sticked with me.

    • @akshat9282
      @akshat9282 6 років тому

      VibratorDefibrilator tell me about that conversation sometime haha

    • @VibratorDefibrilator
      @VibratorDefibrilator 6 років тому +1

      Oh, the exact situation is already in the mist of the past, but one have to know and remember that the doctors are one of the most cynical and dirty-minded people in the world, no exceptions, I think. Slowly, during the years, I came to the conclusion that they have to be... fitted... that way in order to preserve their sanity and to do their noble job. After all, they are close to all human beings from their birth to their deathbed, in illness and health, so nothing sacred and unknown remains there... It is enormous benefit for anybody who is fortuned enough to befriend even one of those noblest of cynics out there.
      As you may already noticed, I'm inclined to be painfully meticulous when I have to explain something. Don't bother even to answer. Wish you good luck.

  • @Bodyknock
    @Bodyknock 6 років тому +2

    I’d be interested in seeing an analytic expansion of the double factorial using something similar to the pi or gamma function definition of factorial which maintains the same recursive relationships on the even and odd integers for the double factorial definition in the video. In other words give a single smooth continuously differentiable function that maintains the same recursive relationships that the positive integers hold.

  • @danielbickford3458
    @danielbickford3458 2 роки тому

    The best description of multiple factorials that I've seen is "the louder you yell at the number the more scared it gets and the smaller it shrinks in on itself"

  • @nicholasleclerc1583
    @nicholasleclerc1583 6 років тому +1

    14:03
    Well, I’m not gonna use factorials, but I am going to use the capital Pi notation, like Signa, but for multiplication, I’m gonna name it CapPi, and so your expression is: lim{x-> +Inf} CapPi{n=1; n=x}((2n+1)/(2n+2)).
    This can be simplified into; lim{x-> +Inf} CapPi{n=1; n=x}(-1/(2n+2)), which means that you multiply more and more with smaller and smaller fractions, which means it will reach zero.

  • @dassapaniranjan6039
    @dassapaniranjan6039 Місяць тому

    Thank you so much for the clear explanation of double factorial!

  • @Courage2976
    @Courage2976 6 років тому +14

    can you do a video on (1/2)!
    You can put it in a calculator and get an answer, but how?

    • @andrewli4484
      @andrewli4484 6 років тому

      KeroseneAndFire The Gamma Function.

    • @paololeaer7838
      @paololeaer7838 6 років тому

      He's already done it

    • @Courage2976
      @Courage2976 6 років тому

      Oh, ha, never saw that video. Thanks!

    • @blackpenredpen
      @blackpenredpen  6 років тому +3

      Wait, I thought I pasted the link for you for it after I hearted the comment. But the link never showed up...
      Did u find it? If not, I will go find it again.

    • @Courage2976
      @Courage2976 6 років тому +2

      blackpenredpen Haha, yeah I found it. I didn’t see the video when it came out I guess. Thanks!

  • @adamp9553
    @adamp9553 4 роки тому +1

    For the final question, an infinite product of n!!/((n+1)!!) approaches zero because the denominator is generally larger, and multiplication makes the resulting fraction so much smaller.

  • @CTJ2619
    @CTJ2619 6 років тому +13

    how about !!! for . 6!!!= . 6*3=18

    • @blackpenredpen
      @blackpenredpen  6 років тому +1

      Carlton Johnson yup

    • @arcioko2142
      @arcioko2142 3 роки тому

      no dont just scrap this notation and use !! for actual iterated factorial

  • @Mario_Altare
    @Mario_Altare 6 років тому +3

    Encore! Encore :-D
    Your videos made me fall in love again with maths

  • @BTD22
    @BTD22 6 років тому +23

    It tends to 0 [1/(2^n)].
    Thanks for made a great channel like that, I really like it

    • @blackpenredpen
      @blackpenredpen  6 років тому +2

      BTD thanks!!

    • @BTD22
      @BTD22 6 років тому

      Wow, this is great! Thanks to you :^)

    • @chrissekely
      @chrissekely 6 років тому +2

      Not necessarily. If you ignore the one and compare term by term:
      3>2
      5>4
      7>6
      And so on. So, seeing it this way, the numerator would be larger and it would not make sense that it should tend to zero.

    • @BTD22
      @BTD22 6 років тому +1

      Chris Sekely I Will try to explain it...
      You can write the denominator like
      2*(2*2)*(2*3)*(2*4)*(2*5)....
      Eventaully all the impar (2K-1) numbers will be simplificated with all the numbers in the numerator.
      Then the denominator will be conformed just for par number (2K) which can be written like 2*2*2*2*2*2.... And that is why the denominator will be (2^n).
      [1/(2^n)].
      I hope that I can explain it well...
      Tell me if you have any question

    • @JohnHaase
      @JohnHaase 6 років тому +1

      You can't rearrange terms in an infinite product arbitrarily. The product does go to 0 but it actually goes as 1/sqrt(k).
      Using the relations presented in the video
      (2k-1)!!/(2k)!! = (2k)!/(2^(2k)*(k!)^2)
      Then using asymptotic Stirling's approximation on this expression
      sqrt(4*pi*k)*(2k/e)^(2k)
      -----------------------------------------------------------
      2^(2k) * ( sqrt(2*pi*k) * (k/e)^k )^2
      Which simplifies to
      1/sqrt(pi*k)
      so it goes to 0 as 1/sqrt(k)
      Using the bounded form of Stirling's formula (en.wikipedia.org/wiki/Stirling%27s_approximation) this holds
      2*sqrt(pi)/e^2 * 1/sqrt(k)

  • @tanelkagan
    @tanelkagan 2 роки тому

    Arthur Schuster also called this an "alternate factorial" which I think is a much better name, because "double factorial" implies you are doing something twice. If I had to come up with a name, I'd call it a "semi-factorial". But hey, that's just me.

  • @josephalacchi9857
    @josephalacchi9857 4 роки тому

    The infinite product can be simplified to lim n->inf of (2n)!/(2^2n * (n!)^2) which is simply the probability of X=n for X~Bin(2n,1/2) which approaches X~Norm(n,n/2) by the Central Limit Theorem implying that P(X=n) approaches 0.

  • @geryz7549
    @geryz7549 3 роки тому +1

    The more factorials, the louder you scream the number

  • @padmasoman9346
    @padmasoman9346 6 років тому

    I think that the answer is 1. The reason is that as n approaches infinity, n/(n+1) approaches 1 and thus by Cauchy's Theorem of the limit of product of terms in a series the answer is 1. I have however worked out the solution by expressing the product as
    (2k)!/(((2k)!!)squared) which can be simplified to (2k)!/(((2 power k).k!)squared) and then further expanding and simplifying as k approaches infinity.

  • @RAJSINGH-of9iy
    @RAJSINGH-of9iy 6 років тому

    We don't have to calculate all these or convert it to a limit, just by intution ( numerator )< (denominator) => 0< ratio < 1 but ratio -> 0 as elements goes to oo in both num. and den..

  • @AdityaGhosh50
    @AdityaGhosh50 6 років тому +2

    Sir, I have a doubt about definite integrals. My question is if I consider a piece wise defined function f(x) which give 0 if x is Rational, f(x) = x if x is irrational. So for some boundaries, is the definite integral of f(x) dx = definite integral of x ? Or not?

    • @francissirizzotti7283
      @francissirizzotti7283 6 років тому

      I don't know - is it always possible to find a rational number between any pair of irrational numbers? I suspect that it is. Even if your range is infinitessimally small, you can stack your two bounds on top of each other and compare them a decimal at a time. Because an irrational number's decimal expansion is infinite, close numbers must start with digits that match exactly, until a point of first difference. That first difference gives us an opportunity to construct a rational number - use the same string of starting digits that matches your bounds, then at the difference, take the digit from the upper end of the bound, then terminate your constructed number. You've constructed a rational number which must be larger than the lower bound, but must also be smaller than the lower bound. In fact, you can now create an infinite number of rational numbers between these two irrationals by using your rational number and adding digits. As long as the digits added remain less than or equal to the corresponding digits in the irrational upper bound. Without continuity - especially the discontinuity that results from an infinite count of missing places - the proof of equivalence of those two integrals would probably prove that the indefinite integrals are equal as well. Which... I'm just not sure how to even think about.

    • @AdityaGhosh50
      @AdityaGhosh50 6 років тому +1

      Francis Sirizzotti Even then. Countability maybe an important factor here. Since it is an integral, we can actually separate the rational points in a particular order, which when multiplied with dx gives 0. The same cannot be said about an uncountable number of irrationals, because they cannot be separated per se.

    • @francissirizzotti7283
      @francissirizzotti7283 6 років тому

      Yeah, I definitely agree this is a possibility. And when you couch it in terms of "does there exist some boundary where integral of f(x) = integral of x", it definitely seems plausible. But my reasoning was more about proving that there are a countably infinite number of rational numbers between _any_ pair of irrational numbers, and since there are also uncountably many irrational numbers in that same interval, and reasoning from differences in infinity would apply to the the space of all real numbers - so your question might be the same as "is Integral f(x) = 1 + C".

    • @gcewing
      @gcewing 6 років тому

      I think Riemann would say the integral is undefined, and Lebesgue would say that removing the rationals makes no difference to the integral, since they have measure zero. See Dr. Peyam's videos on Lebesgue integration.

    • @JohnDixon
      @JohnDixon 6 років тому

      Greg Ewing is right. From a "Riemann" point of view, you can't integrate the function since it isn't continuous at ANY point x. However, since the number of discontinuities is equal to the cardinality of the set of rationals, it is countable and therefore has measure zero; so, from a "Lebesgue" point of view you could technically remove all of the discontinuities and end up with a function identical to f(x) = x, which you could then integrate over a definite interval.

  • @frostcrackle2374
    @frostcrackle2374 6 років тому

    problem at the end can be rewritten as prod(n=1 to k of (2n-1)/2n), which, as n tends to infinity, the product tends towards 0.

  • @tanveerbajwa2682
    @tanveerbajwa2682 4 роки тому +1

    At 9:31 my mind was like😍

  • @ospreytalon8318
    @ospreytalon8318 6 років тому

    The given problem can be very coolly written as the limit (as k tends to infinity) of 2^(-2k) * (2k choose k). It's very clear this is 0, as the LHS term dominates.

  • @Sid-ix5qr
    @Sid-ix5qr 6 років тому +22

    I have made an observation.
    Can anyone figure out the differences between 12:40 and 12:48?

  • @turtlellamacow
    @turtlellamacow 6 років тому +1

    Is it....zero? I rewrote it as e^ln(1*3* .... / 2*4* ....) = e^( ln1 - ln2 + ln3 - ln4 + ...). That series diverges and approaches negative infinity, so the whole thing goes to zero

  • @qarsiseer
    @qarsiseer 6 років тому

    Importantly? n!!! = 3^k * k! also given that n is divisible by 3. It works for any power? of factorial actually: n !^p = p^k * k! given n=pk where k is a whole number. It even works for 1!

  • @Mrwiseguy101690
    @Mrwiseguy101690 6 років тому +6

    I know this doesn't have anything to do with double factorials, but I have a question about that limit at the end. If we were to write the fraction as the product of corresponding terms (i.e. 1/2 * 3/4 * 5/6 * 7/8...), then we can conclude that the limit must be less than 1/2. But if we dismiss the 1 in the numerator because it doesn't change the overall product, and we now chunk the fractions as 3/2 * 5/4 * 7/6..., then we can conclude that the limit must be greater than 3/2. But this is a contradiction. What's the mistake in my reasoning?

    • @user-vm6qx2tu3j
      @user-vm6qx2tu3j 6 років тому +1

      Sponge . That's what I concluded

    • @MagicGonads
      @MagicGonads 6 років тому +9

      You can't reorder infinite series. The correct way to remove the 1 would be to extract the 1/2 otherwise you are effectively multiplying the entire series by 'n' (which is being pushed to infinity)

    • @Mrwiseguy101690
      @Mrwiseguy101690 6 років тому +1

      Ah that makes sense. Thanks for the explanation!

    • @user-vm6qx2tu3j
      @user-vm6qx2tu3j 6 років тому

      Magic Gonads yes

    • @RAJSINGH-of9iy
      @RAJSINGH-of9iy 6 років тому

      Yeah magic gondas is correct, u just can't reorder any terms in multiplication in case of infinite series.Ex:- (1*2*3*.......)/10² = oo but if we chunk into fraction like (1/10²) *(2*3*5*...) < 1 and another form if we take 10^10000 instead of 10² we have (1/10^10000) * (2*3*5....)= 0*oo indeterminate form, hope u get it.. 😊

  • @junipalwitt6970
    @junipalwitt6970 6 років тому

    I could be very wrong here, since I haven't been in math classes in ages, and follow this channel due to my interest in math, but I did it as such. Rewrote it as a limit as n approaches infinity of ((2n-1)!/2n!), which then reduces to limit as n approaches infinity of 1/2n, which seems to approach zero.

  • @JortvilmGritic
    @JortvilmGritic Рік тому

    solution: (1 * 3 * 5 ...) / (2 * 4 * 6 ...) can be created with big PI notation: \prod_{n=1}^{x}\frac{2n-1}{2n}; the solution goes to 0. (1 * 3 * 5 ...) / (2 * 4 * 6 ...) = 0.

  • @helpteacher5640
    @helpteacher5640 6 років тому

    Thanks sir world's best teacher

  • @manojkumarjena7492
    @manojkumarjena7492 3 роки тому

    Expression so good 🙏🙏🙏

  • @gutterball10
    @gutterball10 6 років тому +1

    never heard of such a function!! thanks for the knowledge.

  • @MrCigarro50
    @MrCigarro50 6 років тому

    really elegant notation. Thank you very much.

    • @arcioko2142
      @arcioko2142 3 роки тому

      the second person to say that with the first one being whoever decided to notate as that

  • @L1N3R1D3R
    @L1N3R1D3R 6 років тому

    Question: When you worked out the double factorial for odd numbers, why even bother multiplying by 6 on top and bottom? Leaving it out would make the generalized form in terms of n somewhat nicer, as the (n+1)! on the top would simplify to just n!, and the two (n+1)/2's on the bottom would flip to (n-1)/2's.

  • @justabunga1
    @justabunga1 5 років тому

    @blackpenredpen What about double factorials for something that is a non-integer, 0, and negative numbers? Also, why can’t double factorials be inputted on a calculator?

  • @namangupta1010
    @namangupta1010 6 років тому

    1.) Lim [100(tan^-1(x)/x)]
    (x-->0)
    where [.] represents the greatest integer function
    2.) lim [100(tan(x)/x)]
    (x-->0)
    where [.] represents the greatest integer function

  • @netoskin
    @netoskin 6 років тому

    Everyone loves the good ol' definition of factorial

  • @kotarojujo6365
    @kotarojujo6365 4 роки тому +1

    Idk man but I think the answer for the question at the end is 4/2^n . n!
    Where n is equal to the denominator

  • @Risu0chan
    @Risu0chan 6 років тому

    Noticing each pair of factors is less than 1, the sequence is strictly decreasing, and since it is bounded by 0, we know there is convergence. Now, we can write both the numerator and the denominator as double factorials, (2k-1)!! / (2k)!!, applying the formulas in this video we get a formula with usual factorials, and applying the Stirling equivalence for large numbers, it naturally yields to 1/sqrt(pi*n). The sequence converges to zero.

  • @lythd
    @lythd 5 років тому

    For odd you can just cancel the 6s to get 5!/2^2*2!. In simple you can say n=2k+1 and then n!!=n!/(2^k*k!) or n!!=n!/k!! to be even simpler.

  • @francissirizzotti7283
    @francissirizzotti7283 6 років тому +2

    Every factor in the top can be canceled with its double in the bottom leaving you with 1/(2*4*2*8*2*12*2*16...) which you can multiply pairwise to rewrite as 1/(8*16*24*32...). Using the same trick in the video, this becomes lim x->inf of 1/(8^x * x!), which is pretty trivially, zero.
    ... I think? The pairwise multiplication/canceling on an infinite series seems pretty fishy to me, but I'm not really sure why I should disallow it.

  • @stumbling
    @stumbling 6 років тому

    I saw this solution instantly! Your videos must be rubbing off.
    n!! = n! / (n-1)! * (n-2)! / (n-3)! * (n-4)! ... 1!
    Dividing or multiplying by 1! gives the same answer so it doesn't matter if n was odd or even.
    Not the shortest but I like how it looks. :P

  • @doondoon859
    @doondoon859 6 років тому +6

    For dealing with n!! when n is odd, I prefer to just observe that n! = n!! * (n-1)!! and substitute the formula of n!! with n even.

    • @fCauneau
      @fCauneau 6 років тому

      same I did !!!

  • @DiegoGonzalez-xl9us
    @DiegoGonzalez-xl9us 6 років тому

    Can you explain me why you write the def. of an odd number as: (2k-1) instead of (2k+1)? I mean, what if (k=0)? then it is going to be -1 and you cannot make negative number factorials right? Sorry if I am asking something stupid, but I need to know it.

  • @karakrisundkramm
    @karakrisundkramm 6 років тому

    The product is divergent. You could write it as an infinite product of fractions that are all bigger or equal to one (1 * 3/2 * 5/4 * ...). At the same time you could write it as an infinite product of fractions, that are all smaller than one (1/2 * 3/4 * 5/6 * ...).

    • @MarceloSilva-lh9mh
      @MarceloSilva-lh9mh 5 років тому

      That's why you're not allowed to rearrange the terms.

  • @AlcuBerry
    @AlcuBerry 6 років тому

    Hey, check this out
    After calculating the even case (2k!!), you could just have noticed that
    (2k-1)!!*(2k)!!=(2k)!
    because if you alternated odd numbers with 2k-1and then even numbers with 2k you covered all of them, and then you can divide by 2k!!
    (2k-1)!!=(2k!)/(2k!!) and everything on the right hand side is known.

  • @some1here0
    @some1here0 4 роки тому

    thank you . this video is very helpful

  • @UshaRani-st3yf
    @UshaRani-st3yf 2 роки тому

    Thank you Sir. It was good.
    Now if you can please explain !n and n !!!

  • @YaBoyUneven
    @YaBoyUneven 3 роки тому +1

    I am just a middleschooler, please show mercy

  • @yuvalpaz3752
    @yuvalpaz3752 6 років тому

    We have: n!!=n!/(n-1)!!. setting a_n=n!! we get the recursive relation: a_n=n!/a_(n-1) with a_1=1, putting this into walframAlpha we get:
    a_n = exp((-1)^(n + 1) ( sum_(k=-1)^(n - 1) (-1)^(-k) log((k + 1)!)))=n!!

  • @mastershooter64
    @mastershooter64 4 роки тому +1

    what about fractional factorials?

  • @ConociendoAJesucristo
    @ConociendoAJesucristo Рік тому

    what is the relation recurrence for 0!! and (-1)!! teacher?

  • @PeriOfTheGee
    @PeriOfTheGee 6 років тому

    Also, for odd n, n!! = n!/([n-1]!!)

  • @howardlam6181
    @howardlam6181 3 роки тому

    it arises from integral of x^(2n) exp(-x^2/2) over the real line

  • @Lucaazade
    @Lucaazade 6 років тому +2

    Ps 720! = 2.6 ×10^1746. (The calculator on my phone goes higher than 49000! = 8.4 ×10^208541)

    • @blackpenredpen
      @blackpenredpen  6 років тому +1

      Wow!! What phone do U have?

    • @Lucaazade
      @Lucaazade 6 років тому +1

      Nexus / Android :)

    • @badhbhchadh
      @badhbhchadh 5 років тому

      @@blackpenredpen Wolframalpha server

  • @proto9053
    @proto9053 6 років тому

    We have(2n-1)!!/(2n)!! = (2n)!/(4^n * n!n!) = (2n choose n)/4^n. Now, taking the limit, we have 1/sqrt(πn) -> 0.

  • @xvdvn2216
    @xvdvn2216 6 років тому

    I find it crazy that double factorials are actually LESS insane than normal factorials.

  • @kevincaotong
    @kevincaotong 6 років тому

    We write the product as limit as k approaches infinity of (2k-1)!!/(2k)!!, and plugging in the formulas, we get (((2k)!)/2^k(k)!)/2^k(k)!=((2k)!/(2^{2k}(k!)^2), and applying Stirling's Approximation yields 0. :)

  • @vinceperry769
    @vinceperry769 6 років тому

    I would like to see indefinite integration, integration carried out an indefinite number of times, just to see the symbolism not with ellipses (the language kind not the conic section) therefore, or indefinite composition, again indefinite number of layers of composition, for the compact symbolism. I know the derivative of the indefinite composition would use the TT, product, operator, product of the derivative of the outermost layer, successive factors being derivatives of one more layer in until the innermost layer.

  • @Tletna
    @Tletna 2 роки тому

    If double factorial is factorial every other number, then what is half a factorial? Going down by 1/2 each time? Then, what is a factorial'ed factorial? For example: when n is an integer, what about n(!-n times) more specific example.. so for n = 4, what is n!!!!, is it just n? So, then 4? So then what about n = 0? Just 0?

  • @alberto3071
    @alberto3071 6 років тому +2

    Is there a double factorial function?

  • @meanyoun9691
    @meanyoun9691 6 років тому +1

    Hi i am from Cambodia 👋👋
    I like u blackpenredpen😍😚

    • @AlgyCuber
      @AlgyCuber 6 років тому

      i’m from the country next to you, thailand

  • @sonaroy9493
    @sonaroy9493 4 роки тому

    Also 720 factorial is approximately equal to 2.601*10^1746

  • @pseudo_goose
    @pseudo_goose 5 років тому +1

    Factorio?

  • @alinamalik9687
    @alinamalik9687 Рік тому

    is the aswer to the question at the end "(2k)!"?

  • @قرآنكريم-ش8ز3ص
    @قرآنكريم-ش8ز3ص 3 роки тому

    Thank's very much
    From Palestine 🌍♥️🇵🇸

  • @zeeek3348
    @zeeek3348 6 років тому

    Just for fun, exist something like gamma function for double factOreo?

    • @blackpenredpen
      @blackpenredpen  6 років тому

      Oscar Aguilar
      You will see in my coming video! :)

  • @jaimeduncan6167
    @jaimeduncan6167 6 років тому

    I went down for the odd version, so a little bit uglier but same result. Great explanation.

  • @takyc7883
    @takyc7883 4 роки тому

    So 6!!! Would equal 6x3x1=18? Also how would a half factorial work?

  • @tourneriealexandre7669
    @tourneriealexandre7669 6 років тому

    ln(x)=1/x ??

  • @kor0p
    @kor0p 6 років тому

    This will be limit ,when x go to inf, of ((x+1)!/((4^x)*(x!)^2)) and this is the same as ((x+1)/((x)!*4^(x))) that must go to zero.
    My answer - 0
    Thanks for your channel!

  • @Ny0s
    @Ny0s 4 роки тому

    9:50. Oh, you can pretend nothing happened, but we know!

  • @mtmath1123
    @mtmath1123 5 років тому

    Don’t know why this clip popped up among your great videos, but I know the music played at the end ;-)

  • @crosisbh1451
    @crosisbh1451 6 років тому

    I believe the problem would equal 0. I did some algebra and got the fraction (2k)!/2^(2k)(k!)^2. I wrote out (2k)! and factored out the 2k. It turned out to be 2k(1*(1-1)*(1-2)...). 1-1 = 0, so the whole expression must equal 0, Since the numerator is 0, the whole number is 0. (Hope my logic was right here)

  • @cameodamaneo
    @cameodamaneo 6 років тому +1

    The final answer is undefined, as it's the quotient of two divergent infinite series.

    • @oximas
      @oximas 4 роки тому

      nope it's zero

  • @Mayank-mf7xr
    @Mayank-mf7xr 6 років тому

    the title is a double factorial itself .
    reeeeeeeeeeeeeeeeeeeeeeeeeee

  • @Nerdwithoutglasses
    @Nerdwithoutglasses 4 роки тому

    Well, this is a brilliant move!

  • @tomasmach2272
    @tomasmach2272 Рік тому

    Now I wonder if there is an analytic continuation for this one😅

  • @binatrabinovich4902
    @binatrabinovich4902 3 роки тому

    Plz
    How to do this 2^(n+1) factorial ?

  • @muktagain5763
    @muktagain5763 6 років тому

    Very good job....

  • @its_robbietime1333
    @its_robbietime1333 5 років тому

    Is there something similar to the factorial function except it’s something like 6+5+4+3+2+1

    • @adeelsumar3721
      @adeelsumar3721 5 років тому +1

      It’s_Robbie Time well sort of. the series of adding consecutive natural numbers happens to be the triangular numbers. the formula for the nth triangular number is n(n+1)/2

    • @its_robbietime1333
      @its_robbietime1333 5 років тому

      @@adeelsumar3721 ahh okay

  • @abrarshaikh2254
    @abrarshaikh2254 6 років тому

    How will express double factorial in gamma function?

    • @MagicGonads
      @MagicGonads 6 років тому

      substitute x! = gamma(x+1)

    • @MagicGonads
      @MagicGonads 6 років тому

      www.wolframalpha.com/input/?i=(2%5E(x%2F2)+%CE%93(1+%2B+x%2F2))%2F((pi%2F2)%5E(sin(pi%2F2*x)%2F2))

    • @MagicGonads
      @MagicGonads 6 років тому

      The denominator is just there to make sure that the sqrt(pi/2) doesn't occur on the odd numbers due to using the gamma function rather than a product operator

  • @BUSHhoehh
    @BUSHhoehh 6 років тому

    Okay that's cool and all but have you thought about a half-factorial? Be interesting to see and think about that

    • @BUSHhoehh
      @BUSHhoehh 6 років тому

      Like peyam did with a half derivative, except a half factorial

    • @arcioko2142
      @arcioko2142 3 роки тому

      double factorial should be called half factorial and be notated as something else because i hate the notation with every single fucking part of my heart

  • @sophiegrey9576
    @sophiegrey9576 6 років тому

    I thought that was just a really good chess move.

  • @Lifefinder15
    @Lifefinder15 4 роки тому

    well done sir from Pakistan

  • @GodbornNoven
    @GodbornNoven 2 роки тому

    n!!=n(n-y)(n-2y)(n-3y)...
    where y = 2.
    6!!=6(6-2)(6-4)
    xy=z'(the one that ends the series)
    it ends at n-z'>0
    if odd, then n-z'=1
    if even, then n-z'=2
    two examples for reference.
    6 and 5
    6!!=6(6-2)(6-4)=6x4x2
    6-4>0
    since 6 is even then the series ends with a two
    5!!=5(5-2)(5-4)
    odd series so it ends with 1.

  • @lorenzoiannuzzi3937
    @lorenzoiannuzzi3937 6 років тому

    is it 0?

  • @TheMauror22
    @TheMauror22 6 років тому

    So if we had instead a triple factorial, we would just decrease the numbers by 3 and so on?

  • @kapildeo8221
    @kapildeo8221 6 років тому

    I think lim n tends to infinity (2n-1)/2n=1 may be right answer....

  • @thatonepersonyouknowtheone7781
    @thatonepersonyouknowtheone7781 2 роки тому

    the louder you shout, the quieter you're heard