McIntyre MathsMathematics Extension 2
Extension 2 › Proof › Lesson 9

Lesson 9 · Further induction III: induction in context

Mathematics Extension 2 · Proof · Outcome ME2-12-01 · Term 4 · Week 5
Learning intentionTo apply mathematical induction in a range of contexts: recursive sequences, derivatives, trigonometric identities, the binomial theorem, logarithms and geometry.
Success criteria · I can…
  • prove a formula for a recursively defined sequence
  • prove results about derivatives and trigonometric products or sums
  • prove Pascal’s identity and the binomial theorem by induction
  • prove logarithm results and simple geometric results by induction

Lesson flow

Warm-up · Part A

Slide 2
(a) If a1=2a_1=2 and an+1=3an−2a_{n+1}=3a_n-2, find a2a_2, a3a_3, a4a_4.
Show answerHide answer
Answer
a2=4a_2=4, a3=10a_3=10, a4=28a_4=28. A pattern: an=3n−1+1a_n=3^{n-1}+1.
Why
Apply the rule repeatedly: 3(2)−2=43(2)-2=4, 3(4)−2=103(4)-2=10, 3(10)−2=283(10)-2=28. These equal 31+13^1+1, 32+13^2+1, 33+13^3+1.
(b) Differentiate xexxe^x, then differentiate the answer again.
Show answerHide answer
Answer
(x+1)ex(x+1)e^x, then (x+2)ex(x+2)e^x. A pattern: (x+n)ex(x+n)e^x.
Why
Product rule: ex+xex=(x+1)exe^x+xe^x=(x+1)e^x. Again: ex+(x+1)ex=(x+2)exe^x+(x+1)e^x=(x+2)e^x.
(c) Write sin⁡2A\sin2A and cos⁡2A\cos2A in terms of sin⁡A\sin A and cos⁡A\cos A.
Show answerHide answer
Answer
sin⁡2A=2sin⁡Acos⁡A\sin2A=2\sin A\cos A and cos⁡2A=cos⁡2A−sin⁡2A=2cos⁡2A−1=1−2sin⁡2A\cos2A=\cos^2A-\sin^2A=2\cos^2A-1=1-2\sin^2A.
Why
These are the double angle formulas. The three forms of cos⁡2A\cos2A come from sin⁡2A+cos⁡2A=1\sin^2A+\cos^2A=1.
(d) Show that 1+cos⁡θ=2cos⁡2θ21+\cos\theta=2\cos^2\dfrac\theta2.
Show answerHide answer
Answer
Put A=θ2A=\dfrac\theta2 in cos⁡2A=2cos⁡2A−1\cos2A=2\cos^2A-1: cos⁡θ=2cos⁡2θ2−1\cos\theta=2\cos^2\dfrac\theta2-1, so 1+cos⁡θ=2cos⁡2θ21+\cos\theta=2\cos^2\dfrac\theta2.
Why
Parts (a), (b) and (d) are all used later in the lesson.

Recursive sequences

Slide 3
Definition
Proving a formula for a recursively defined sequence

A sequence is defined recursively when each term is given in terms of earlier terms, for example a1=2a_1=2, an+1=3an−2a_{n+1}=3a_n-2.

To prove a formula for ana_n: check a1a_1; assume the formula for aka_k; then substitute it into the recursion to get ak+1a_{k+1}.

Example 1

Prove by induction that if a1=2a_1=2 and an+1=3an−2a_{n+1}=3a_n-2, then an=3n−1+1a_n=3^{n-1}+1 for all integers n≥1n\ge1.
Show answerHide answer
Answer
Let P(n)P(n) be the statement an=3n−1+1a_n=3^{n-1}+1.
Step 1. P(1)P(1): 30+1=2=a13^0+1=2=a_1, so P(1)P(1) is true.
Step 2. Assume P(k)P(k): ak=3k−1+1a_k=3^{k-1}+1 for some integer k≥1k\ge1.
Step 3. ak+1=3ak−2=3(3k−1+1)−2=3k+1.a_{k+1}=3a_k-2=3\left(3^{k-1}+1\right)-2=3^k+1. So P(k+1)P(k+1) is true when P(k)P(k) is true.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
The assumption is used inside the recursion: replace aka_k in ak+1=3ak−2a_{k+1}=3a_k-2 by the assumed formula. This differs from series, where you add a term to the assumed sum.

Example 2 · A recursive fraction

Slide 4

Example 2

A sequence is defined by u1=1u_1=1 and un+1=un1+unu_{n+1}=\dfrac{u_n}{1+u_n}. Find u2u_2, u3u_3, u4u_4, make a conjecture for unu_n, and prove it by induction.
Show answerHide answer
Answer
u2=12u_2=\dfrac12, u3=13u_3=\dfrac13, u4=14u_4=\dfrac14. Conjecture: un=1nu_n=\dfrac1n.
Step 1. u1=1=11u_1=1=\dfrac11.
Step 2. Assume uk=1ku_k=\dfrac1k for some integer k≥1k\ge1.
Step 3. uk+1=uk1+uk=1/k1+1/k=1/k(k+1)/k=1k+1.u_{k+1}=\frac{u_k}{1+u_k}=\frac{1/k}{1+1/k}=\frac{1/k}{(k+1)/k}=\frac1{k+1}.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
Conjecture first from the pattern, then prove it: this is the reasoning and communication the syllabus asks for. In Step 3, multiply top and bottom of the complex fraction by kk.

HSC practice · A nested square root

Slide 5
2022 HSC Q13(b)The sequence ana_n is defined by a1=2a_1=\sqrt2 and an+1=2+ana_{n+1}=\sqrt{2+a_n} for n≥1n\ge1. Prove by induction that an=2cos⁡(π2n+1)a_n=2\cos\left(\dfrac\pi{2^{n+1}}\right) for all integers n≥1n\ge1. You may use 2+2cos⁡θ=2cos⁡θ2\sqrt{2+2\cos\theta}=2\cos\dfrac\theta2 for 0≤θ≤π0\le\theta\le\pi (warm-up (d)).
Show answerHide answer
Answer
Step 1. P(1)P(1): 2cos⁡π4=2×22=2=a12\cos\dfrac\pi4=2\times\dfrac{\sqrt2}2=\sqrt2=a_1, so P(1)P(1) is true.
Step 2. Assume ak=2cos⁡π2k+1a_k=2\cos\dfrac\pi{2^{k+1}} for some integer k≥1k\ge1.
Step 3. ak+1=2+2cos⁡π2k+1=4cos⁡2π2k+2=2cos⁡π2k+2,a_{k+1}=\sqrt{2+2\cos\frac\pi{2^{k+1}}}=\sqrt{4\cos^2\frac\pi{2^{k+2}}}=2\cos\frac\pi{2^{k+2}}, taking the positive root since 0<π2k+2<π20<\dfrac\pi{2^{k+2}}<\dfrac\pi2, so the cosine is positive.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
Use the given identity with θ=π2k+1\theta=\dfrac\pi{2^{k+1}}, which lies in [0,π][0,\pi]. Then θ2=π2k+2\dfrac\theta2=\dfrac\pi{2^{k+2}}. The positive square root must be justified: 4cos⁡2ϕ=2∣cos⁡ϕ∣=2cos⁡ϕ\sqrt{4\cos^2\phi}=2|\cos\phi|=2\cos\phi only because cos⁡ϕ>0\cos\phi>0. (Wording adapted from the paper.)

Example 3 · Calculus and trigonometry

Slide 6

Example 3

(a) Prove by induction that dndxn(xex)=(x+n)ex\dfrac{d^n}{dx^n}\left(xe^x\right)=(x+n)e^x for all integers n≥1n\ge1.
Show answerHide answer
Answer
Step 1. P(1)P(1): ddx(xex)=ex+xex=(x+1)ex\dfrac d{dx}\left(xe^x\right)=e^x+xe^x=(x+1)e^x.
Step 2. Assume dkdxk(xex)=(x+k)ex\dfrac{d^k}{dx^k}\left(xe^x\right)=(x+k)e^x for some integer k≥1k\ge1.
Step 3. dk+1dxk+1(xex)=ddx[(x+k)ex]=ex+(x+k)ex=(x+k+1)ex.\frac{d^{k+1}}{dx^{k+1}}\left(xe^x\right)=\frac d{dx}\left[(x+k)e^x\right]=e^x+(x+k)e^x=(x+k+1)e^x.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
The (k+1)(k+1)th derivative is the derivative of the kkth derivative. Use the product rule on (x+k)ex(x+k)e^x, treating kk as a constant.
(b) Prove by induction that, for sin⁡x≠0\sin x\ne0, cos⁡xcos⁡2xcos⁡4x⋯cos⁡(2n−1x)=sin⁡(2nx)2nsin⁡x\cos x\cos2x\cos4x\cdots\cos\left(2^{n-1}x\right)=\dfrac{\sin\left(2^nx\right)}{2^n\sin x}.
Show answerHide answer
Answer
Step 1. P(1)P(1): sin⁡2x2sin⁡x=2sin⁡xcos⁡x2sin⁡x=cos⁡x\dfrac{\sin2x}{2\sin x}=\dfrac{2\sin x\cos x}{2\sin x}=\cos x.
Step 2. Assume the product up to cos⁡(2k−1x)\cos\left(2^{k-1}x\right) equals sin⁡(2kx)2ksin⁡x\dfrac{\sin\left(2^kx\right)}{2^k\sin x} for some integer k≥1k\ge1.
Step 3. Multiply by cos⁡(2kx)\cos\left(2^kx\right): sin⁡(2kx)cos⁡(2kx)2ksin⁡x=12sin⁡(2k+1x)2ksin⁡x=sin⁡(2k+1x)2k+1sin⁡x.\frac{\sin(2^kx)\cos(2^kx)}{2^k\sin x}=\frac{\tfrac12\sin\left(2^{k+1}x\right)}{2^k\sin x}=\frac{\sin\left(2^{k+1}x\right)}{2^{k+1}\sin x}.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
The (k+1)(k+1)th product is the kkth product times one new factor, cos⁡(2kx)\cos(2^kx). The double angle formula with A=2kxA=2^kx gives sin⁡Acos⁡A=12sin⁡2A\sin A\cos A=\tfrac12\sin2A.

HSC practice · A trigonometric sum

Slide 7
2019 HSC Q14(c)(i) Show that cot⁡x−cot⁡2x=csc⁡2x\cot x-\cot2x=\csc2x.
(ii) Hence prove by induction that for all integers n≥1n\ge1, ∑r=1ncsc⁡(2rx)=cot⁡x−cot⁡(2nx).\sum_{r=1}^{n}\csc\left(2^rx\right)=\cot x-\cot\left(2^nx\right).
Show answerHide answer
Answer
(i) cot⁡x−cot⁡2x=cos⁡xsin⁡x−cos⁡2x2sin⁡xcos⁡x=2cos⁡2x−cos⁡2x2sin⁡xcos⁡x=1sin⁡2x=csc⁡2x.\cot x-\cot2x=\frac{\cos x}{\sin x}-\frac{\cos2x}{2\sin x\cos x}=\frac{2\cos^2x-\cos2x}{2\sin x\cos x}=\frac1{\sin2x}=\csc2x. (ii) Step 1. P(1)P(1): csc⁡2x=cot⁡x−cot⁡2x\csc2x=\cot x-\cot2x by (i).
Step 2. Assume ∑r=1kcsc⁡(2rx)=cot⁡x−cot⁡(2kx)\displaystyle\sum_{r=1}^{k}\csc\left(2^rx\right)=\cot x-\cot\left(2^kx\right) for some integer k≥1k\ge1.
Step 3. Add csc⁡(2k+1x)=cot⁡(2kx)−cot⁡(2k+1x)\csc\left(2^{k+1}x\right)=\cot\left(2^kx\right)-\cot\left(2^{k+1}x\right), which is (i) with xx replaced by 2kx2^kx: ∑r=1k+1csc⁡(2rx)=cot⁡x−cot⁡(2kx)+cot⁡(2kx)−cot⁡(2k+1x)=cot⁡x−cot⁡(2k+1x).\sum_{r=1}^{k+1}\csc\left(2^rx\right)=\cot x-\cot\left(2^kx\right)+\cot\left(2^kx\right)-\cot\left(2^{k+1}x\right)=\cot x-\cot\left(2^{k+1}x\right).
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
In (i), cos⁡2x=2cos⁡2x−1\cos2x=2\cos^2x-1 makes the numerator 11, and 2sin⁡xcos⁡x=sin⁡2x2\sin x\cos x=\sin2x. The key idea in (ii) is “(i) with xx replaced by 2kx2^kx”, which students often miss. (Cosec is written csc here; both are fine. Wording adapted from the paper.)

Spot the flaw · What’s missing?

Slide 8

A student wrote this. Find the flaw, explain why it matters, then write a correct version.

Student’s claim

If a1=0a_1=0 and an+1=2+ana_{n+1}=\sqrt{2+a_n}, then an=2cos⁡π2na_n=2\cos\dfrac\pi{2^n} for all n≥1n\ge1.

  1. n=1n=1: 2cos⁡π2=0=a12\cos\dfrac\pi2=0=a_1. Assume ak=2cos⁡π2ka_k=2\cos\dfrac\pi{2^k}.
  2. ak+1=2+2cos⁡π2k=4cos⁡2π2k+1=2cos⁡π2k+1a_{k+1}=\sqrt{2+2\cos\dfrac\pi{2^k}}=\sqrt{4\cos^2\dfrac\pi{2^{k+1}}}=2\cos\dfrac\pi{2^{k+1}}.
  3. Hence true for all n≥1n\ge1 by induction. □
Spot the flawWhat is missing from this proof, and why would a marker deduct for it?
Show answerHide answer
Answer
Nothing written is false, but three things are missing.
1. The step 2+2cos⁡2θ=4cos⁡2θ2+2\cos2\theta=4\cos^2\theta (from cos⁡2θ=2cos⁡2θ−1\cos2\theta=2\cos^2\theta-1, with θ=π2k+1\theta=\dfrac\pi{2^{k+1}}) is not shown.
2. 4cos⁡2θ=2∣cos⁡θ∣\sqrt{4\cos^2\theta}=2|\cos\theta|, so we must justify cos⁡π2k+1>0\cos\dfrac\pi{2^{k+1}}>0. This is true because 0<π2k+1≤π4<π20<\dfrac\pi{2^{k+1}}\le\dfrac\pi4<\dfrac\pi2 for k≥1k\ge1.
3. There is no conclusion naming P(k+1)P(k+1) and the start value: “P(1)P(1) is true, and P(k)⇒P(k+1)P(k)\Rightarrow P(k+1), so P(n)P(n) is true for all n≥1n\ge1 by induction.”
Why
Compare the 2022 HSC question earlier in this lesson, which has the same structure with a1=2a_1=\sqrt2. In an induction proof every step must be justified, in particular the choice of the positive square root, and the conclusion must be stated.

Warm-up · Part B

Slide 9

A fresh start for the second half of the lesson.

(a) Evaluate (52)+(53)\dbinom52+\dbinom53 and (63)\dbinom63. What do you notice?
Show answerHide answer
Answer
10+10=20=(63)10+10=20=\dbinom63.
Why
This is Pascal’s identity: (52)+(53)=(63)\dbinom52+\dbinom53=\dbinom63.
(b) Expand (1+x)4(1+x)^4 using Pascal’s triangle.
Show answerHide answer
Answer
1+4x+6x2+4x3+x41+4x+6x^2+4x^3+x^4.
Why
The row of Pascal’s triangle 1, 4, 6, 4, 1 gives the coefficients.
(c) Simplify ln⁡(1−1r2)\ln\left(1-\dfrac1{r^2}\right) as a single log of a product of two fractions.
Show answerHide answer
Answer
ln⁡[r−1r×r+1r]\ln\left[\dfrac{r-1}r\times\dfrac{r+1}r\right].
Why
1−1r2=r2−1r2=(r−1)(r+1)r21-\dfrac1{r^2}=\dfrac{r^2-1}{r^2}=\dfrac{(r-1)(r+1)}{r^2}, which is the product r−1r×r+1r\dfrac{r-1}r\times\dfrac{r+1}r. Equivalently ln⁡r−1r+ln⁡r+1r\ln\dfrac{r-1}r+\ln\dfrac{r+1}r.
(d) How many diagonals does a hexagon have?
Show answerHide answer
Answer
9.
Why
Each of the 6 vertices joins to 3 non-adjacent vertices, and each diagonal is counted twice: 6×32=9\dfrac{6\times3}2=9.

The binomial theorem by induction

Slide 10
Theorem
Pascal’s identity and the binomial theorem

Pascal’s identity: (kr−1)+(kr)=(k+1r),1≤r≤k.\binom k{r-1}+\binom kr=\binom{k+1}r,\qquad 1\le r\le k.

Binomial theorem: (1+x)n=∑r=0n(nr)xrfor all integers n≥1.(1+x)^n=\sum_{r=0}^{n}\binom nrx^r\quad\text{for all integers }n\ge1.

Pascal’s identity is the key step: it combines the coefficients of xrx^r from (1+x)k×1(1+x)^k\times1 and (1+x)k×x(1+x)^k\times x.

Example 4

(a) Prove Pascal’s identity using the factorial formula.
Show answerHide answer
Answer
Using (nr)=n!r!(n−r)!\dbinom nr=\dfrac{n!}{r!(n-r)!}: (kr−1)+(kr)=k!(r−1)!(k−r+1)!+k!r!(k−r)!=k![r+(k−r+1)]r!(k−r+1)!=(k+1)!r!(k+1−r)!=(k+1r).\binom k{r-1}+\binom kr=\frac{k!}{(r-1)!(k-r+1)!}+\frac{k!}{r!(k-r)!}=\frac{k!\left[r+(k-r+1)\right]}{r!(k-r+1)!}=\frac{(k+1)!}{r!(k+1-r)!}=\binom{k+1}r.
Why
Use the common denominator r!(k−r+1)!r!(k-r+1)!. The first fraction needs a factor rr (since r!=r (r−1)!r!=r\,(r-1)!) and the second needs (k−r+1)(k-r+1) (since (k−r+1)!=(k−r+1)(k−r)!(k-r+1)!=(k-r+1)(k-r)!). The numerator becomes k!(k+1)k!(k+1).
(b) Prove the binomial theorem by induction.
Show answerHide answer
Answer
Step 1. P(1)P(1): (1+x)1=1+x=(10)+(11)x(1+x)^1=1+x=\dbinom10+\dbinom11x.
Step 2. Assume (1+x)k=∑r=0k(kr)xr(1+x)^k=\displaystyle\sum_{r=0}^{k}\binom krx^r for some integer k≥1k\ge1.
Step 3. (1+x)k+1=(1+x)∑r=0k(kr)xr=∑r=0k(kr)xr+∑r=0k(kr)xr+1.(1+x)^{k+1}=(1+x)\sum_{r=0}^{k}\binom krx^r=\sum_{r=0}^{k}\binom krx^r+\sum_{r=0}^{k}\binom krx^{r+1}. The coefficient of xrx^r for 1≤r≤k1\le r\le k is (kr)+(kr−1)=(k+1r)\dbinom kr+\dbinom k{r-1}=\dbinom{k+1}r by (a). The constant term is 1=(k+10)1=\dbinom{k+1}0 and the coefficient of xk+1x^{k+1} is 1=(k+1k+1)1=\dbinom{k+1}{k+1}. Hence (1+x)k+1=∑r=0k+1(k+1r)xr(1+x)^{k+1}=\displaystyle\sum_{r=0}^{k+1}\binom{k+1}rx^r.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
Write out a few terms rather than relying only on the sigma notation: the first sum gives the coefficients of xrx^r and the second sum, shifted along by one power, gives the coefficients of xrx^{r} from (kr−1)\dbinom k{r-1}. The binomial theorem was used in Extension 1; this is where it is proved.

Example 5 · Induction with logarithms

Slide 11

Example 5

(a) Given ln⁡(ab)=ln⁡a+ln⁡b\ln(ab)=\ln a+\ln b, prove by induction that for positive a1,…,ana_1,\dots,a_n: ln⁡(a1a2⋯an)=∑i=1nln⁡ai\ln\left(a_1a_2\cdots a_n\right)=\displaystyle\sum_{i=1}^{n}\ln a_i.
Show answerHide answer
Answer
Step 1. P(1)P(1): ln⁡a1=ln⁡a1\ln a_1=\ln a_1.
Step 2. Assume ln⁡(a1⋯ak)=∑i=1kln⁡ai\ln\left(a_1\cdots a_k\right)=\displaystyle\sum_{i=1}^{k}\ln a_i for some integer k≥1k\ge1.
Step 3. ln⁡(a1⋯akak+1)=ln⁡(a1⋯ak)+ln⁡ak+1=∑i=1kln⁡ai+ln⁡ak+1=∑i=1k+1ln⁡ai.\ln\left(a_1\cdots a_ka_{k+1}\right)=\ln\left(a_1\cdots a_k\right)+\ln a_{k+1}=\sum_{i=1}^{k}\ln a_i+\ln a_{k+1}=\sum_{i=1}^{k+1}\ln a_i.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
The log law ln⁡(ab)=ln⁡a+ln⁡b\ln(ab)=\ln a+\ln b is applied with a=a1⋯aka=a_1\cdots a_k and b=ak+1b=a_{k+1}. Then the assumption replaces ln⁡(a1⋯ak)\ln(a_1\cdots a_k).
(b) Prove by induction that for all integers n≥2n\ge2: ∑r=2nln⁡(1−1r2)=ln⁡(n+12n)\displaystyle\sum_{r=2}^{n}\ln\left(1-\frac1{r^2}\right)=\ln\left(\frac{n+1}{2n}\right).
Show answerHide answer
Answer
Step 1. P(2)P(2): ln⁡(1−14)=ln⁡34=ln⁡32×2\ln\left(1-\dfrac14\right)=\ln\dfrac34=\ln\dfrac{3}{2\times2}, true.
Step 2. Assume the result for some integer k≥2k\ge2.
Step 3. ∑r=2k+1ln⁡(1−1r2)=ln⁡k+12k+ln⁡(1−1(k+1)2)=ln⁡[k+12k×k(k+2)(k+1)2]=ln⁡k+22(k+1).\sum_{r=2}^{k+1}\ln\left(1-\frac1{r^2}\right)=\ln\frac{k+1}{2k}+\ln\left(1-\frac1{(k+1)^2}\right)=\ln\left[\frac{k+1}{2k}\times\frac{k(k+2)}{(k+1)^2}\right]=\ln\frac{k+2}{2(k+1)}.
Step 4. True for n=2n=2, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥2n\ge2 by induction.
Why
The starting value is n=2n=2. Use 1−1(k+1)2=(k+1)2−1(k+1)2=k(k+2)(k+1)21-\dfrac1{(k+1)^2}=\dfrac{(k+1)^2-1}{(k+1)^2}=\dfrac{k(k+2)}{(k+1)^2}. A non-induction method is to telescope using warm-up (c).

Example 6 · Induction in geometry

Slide 12

Think: when a new vertex is added between two neighbouring vertices, which new diagonals appear?

Example 6

Prove by induction that a convex polygon with nn sides has n(n−3)2\dfrac{n(n-3)}2 diagonals, for all integers n≥3n\ge3.
Show answerHide answer
Answer
Step 1. P(3)P(3): a triangle has 0 diagonals and 3×02=0\dfrac{3\times0}2=0.
Step 2. Assume a convex kk-gon has k(k−3)2\dfrac{k(k-3)}2 diagonals, for some integer k≥3k\ge3.
Step 3. Add a vertex VV outside the edge ABAB to make a convex (k+1)(k+1)-gon. The new diagonals are VV joined to the k−2k-2 vertices other than AA and BB, and the old side ABAB is now a diagonal: k−1k-1 new diagonals. k(k−3)2+k−1=k2−k−22=(k+1)(k−2)2=(k+1)((k+1)−3)2.\frac{k(k-3)}2+k-1=\frac{k^2-k-2}2=\frac{(k+1)(k-2)}2=\frac{(k+1)\big((k+1)-3\big)}2.
Step 4. True for n=3n=3, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥3n\ge3 by induction.
Why
In a geometry proof, Step 3 must say in words what the (k+1)(k+1)th step adds. Draw a hexagon becoming a heptagon: VV is adjacent only to AA and BB, so it makes diagonals to the other k−2k-2 vertices, and ABAB stops being a side.

Exit ticket

Slide 13
A sequence is defined by a1=5a_1=5 and an+1=2an−3a_{n+1}=2a_n-3. Prove by induction that an=2n+3a_n=2^n+3 for all integers n≥1n\ge1.
Show answerHide answer
Answer
Step 1. P(1)P(1): 2+3=5=a12+3=5=a_1, so P(1)P(1) is true.
Step 2. Assume ak=2k+3a_k=2^k+3 for some integer k≥1k\ge1.
Step 3. ak+1=2ak−3=2(2k+3)−3=2k+1+3a_{k+1}=2a_k-3=2\left(2^k+3\right)-3=2^{k+1}+3. So P(k+1)P(k+1) is true when P(k)P(k) is true.
Step 4. True for n=1n=1, and true for n=k⇒n=k+1n=k\Rightarrow n=k+1, so true for all integers n≥1n\ge1 by induction.
Why
Substitute the assumption into the recursion. Do not use the formula for ak+1a_{k+1} in Step 3, as that is what you are proving. Check: a2=7=22+3a_2=7=2^2+3.