One seemingly trivial yet useful strategy for doing math or problem solving in general is wishful thinking.
Wishful thinking is when you simplify/change the problem to another problem that you wish it looked like so you could solve it easily. After getting your "wish", you try to build on your evolved form to answer the original problem.
The MAA describes wishful thinking by this question in this article: "If you could change the question, what change would make it ridiculously easy?"
Example of Wishful Thinking
A simple example in which you could use wishful thinking would be to factor x4+x2+1. It's a fourth degree polynomial, which is hard to solve, so you wished it was a quadratic because quadratics are easier to solve. But wait! It's an even polynomial, so the polynomial could be rewritten as a quadratic: (x2)2+x2+1.
If you define y=x2, you can rewrite the polynomial as y2+y+1. Guess factoring leads nowhere, and you could only wish that it was a perfect square. Aha! We could force the polynomial to contain a perfect square by rewriting the expression as (y2+2y+1)−y=(y+1)2−y. Since y=x2, you can use difference of squares to factor it.
For more resources on wishful thinking, you can watch this video by the Mathematical Association of America here, or read The Art and Craft of Problem Solving (Chapter 2) by Paul Zeitz.
Many times when factoring polynomials, or solving complicated algebra equations, it is useful to make a clever substitution to make life easier for us.
Example
Solve x4+5x2+6=0.
This is a quartic. Normally, we try to find two binomials that multiply to this. With some guessing and checking, we find that (x2+3)(x2+2)=0 so our roots are ±3i and ±2i. However, there is a simpler way to do this. Let a=x2. We can express the equation in terms of a. x4+5x2+6=(x2)2+5(x2)+6=a2+5a+6. Now this is a quadratic we know how to handle. We can factor this as (a+2)(a+3)=0 so a=−2,−3. Solving for x yields
x2=−2⟹xx2=−3⟹x=±2i=±3i
This is a much easier way to do this. Of course this is just the beginning of how substitution can greatly simplify problems.
Find all values of x that satisfy 4x−3+4x−310=7.
Express your answer as an ordered list with no spaces. Use fractions if necessary.
The solution to the previous problem is as follows. Let q=4x−3. Now all we have to solve is q+q10=7 which is simple. Multiplying by q for each side yields q2+10=7q⟹q2−7q+10=0⟹(q−5)(q−2)=0⟹q=2,5.
Now we equate each value of q to 4x−3. So q=2⟹2=4x−3⟹4=4x−3⟹x=47q=5⟹5=4x−3⟹25=4x−3⟹x=7. Therefore, x=47,7.
Find all values of x such that 12(logx)2+11(logx)+2=0.
Assume that log is the base 10 logarithm.
Approximate to the nearest hundreds and as an ordered list.
Solution to previous problem
Let u=logx. We have
12u2+11u+2=0⟹12u2+3u+8u+2=0⟹3u(4u+1)+2(4u+1)=0⟹(4u+1)(3u+2)=0⟹u=−41,−32⟹{logx=−41logx=−32⟹{x=10−1/4x=10−2/3⟹{x≈0.56x≈0.22
In this article we are going to talk about a simple yet powerful tactic, massaging. That is a pretty bad name, doesn't it? (Maybe you think it is a good name, I don't care) But actually, if you massage an expression or an equation or even a term, you are "loosen them up" so that you can solve it easier (that's why I said that this is a great tactic). But you must have no idea on how to massage an expression by only reading the long and trashy text above, right? So, let's look at an example, shall we?
Example. Let A=n=1∑10,000n1. Find ⌊A⌋ without using a calculator. (Even you use a calculator, you also have no enough time to press up to 10,000, except you use Wolfram Alpha :P)
Solution. We can use massaging here since we don't need an exact value, luckily. Through massaging, we can make the sum telescope. We have
n1=2n2<n+n−12=2(n−n−1)
(In this step, we rationalised the denominator)
On the other hand,
n1>n+1+n2=2(n+1−n)
Implying
2(n+1−n)<n1<2(n−n−1)
That looks great, right? We had just bound the deadly term n1 above and below by the terms that will telescope greatly when summed. Therefore,
n=1∑10,0002(n+1−n)=210,001−2 and n=1∑10,0002(n−n−1)=210,000−0
Implying
210,001−2<n1<210,000−0⟹⌊A⌋=198 or 199
Let me ask you something, which is the correct answer, 198 or 199, in your opinion? You can't decide it right? So, let's refine this estimate! When n=1, we get 2(2−1)<1<21. Lower limit is good, but that's too terrible for upper limit! How about we start the summation at n=2? We have
A=1+n=2∑10,000n1
And we estimate
210001−22<n=2∑10,000n1<210,000−21
The lower limit is about 197, while the upper limit is 198. So, A is somewhere at 198 and 199, implying ⌊A⌋=198.