How do I prove this combinatorial identity Unicorn Meta Zoo #1: Why another podcast? Announcing the arrival of Valued Associate #679: Cesar ManaraAlternative combinatorial proof for $sumlimits_r=0^nbinomnrbinomm+rn=sumlimits_r=0^nbinomnrbinommr2^r$Number of way to invite my friends over for dinner on 5 nightsHelp with how to prepare the inductive step of a strong induction exercise.Source and/or combinatorial interpretation for $F_n+k = sum_i=0^k binomkiF_n-i$Another combinatorial identity of McKayExpectation of the fraction a random function covers its rangeFind a conjecture for $F_1+F_2+…+F_n$Proof involving Fibonacci number and binomial coefficientCombinatorial proof of $n choose 1 + n choose 3 +cdots = n choose 0 + n choose 2+cdots$Generating function of even Fibonacci numbers

How exactly does Hawking radiation decrease the mass of black holes?

How do I prove this combinatorial identity

Philosophical question on logistic regression: why isn't the optimal threshold value trained?

A faster way to compute the largest prime factor

What does **function** actually mean in music?

Conditionally enable edit in lightning:datatable

What to do with someone that cheated their way through university and a PhD program?

What's the difference between using dependency injection with a container and using a service locator?

Why did C use the -> operator instead of reusing the . operator?

Will I lose my paid in full property

A strange hotel

How do I check if a string is entirely made of the same substring?

Air bladders in bat-like skin wings for better lift?

What was Apollo 13's "Little Jolt" after MECO?

How long after the last departure shall the airport stay open for an emergency return?

What does a straight horizontal line above a few notes, after a changed tempo mean?

How to keep bees out of canned beverages?

Why didn't the Space Shuttle bounce back into space as many times as possible so as to lose a lot of kinetic energy up there?

Is Diceware more secure than a long passphrase?

Could moose/elk survive in the Amazon forest?

Which big number is bigger?

Scheduling based problem

What makes accurate emulation of old systems a difficult task?

Obeylines and gappto from etoolbox



How do I prove this combinatorial identity



Unicorn Meta Zoo #1: Why another podcast?
Announcing the arrival of Valued Associate #679: Cesar ManaraAlternative combinatorial proof for $sumlimits_r=0^nbinomnrbinomm+rn=sumlimits_r=0^nbinomnrbinommr2^r$Number of way to invite my friends over for dinner on 5 nightsHelp with how to prepare the inductive step of a strong induction exercise.Source and/or combinatorial interpretation for $F_n+k = sum_i=0^k binomkiF_n-i$Another combinatorial identity of McKayExpectation of the fraction a random function covers its rangeFind a conjecture for $F_1+F_2+…+F_n$Proof involving Fibonacci number and binomial coefficientCombinatorial proof of $n choose 1 + n choose 3 +cdots = n choose 0 + n choose 2+cdots$Generating function of even Fibonacci numbers










12












$begingroup$


Show that $2n choose n + 32n-1 choose n + 3^22n-2 choose n + cdots + 3^nn choose n \ = 2n+1 choose n+1 + 22n+1 choose n+2 + 2^22n+1 choose n+3 + cdots + 2^n2n+1 choose 2n+1$



One way that I did it was to use the idea of generating functions.
For the left hand side expression, I can find 2 functions. Consider;



$$f_1 (x) = frac1(1-3x) \ = 1 + 3^1x + 3^2x^2 + 3^3x^3 + cdots + 3^nx^n + cdots \ f_2(x) = frac1(1-x)^n+1 \ = n choose n + n+1 choose nx + n+2 choose nx^2 + cdots + 2n-1 choose nx^n-1 + 2n choose nx^n + cdots + $$



Consider the coefficient of $x^n$ in the expansion of $f_1 (x) . f_2 (x)$. Then the coefficient will be the expression on the left hand side.



Now we further consider 2 functions for the right-hand side expression.



Consider;



$$f_3 (x) = frac 1(1-2x) \ = 1 + 2^1x + 2^2x^2 + cdots + 2^n-1x^n-1 + 2^nx^n + cdots \ f_4 (x) = (1+x)^2n+1 \= 1 + 2n+1 choose 1x + cdots + 2n+1 choose n-1x^n-1 + 2n+1 choose nx^n +cdots + 2n+1 choose 0x^2n +1 \ = 2n+1 choose 2n+1 + 2n+1 choose 2nx + 2n+1 choose 2n-1x^2 + cdots + 2n+1 choose n+2x^n-1 + 2n+1 choose n+1x^n + \
+ 2n+1 choose nx^n+1 +cdots + 2n+1 choose 0x^2n +1$$



Hence the coefficient of $x^n$ is the coefficient of $x^n$ in the expansion of $f_3(x) . f_4(x)$



This is what I managed to do so far. I'm not sure if $f_1(x) .f_2(x) = f_3(x).f_4(x)$. If the two functions are indeed equal, then I can conclude that their coefficient of $x^n$ must be equal, which will immediately answer the question. If they are equal, how do I show that they are?



If the two functions are not equal? How do I proceed to show this question?



Edit: It might not be true that the product of the two functions are equal. I tried substituting $x=0.1, n=1$. Seems like the two values are not equal. How do I proceed with this question?










share|cite|improve this question











$endgroup$











  • $begingroup$
    The two functions are not equal. In general for rational expressions, ie fractions where numerator and denominator are polynomials, if $a(x)/b(x)=c(x)/d(x)$ for all $x$ (ie expressions are identical), then you must have the polynomial equality $a(x)d(x)=b(x)c(x)$ which is only true if the two products are the same polynomial. If both fractions, $a(x)/b(x)$ and $c(x)/d(x)$, are without common factors, this is only true if $a(x)=kcdot c(x)$ and $b(x)=kcdot d(x)$ for some constant $k$.
    $endgroup$
    – Einar Rødland
    12 hours ago










  • $begingroup$
    Noted! Thanks for the explanation!
    $endgroup$
    – Icycarus
    12 hours ago















12












$begingroup$


Show that $2n choose n + 32n-1 choose n + 3^22n-2 choose n + cdots + 3^nn choose n \ = 2n+1 choose n+1 + 22n+1 choose n+2 + 2^22n+1 choose n+3 + cdots + 2^n2n+1 choose 2n+1$



One way that I did it was to use the idea of generating functions.
For the left hand side expression, I can find 2 functions. Consider;



$$f_1 (x) = frac1(1-3x) \ = 1 + 3^1x + 3^2x^2 + 3^3x^3 + cdots + 3^nx^n + cdots \ f_2(x) = frac1(1-x)^n+1 \ = n choose n + n+1 choose nx + n+2 choose nx^2 + cdots + 2n-1 choose nx^n-1 + 2n choose nx^n + cdots + $$



Consider the coefficient of $x^n$ in the expansion of $f_1 (x) . f_2 (x)$. Then the coefficient will be the expression on the left hand side.



Now we further consider 2 functions for the right-hand side expression.



Consider;



$$f_3 (x) = frac 1(1-2x) \ = 1 + 2^1x + 2^2x^2 + cdots + 2^n-1x^n-1 + 2^nx^n + cdots \ f_4 (x) = (1+x)^2n+1 \= 1 + 2n+1 choose 1x + cdots + 2n+1 choose n-1x^n-1 + 2n+1 choose nx^n +cdots + 2n+1 choose 0x^2n +1 \ = 2n+1 choose 2n+1 + 2n+1 choose 2nx + 2n+1 choose 2n-1x^2 + cdots + 2n+1 choose n+2x^n-1 + 2n+1 choose n+1x^n + \
+ 2n+1 choose nx^n+1 +cdots + 2n+1 choose 0x^2n +1$$



Hence the coefficient of $x^n$ is the coefficient of $x^n$ in the expansion of $f_3(x) . f_4(x)$



This is what I managed to do so far. I'm not sure if $f_1(x) .f_2(x) = f_3(x).f_4(x)$. If the two functions are indeed equal, then I can conclude that their coefficient of $x^n$ must be equal, which will immediately answer the question. If they are equal, how do I show that they are?



If the two functions are not equal? How do I proceed to show this question?



Edit: It might not be true that the product of the two functions are equal. I tried substituting $x=0.1, n=1$. Seems like the two values are not equal. How do I proceed with this question?










share|cite|improve this question











$endgroup$











  • $begingroup$
    The two functions are not equal. In general for rational expressions, ie fractions where numerator and denominator are polynomials, if $a(x)/b(x)=c(x)/d(x)$ for all $x$ (ie expressions are identical), then you must have the polynomial equality $a(x)d(x)=b(x)c(x)$ which is only true if the two products are the same polynomial. If both fractions, $a(x)/b(x)$ and $c(x)/d(x)$, are without common factors, this is only true if $a(x)=kcdot c(x)$ and $b(x)=kcdot d(x)$ for some constant $k$.
    $endgroup$
    – Einar Rødland
    12 hours ago










  • $begingroup$
    Noted! Thanks for the explanation!
    $endgroup$
    – Icycarus
    12 hours ago













12












12








12


5



$begingroup$


Show that $2n choose n + 32n-1 choose n + 3^22n-2 choose n + cdots + 3^nn choose n \ = 2n+1 choose n+1 + 22n+1 choose n+2 + 2^22n+1 choose n+3 + cdots + 2^n2n+1 choose 2n+1$



One way that I did it was to use the idea of generating functions.
For the left hand side expression, I can find 2 functions. Consider;



$$f_1 (x) = frac1(1-3x) \ = 1 + 3^1x + 3^2x^2 + 3^3x^3 + cdots + 3^nx^n + cdots \ f_2(x) = frac1(1-x)^n+1 \ = n choose n + n+1 choose nx + n+2 choose nx^2 + cdots + 2n-1 choose nx^n-1 + 2n choose nx^n + cdots + $$



Consider the coefficient of $x^n$ in the expansion of $f_1 (x) . f_2 (x)$. Then the coefficient will be the expression on the left hand side.



Now we further consider 2 functions for the right-hand side expression.



Consider;



$$f_3 (x) = frac 1(1-2x) \ = 1 + 2^1x + 2^2x^2 + cdots + 2^n-1x^n-1 + 2^nx^n + cdots \ f_4 (x) = (1+x)^2n+1 \= 1 + 2n+1 choose 1x + cdots + 2n+1 choose n-1x^n-1 + 2n+1 choose nx^n +cdots + 2n+1 choose 0x^2n +1 \ = 2n+1 choose 2n+1 + 2n+1 choose 2nx + 2n+1 choose 2n-1x^2 + cdots + 2n+1 choose n+2x^n-1 + 2n+1 choose n+1x^n + \
+ 2n+1 choose nx^n+1 +cdots + 2n+1 choose 0x^2n +1$$



Hence the coefficient of $x^n$ is the coefficient of $x^n$ in the expansion of $f_3(x) . f_4(x)$



This is what I managed to do so far. I'm not sure if $f_1(x) .f_2(x) = f_3(x).f_4(x)$. If the two functions are indeed equal, then I can conclude that their coefficient of $x^n$ must be equal, which will immediately answer the question. If they are equal, how do I show that they are?



If the two functions are not equal? How do I proceed to show this question?



Edit: It might not be true that the product of the two functions are equal. I tried substituting $x=0.1, n=1$. Seems like the two values are not equal. How do I proceed with this question?










share|cite|improve this question











$endgroup$




Show that $2n choose n + 32n-1 choose n + 3^22n-2 choose n + cdots + 3^nn choose n \ = 2n+1 choose n+1 + 22n+1 choose n+2 + 2^22n+1 choose n+3 + cdots + 2^n2n+1 choose 2n+1$



One way that I did it was to use the idea of generating functions.
For the left hand side expression, I can find 2 functions. Consider;



$$f_1 (x) = frac1(1-3x) \ = 1 + 3^1x + 3^2x^2 + 3^3x^3 + cdots + 3^nx^n + cdots \ f_2(x) = frac1(1-x)^n+1 \ = n choose n + n+1 choose nx + n+2 choose nx^2 + cdots + 2n-1 choose nx^n-1 + 2n choose nx^n + cdots + $$



Consider the coefficient of $x^n$ in the expansion of $f_1 (x) . f_2 (x)$. Then the coefficient will be the expression on the left hand side.



Now we further consider 2 functions for the right-hand side expression.



Consider;



$$f_3 (x) = frac 1(1-2x) \ = 1 + 2^1x + 2^2x^2 + cdots + 2^n-1x^n-1 + 2^nx^n + cdots \ f_4 (x) = (1+x)^2n+1 \= 1 + 2n+1 choose 1x + cdots + 2n+1 choose n-1x^n-1 + 2n+1 choose nx^n +cdots + 2n+1 choose 0x^2n +1 \ = 2n+1 choose 2n+1 + 2n+1 choose 2nx + 2n+1 choose 2n-1x^2 + cdots + 2n+1 choose n+2x^n-1 + 2n+1 choose n+1x^n + \
+ 2n+1 choose nx^n+1 +cdots + 2n+1 choose 0x^2n +1$$



Hence the coefficient of $x^n$ is the coefficient of $x^n$ in the expansion of $f_3(x) . f_4(x)$



This is what I managed to do so far. I'm not sure if $f_1(x) .f_2(x) = f_3(x).f_4(x)$. If the two functions are indeed equal, then I can conclude that their coefficient of $x^n$ must be equal, which will immediately answer the question. If they are equal, how do I show that they are?



If the two functions are not equal? How do I proceed to show this question?



Edit: It might not be true that the product of the two functions are equal. I tried substituting $x=0.1, n=1$. Seems like the two values are not equal. How do I proceed with this question?







combinatorics induction binomial-coefficients generating-functions combinatorial-proofs






share|cite|improve this question















share|cite|improve this question













share|cite|improve this question




share|cite|improve this question








edited 4 hours ago









Andreas

8,5111137




8,5111137










asked 13 hours ago









IcycarusIcycarus

5421315




5421315











  • $begingroup$
    The two functions are not equal. In general for rational expressions, ie fractions where numerator and denominator are polynomials, if $a(x)/b(x)=c(x)/d(x)$ for all $x$ (ie expressions are identical), then you must have the polynomial equality $a(x)d(x)=b(x)c(x)$ which is only true if the two products are the same polynomial. If both fractions, $a(x)/b(x)$ and $c(x)/d(x)$, are without common factors, this is only true if $a(x)=kcdot c(x)$ and $b(x)=kcdot d(x)$ for some constant $k$.
    $endgroup$
    – Einar Rødland
    12 hours ago










  • $begingroup$
    Noted! Thanks for the explanation!
    $endgroup$
    – Icycarus
    12 hours ago
















  • $begingroup$
    The two functions are not equal. In general for rational expressions, ie fractions where numerator and denominator are polynomials, if $a(x)/b(x)=c(x)/d(x)$ for all $x$ (ie expressions are identical), then you must have the polynomial equality $a(x)d(x)=b(x)c(x)$ which is only true if the two products are the same polynomial. If both fractions, $a(x)/b(x)$ and $c(x)/d(x)$, are without common factors, this is only true if $a(x)=kcdot c(x)$ and $b(x)=kcdot d(x)$ for some constant $k$.
    $endgroup$
    – Einar Rødland
    12 hours ago










  • $begingroup$
    Noted! Thanks for the explanation!
    $endgroup$
    – Icycarus
    12 hours ago















$begingroup$
The two functions are not equal. In general for rational expressions, ie fractions where numerator and denominator are polynomials, if $a(x)/b(x)=c(x)/d(x)$ for all $x$ (ie expressions are identical), then you must have the polynomial equality $a(x)d(x)=b(x)c(x)$ which is only true if the two products are the same polynomial. If both fractions, $a(x)/b(x)$ and $c(x)/d(x)$, are without common factors, this is only true if $a(x)=kcdot c(x)$ and $b(x)=kcdot d(x)$ for some constant $k$.
$endgroup$
– Einar Rødland
12 hours ago




$begingroup$
The two functions are not equal. In general for rational expressions, ie fractions where numerator and denominator are polynomials, if $a(x)/b(x)=c(x)/d(x)$ for all $x$ (ie expressions are identical), then you must have the polynomial equality $a(x)d(x)=b(x)c(x)$ which is only true if the two products are the same polynomial. If both fractions, $a(x)/b(x)$ and $c(x)/d(x)$, are without common factors, this is only true if $a(x)=kcdot c(x)$ and $b(x)=kcdot d(x)$ for some constant $k$.
$endgroup$
– Einar Rødland
12 hours ago












$begingroup$
Noted! Thanks for the explanation!
$endgroup$
– Icycarus
12 hours ago




$begingroup$
Noted! Thanks for the explanation!
$endgroup$
– Icycarus
12 hours ago










3 Answers
3






active

oldest

votes


















15












$begingroup$

Here is a combinatorial proof. Both sides of the equation answer the following question:




How many sequences are there of length $2n+1$, with entries in $0,1,2$, such that



  • at least one of the entries is a $2$, and

  • there are exactly $n$ zeroes to the left of the leftmost $2$?



LHS:



Suppose the leftmost $2$ occurs in spot $k+1$. Among the $k$ spots before hand, you must choose $n$ of the entries to be zero. The $2n+1-(k+1)=2n-k$ spots afterward can be anything. There are $binomkn3^2n-k$ ways to do this. Then sum over $k$.



RHS:



Suppose there are $j$ entries which are equal to $0$ or $2$. Choose those entries which are equal to $0$ or $2$ in $binom2n+1j$ ways. The leftmost $n$ of these entries must be zero, the $(n+1)^st$ entry must be two, then the remaining $j-(n+1)$ entries can be chosen freely among $0$ and $2$. There are $binom2n+1j2^j-(n+1)$ ways to do this, then sum over $j$.






share|cite|improve this answer









$endgroup$








  • 2




    $begingroup$
    How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
    $endgroup$
    – Icycarus
    12 hours ago






  • 5




    $begingroup$
    @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
    $endgroup$
    – Mike Earnest
    12 hours ago






  • 2




    $begingroup$
    @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
    $endgroup$
    – Mitch
    6 hours ago


















2












$begingroup$

Using your functions, consider
$$
3^n f_2(frac13) = 3^n frac1(1-frac13)^n+1 = frac32 (frac92)^n\ = n choose n3^n + n+1 choose n3^n-1 + cdots + 2n choose n + colorred 2n +1 choose n 3^-1+ cdots
$$

and further
$$
2^n f_4 (frac12) = 2^n (frac32)^2n+1 = frac32 (frac92)^n \= 2n+1 choose 2n+12^n + 2n+1 choose 2n2^n-1 + cdots + 2n+1 choose n+1 + colorred 2n +1 choose n 2^-1+ cdots + 2n +1 choose 0 2^-n-1
$$

The two expressions both equal $frac32 (frac92)^n$, and the first $n+1$ many terms represent the LHS and RHS of the original question. The terms in red are extra terms: once it is established that these terms also equal, the questions is solved. That is, show that
$$
sum_k=1^infty2n +k choose n 3^-k = sum_k=1^n+12n +1 choose n+1-k 2^-k
$$






share|cite|improve this answer











$endgroup$








  • 2




    $begingroup$
    I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
    $endgroup$
    – Mike Earnest
    12 hours ago










  • $begingroup$
    @MikeEarnest I resolved it and put it in the main text.
    $endgroup$
    – Andreas
    4 hours ago


















1












$begingroup$

We seek to show that



$$sum_q=0^n 2n-qchoose n 3^q
= sum_q=0^n 2n+1choose n+1+q 2^q.$$



We have for the LHS



$$sum_q=0^n 2n-qchoose n-q 3^q
= sum_q=0^n 3^q [z^n-q] (1+z)^2n-q
\ = [z^n] (1+z)^2n sum_q=0^n 3^q z^q (1+z)^-q.$$



The coefficient extractor controls the range and we obtain



$$[z^n] (1+z)^2n sum_qge 0 3^q z^q (1+z)^-q
= [z^n] (1+z)^2n frac11-3z/(1+z)
\ = [z^n] (1+z)^2n+1 frac11-2z.$$



We could conclude at this point by inspection. Continuing anyway we
get for the RHS



$$sum_q=0^n 2n+1choose n-q 2^q
= sum_q=0^n 2^q [z^n-q] (1+z)^2n+1
\ = [z^n] (1+z)^2n+1 sum_q=0^n 2^q z^q.$$



The coefficient extractor once more controls the range and we obtain



$$[z^n] (1+z)^2n+1 sum_qge 0 2^q z^q
= [z^n] (1+z)^2n+1 frac11-2z.$$



The two generating functions are the same and we have equality for LHS
and RHS.






share|cite|improve this answer









$endgroup$













    Your Answer








    StackExchange.ready(function()
    var channelOptions =
    tags: "".split(" "),
    id: "69"
    ;
    initTagRenderer("".split(" "), "".split(" "), channelOptions);

    StackExchange.using("externalEditor", function()
    // Have to fire editor after snippets, if snippets enabled
    if (StackExchange.settings.snippets.snippetsEnabled)
    StackExchange.using("snippets", function()
    createEditor();
    );

    else
    createEditor();

    );

    function createEditor()
    StackExchange.prepareEditor(
    heartbeatType: 'answer',
    autoActivateHeartbeat: false,
    convertImagesToLinks: true,
    noModals: true,
    showLowRepImageUploadWarning: true,
    reputationToPostImages: 10,
    bindNavPrevention: true,
    postfix: "",
    imageUploader:
    brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
    contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
    allowUrls: true
    ,
    noCode: true, onDemand: true,
    discardSelector: ".discard-answer"
    ,immediatelyShowMarkdownHelp:true
    );



    );













    draft saved

    draft discarded


















    StackExchange.ready(
    function ()
    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3201479%2fhow-do-i-prove-this-combinatorial-identity%23new-answer', 'question_page');

    );

    Post as a guest















    Required, but never shown

























    3 Answers
    3






    active

    oldest

    votes








    3 Answers
    3






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes









    15












    $begingroup$

    Here is a combinatorial proof. Both sides of the equation answer the following question:




    How many sequences are there of length $2n+1$, with entries in $0,1,2$, such that



    • at least one of the entries is a $2$, and

    • there are exactly $n$ zeroes to the left of the leftmost $2$?



    LHS:



    Suppose the leftmost $2$ occurs in spot $k+1$. Among the $k$ spots before hand, you must choose $n$ of the entries to be zero. The $2n+1-(k+1)=2n-k$ spots afterward can be anything. There are $binomkn3^2n-k$ ways to do this. Then sum over $k$.



    RHS:



    Suppose there are $j$ entries which are equal to $0$ or $2$. Choose those entries which are equal to $0$ or $2$ in $binom2n+1j$ ways. The leftmost $n$ of these entries must be zero, the $(n+1)^st$ entry must be two, then the remaining $j-(n+1)$ entries can be chosen freely among $0$ and $2$. There are $binom2n+1j2^j-(n+1)$ ways to do this, then sum over $j$.






    share|cite|improve this answer









    $endgroup$








    • 2




      $begingroup$
      How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
      $endgroup$
      – Icycarus
      12 hours ago






    • 5




      $begingroup$
      @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
      $endgroup$
      – Mike Earnest
      12 hours ago






    • 2




      $begingroup$
      @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
      $endgroup$
      – Mitch
      6 hours ago















    15












    $begingroup$

    Here is a combinatorial proof. Both sides of the equation answer the following question:




    How many sequences are there of length $2n+1$, with entries in $0,1,2$, such that



    • at least one of the entries is a $2$, and

    • there are exactly $n$ zeroes to the left of the leftmost $2$?



    LHS:



    Suppose the leftmost $2$ occurs in spot $k+1$. Among the $k$ spots before hand, you must choose $n$ of the entries to be zero. The $2n+1-(k+1)=2n-k$ spots afterward can be anything. There are $binomkn3^2n-k$ ways to do this. Then sum over $k$.



    RHS:



    Suppose there are $j$ entries which are equal to $0$ or $2$. Choose those entries which are equal to $0$ or $2$ in $binom2n+1j$ ways. The leftmost $n$ of these entries must be zero, the $(n+1)^st$ entry must be two, then the remaining $j-(n+1)$ entries can be chosen freely among $0$ and $2$. There are $binom2n+1j2^j-(n+1)$ ways to do this, then sum over $j$.






    share|cite|improve this answer









    $endgroup$








    • 2




      $begingroup$
      How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
      $endgroup$
      – Icycarus
      12 hours ago






    • 5




      $begingroup$
      @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
      $endgroup$
      – Mike Earnest
      12 hours ago






    • 2




      $begingroup$
      @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
      $endgroup$
      – Mitch
      6 hours ago













    15












    15








    15





    $begingroup$

    Here is a combinatorial proof. Both sides of the equation answer the following question:




    How many sequences are there of length $2n+1$, with entries in $0,1,2$, such that



    • at least one of the entries is a $2$, and

    • there are exactly $n$ zeroes to the left of the leftmost $2$?



    LHS:



    Suppose the leftmost $2$ occurs in spot $k+1$. Among the $k$ spots before hand, you must choose $n$ of the entries to be zero. The $2n+1-(k+1)=2n-k$ spots afterward can be anything. There are $binomkn3^2n-k$ ways to do this. Then sum over $k$.



    RHS:



    Suppose there are $j$ entries which are equal to $0$ or $2$. Choose those entries which are equal to $0$ or $2$ in $binom2n+1j$ ways. The leftmost $n$ of these entries must be zero, the $(n+1)^st$ entry must be two, then the remaining $j-(n+1)$ entries can be chosen freely among $0$ and $2$. There are $binom2n+1j2^j-(n+1)$ ways to do this, then sum over $j$.






    share|cite|improve this answer









    $endgroup$



    Here is a combinatorial proof. Both sides of the equation answer the following question:




    How many sequences are there of length $2n+1$, with entries in $0,1,2$, such that



    • at least one of the entries is a $2$, and

    • there are exactly $n$ zeroes to the left of the leftmost $2$?



    LHS:



    Suppose the leftmost $2$ occurs in spot $k+1$. Among the $k$ spots before hand, you must choose $n$ of the entries to be zero. The $2n+1-(k+1)=2n-k$ spots afterward can be anything. There are $binomkn3^2n-k$ ways to do this. Then sum over $k$.



    RHS:



    Suppose there are $j$ entries which are equal to $0$ or $2$. Choose those entries which are equal to $0$ or $2$ in $binom2n+1j$ ways. The leftmost $n$ of these entries must be zero, the $(n+1)^st$ entry must be two, then the remaining $j-(n+1)$ entries can be chosen freely among $0$ and $2$. There are $binom2n+1j2^j-(n+1)$ ways to do this, then sum over $j$.







    share|cite|improve this answer












    share|cite|improve this answer



    share|cite|improve this answer










    answered 12 hours ago









    Mike EarnestMike Earnest

    28.6k22255




    28.6k22255







    • 2




      $begingroup$
      How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
      $endgroup$
      – Icycarus
      12 hours ago






    • 5




      $begingroup$
      @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
      $endgroup$
      – Mike Earnest
      12 hours ago






    • 2




      $begingroup$
      @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
      $endgroup$
      – Mitch
      6 hours ago












    • 2




      $begingroup$
      How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
      $endgroup$
      – Icycarus
      12 hours ago






    • 5




      $begingroup$
      @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
      $endgroup$
      – Mike Earnest
      12 hours ago






    • 2




      $begingroup$
      @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
      $endgroup$
      – Mitch
      6 hours ago







    2




    2




    $begingroup$
    How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
    $endgroup$
    – Icycarus
    12 hours ago




    $begingroup$
    How did you get to the process of the answer? The way that you thought of the answer is quite a unique way and I was wondering if you can share how you managed to think about this solution
    $endgroup$
    – Icycarus
    12 hours ago




    5




    5




    $begingroup$
    @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
    $endgroup$
    – Mike Earnest
    12 hours ago




    $begingroup$
    @Icycarus The LHS has a fixed lower index and changing upper index. This reminded me of the Hockey stick identity, whose proof involves conditioning on where the largest element of a subset lies. Since there was a $3^i$ afterwards, I figured ternary sequences had to be involved somehow.
    $endgroup$
    – Mike Earnest
    12 hours ago




    2




    2




    $begingroup$
    @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
    $endgroup$
    – Mitch
    6 hours ago




    $begingroup$
    @Icycarus The method of counting combinatorial objects is called 'combinatrial proof'. Most of math taught in school is concerned with manipulating algebraic structures; combinatorial identities are actually intended to be mapped to ... let's say somewhat real-world arrangements of objects. And the kind of proof used here involves manipulation of those arrangements rather than straight algebra. You'll get the feel for it with some much more basic identity proofs. There's often an 'easy' map from algebraic op to combinatorial op.
    $endgroup$
    – Mitch
    6 hours ago











    2












    $begingroup$

    Using your functions, consider
    $$
    3^n f_2(frac13) = 3^n frac1(1-frac13)^n+1 = frac32 (frac92)^n\ = n choose n3^n + n+1 choose n3^n-1 + cdots + 2n choose n + colorred 2n +1 choose n 3^-1+ cdots
    $$

    and further
    $$
    2^n f_4 (frac12) = 2^n (frac32)^2n+1 = frac32 (frac92)^n \= 2n+1 choose 2n+12^n + 2n+1 choose 2n2^n-1 + cdots + 2n+1 choose n+1 + colorred 2n +1 choose n 2^-1+ cdots + 2n +1 choose 0 2^-n-1
    $$

    The two expressions both equal $frac32 (frac92)^n$, and the first $n+1$ many terms represent the LHS and RHS of the original question. The terms in red are extra terms: once it is established that these terms also equal, the questions is solved. That is, show that
    $$
    sum_k=1^infty2n +k choose n 3^-k = sum_k=1^n+12n +1 choose n+1-k 2^-k
    $$






    share|cite|improve this answer











    $endgroup$








    • 2




      $begingroup$
      I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
      $endgroup$
      – Mike Earnest
      12 hours ago










    • $begingroup$
      @MikeEarnest I resolved it and put it in the main text.
      $endgroup$
      – Andreas
      4 hours ago















    2












    $begingroup$

    Using your functions, consider
    $$
    3^n f_2(frac13) = 3^n frac1(1-frac13)^n+1 = frac32 (frac92)^n\ = n choose n3^n + n+1 choose n3^n-1 + cdots + 2n choose n + colorred 2n +1 choose n 3^-1+ cdots
    $$

    and further
    $$
    2^n f_4 (frac12) = 2^n (frac32)^2n+1 = frac32 (frac92)^n \= 2n+1 choose 2n+12^n + 2n+1 choose 2n2^n-1 + cdots + 2n+1 choose n+1 + colorred 2n +1 choose n 2^-1+ cdots + 2n +1 choose 0 2^-n-1
    $$

    The two expressions both equal $frac32 (frac92)^n$, and the first $n+1$ many terms represent the LHS and RHS of the original question. The terms in red are extra terms: once it is established that these terms also equal, the questions is solved. That is, show that
    $$
    sum_k=1^infty2n +k choose n 3^-k = sum_k=1^n+12n +1 choose n+1-k 2^-k
    $$






    share|cite|improve this answer











    $endgroup$








    • 2




      $begingroup$
      I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
      $endgroup$
      – Mike Earnest
      12 hours ago










    • $begingroup$
      @MikeEarnest I resolved it and put it in the main text.
      $endgroup$
      – Andreas
      4 hours ago













    2












    2








    2





    $begingroup$

    Using your functions, consider
    $$
    3^n f_2(frac13) = 3^n frac1(1-frac13)^n+1 = frac32 (frac92)^n\ = n choose n3^n + n+1 choose n3^n-1 + cdots + 2n choose n + colorred 2n +1 choose n 3^-1+ cdots
    $$

    and further
    $$
    2^n f_4 (frac12) = 2^n (frac32)^2n+1 = frac32 (frac92)^n \= 2n+1 choose 2n+12^n + 2n+1 choose 2n2^n-1 + cdots + 2n+1 choose n+1 + colorred 2n +1 choose n 2^-1+ cdots + 2n +1 choose 0 2^-n-1
    $$

    The two expressions both equal $frac32 (frac92)^n$, and the first $n+1$ many terms represent the LHS and RHS of the original question. The terms in red are extra terms: once it is established that these terms also equal, the questions is solved. That is, show that
    $$
    sum_k=1^infty2n +k choose n 3^-k = sum_k=1^n+12n +1 choose n+1-k 2^-k
    $$






    share|cite|improve this answer











    $endgroup$



    Using your functions, consider
    $$
    3^n f_2(frac13) = 3^n frac1(1-frac13)^n+1 = frac32 (frac92)^n\ = n choose n3^n + n+1 choose n3^n-1 + cdots + 2n choose n + colorred 2n +1 choose n 3^-1+ cdots
    $$

    and further
    $$
    2^n f_4 (frac12) = 2^n (frac32)^2n+1 = frac32 (frac92)^n \= 2n+1 choose 2n+12^n + 2n+1 choose 2n2^n-1 + cdots + 2n+1 choose n+1 + colorred 2n +1 choose n 2^-1+ cdots + 2n +1 choose 0 2^-n-1
    $$

    The two expressions both equal $frac32 (frac92)^n$, and the first $n+1$ many terms represent the LHS and RHS of the original question. The terms in red are extra terms: once it is established that these terms also equal, the questions is solved. That is, show that
    $$
    sum_k=1^infty2n +k choose n 3^-k = sum_k=1^n+12n +1 choose n+1-k 2^-k
    $$







    share|cite|improve this answer














    share|cite|improve this answer



    share|cite|improve this answer








    edited 4 hours ago

























    answered 12 hours ago









    AndreasAndreas

    8,5111137




    8,5111137







    • 2




      $begingroup$
      I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
      $endgroup$
      – Mike Earnest
      12 hours ago










    • $begingroup$
      @MikeEarnest I resolved it and put it in the main text.
      $endgroup$
      – Andreas
      4 hours ago












    • 2




      $begingroup$
      I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
      $endgroup$
      – Mike Earnest
      12 hours ago










    • $begingroup$
      @MikeEarnest I resolved it and put it in the main text.
      $endgroup$
      – Andreas
      4 hours ago







    2




    2




    $begingroup$
    I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
    $endgroup$
    – Mike Earnest
    12 hours ago




    $begingroup$
    I think something is awry. For $3^nf_2(1/3)$, the sum is infinite, but OP’s sum is finite.
    $endgroup$
    – Mike Earnest
    12 hours ago












    $begingroup$
    @MikeEarnest I resolved it and put it in the main text.
    $endgroup$
    – Andreas
    4 hours ago




    $begingroup$
    @MikeEarnest I resolved it and put it in the main text.
    $endgroup$
    – Andreas
    4 hours ago











    1












    $begingroup$

    We seek to show that



    $$sum_q=0^n 2n-qchoose n 3^q
    = sum_q=0^n 2n+1choose n+1+q 2^q.$$



    We have for the LHS



    $$sum_q=0^n 2n-qchoose n-q 3^q
    = sum_q=0^n 3^q [z^n-q] (1+z)^2n-q
    \ = [z^n] (1+z)^2n sum_q=0^n 3^q z^q (1+z)^-q.$$



    The coefficient extractor controls the range and we obtain



    $$[z^n] (1+z)^2n sum_qge 0 3^q z^q (1+z)^-q
    = [z^n] (1+z)^2n frac11-3z/(1+z)
    \ = [z^n] (1+z)^2n+1 frac11-2z.$$



    We could conclude at this point by inspection. Continuing anyway we
    get for the RHS



    $$sum_q=0^n 2n+1choose n-q 2^q
    = sum_q=0^n 2^q [z^n-q] (1+z)^2n+1
    \ = [z^n] (1+z)^2n+1 sum_q=0^n 2^q z^q.$$



    The coefficient extractor once more controls the range and we obtain



    $$[z^n] (1+z)^2n+1 sum_qge 0 2^q z^q
    = [z^n] (1+z)^2n+1 frac11-2z.$$



    The two generating functions are the same and we have equality for LHS
    and RHS.






    share|cite|improve this answer









    $endgroup$

















      1












      $begingroup$

      We seek to show that



      $$sum_q=0^n 2n-qchoose n 3^q
      = sum_q=0^n 2n+1choose n+1+q 2^q.$$



      We have for the LHS



      $$sum_q=0^n 2n-qchoose n-q 3^q
      = sum_q=0^n 3^q [z^n-q] (1+z)^2n-q
      \ = [z^n] (1+z)^2n sum_q=0^n 3^q z^q (1+z)^-q.$$



      The coefficient extractor controls the range and we obtain



      $$[z^n] (1+z)^2n sum_qge 0 3^q z^q (1+z)^-q
      = [z^n] (1+z)^2n frac11-3z/(1+z)
      \ = [z^n] (1+z)^2n+1 frac11-2z.$$



      We could conclude at this point by inspection. Continuing anyway we
      get for the RHS



      $$sum_q=0^n 2n+1choose n-q 2^q
      = sum_q=0^n 2^q [z^n-q] (1+z)^2n+1
      \ = [z^n] (1+z)^2n+1 sum_q=0^n 2^q z^q.$$



      The coefficient extractor once more controls the range and we obtain



      $$[z^n] (1+z)^2n+1 sum_qge 0 2^q z^q
      = [z^n] (1+z)^2n+1 frac11-2z.$$



      The two generating functions are the same and we have equality for LHS
      and RHS.






      share|cite|improve this answer









      $endgroup$















        1












        1








        1





        $begingroup$

        We seek to show that



        $$sum_q=0^n 2n-qchoose n 3^q
        = sum_q=0^n 2n+1choose n+1+q 2^q.$$



        We have for the LHS



        $$sum_q=0^n 2n-qchoose n-q 3^q
        = sum_q=0^n 3^q [z^n-q] (1+z)^2n-q
        \ = [z^n] (1+z)^2n sum_q=0^n 3^q z^q (1+z)^-q.$$



        The coefficient extractor controls the range and we obtain



        $$[z^n] (1+z)^2n sum_qge 0 3^q z^q (1+z)^-q
        = [z^n] (1+z)^2n frac11-3z/(1+z)
        \ = [z^n] (1+z)^2n+1 frac11-2z.$$



        We could conclude at this point by inspection. Continuing anyway we
        get for the RHS



        $$sum_q=0^n 2n+1choose n-q 2^q
        = sum_q=0^n 2^q [z^n-q] (1+z)^2n+1
        \ = [z^n] (1+z)^2n+1 sum_q=0^n 2^q z^q.$$



        The coefficient extractor once more controls the range and we obtain



        $$[z^n] (1+z)^2n+1 sum_qge 0 2^q z^q
        = [z^n] (1+z)^2n+1 frac11-2z.$$



        The two generating functions are the same and we have equality for LHS
        and RHS.






        share|cite|improve this answer









        $endgroup$



        We seek to show that



        $$sum_q=0^n 2n-qchoose n 3^q
        = sum_q=0^n 2n+1choose n+1+q 2^q.$$



        We have for the LHS



        $$sum_q=0^n 2n-qchoose n-q 3^q
        = sum_q=0^n 3^q [z^n-q] (1+z)^2n-q
        \ = [z^n] (1+z)^2n sum_q=0^n 3^q z^q (1+z)^-q.$$



        The coefficient extractor controls the range and we obtain



        $$[z^n] (1+z)^2n sum_qge 0 3^q z^q (1+z)^-q
        = [z^n] (1+z)^2n frac11-3z/(1+z)
        \ = [z^n] (1+z)^2n+1 frac11-2z.$$



        We could conclude at this point by inspection. Continuing anyway we
        get for the RHS



        $$sum_q=0^n 2n+1choose n-q 2^q
        = sum_q=0^n 2^q [z^n-q] (1+z)^2n+1
        \ = [z^n] (1+z)^2n+1 sum_q=0^n 2^q z^q.$$



        The coefficient extractor once more controls the range and we obtain



        $$[z^n] (1+z)^2n+1 sum_qge 0 2^q z^q
        = [z^n] (1+z)^2n+1 frac11-2z.$$



        The two generating functions are the same and we have equality for LHS
        and RHS.







        share|cite|improve this answer












        share|cite|improve this answer



        share|cite|improve this answer










        answered 2 hours ago









        Marko RiedelMarko Riedel

        41.8k341112




        41.8k341112



























            draft saved

            draft discarded
















































            Thanks for contributing an answer to Mathematics Stack Exchange!


            • Please be sure to answer the question. Provide details and share your research!

            But avoid


            • Asking for help, clarification, or responding to other answers.

            • Making statements based on opinion; back them up with references or personal experience.

            Use MathJax to format equations. MathJax reference.


            To learn more, see our tips on writing great answers.




            draft saved


            draft discarded














            StackExchange.ready(
            function ()
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3201479%2fhow-do-i-prove-this-combinatorial-identity%23new-answer', 'question_page');

            );

            Post as a guest















            Required, but never shown





















































            Required, but never shown














            Required, but never shown












            Required, but never shown







            Required, but never shown

































            Required, but never shown














            Required, but never shown












            Required, but never shown







            Required, but never shown







            Popular posts from this blog

            六本木駅

            Integral that is continuous and looks like it converges to a geometric seriesTesting if a geometric series converges by taking limit to infinitySummation of arithmetic-geometric series of higher orderGeometric series with polynomial exponentHow to Recognize a Geometric SeriesShowing an integral equality with series over the integersDiscontinuity of a series of continuous functionsReasons why a Series ConvergesSum of infinite geometric series with two terms in summationUsing geometric series for computing IntegralsLimit of geometric series sum when $r = 1$

            Joseph Lister