You are on page 1of 29

SEPTEMBER 24, 2012

MANHATTAN WEEKLY
CHALLENGE

1. In the grid above, the variables a through p are each equal to 2, 3, 5, or 7, with exactly
one occurrence of each value in any row and in any column. What is the value of fgjk?
a
e
i
m

b
f
j
n

c
g
k
o

D
H
L
p

(1) bcehilno = 354


(2) (2.252)afkp = dgjm
ANSWER
Although there are 16 different variables in this 4x4 grid, the values of these variables
are severely restricted: they can only be 2, 3, 5, or 7 (you should recognize these as
primes), and only one of each can exist in any row or column. You are asked for the
value of the product fgjk, which is the 2x2 grid at the center.
Statement 1: SUFFICIENT. Note that 354 = 5474, so four of the variables must be 5s,
while the other four must be 7s. You are not allowed to put two of the same values in
any row or column, so a possible arrangement (there are many) looks like this:
5
5
7

7
7
5

7
5
The central grid must be composed of two 2s and two 3s, since each row and each
column must have exactly one of each possible value, and youve already used up the
5s and 7s. Although you cannot be sure which particular variable in the central 2x2 grid
is which value, you know that the product is 36 (2233).
Statement 2: SUFFICIENT. First, use FDP connections: rewrite 2.25 as 9/4, or (3/2)2.
Thus, 2.252 = (3/2)4. Cross-multiplying by 24, you get 34afkp = 24dgjm.
Since the only possible values of the variables are 2, 3, 5, and 7, the only possibilities for
the values of the variables are a = f = k = p = 2, while d = g = j = m = 3.
This tells us that fgjk = 36.
The correct answer is D.

2. For positive integers n, the integer part of the nth term of sequence A equals n, while
the infinite decimal part of the nth term is constructed in order out of the consecutive
positive multiples of n, beginning with 2n. For instance, A1 = 1.2345678..., while A2 =
2.4681012... The sum of the first seven terms of sequence A is between
A. 28 and 29
B. 29 and 30
C. 30 and 31
D. 31 and 32
E. 32 and 33
ANSWER:
First, construct the first seven terms, though only out to a few decimal places, following
the given pattern.
A1 = 1.23...
A2 = 2.46...
A3 = 3.69...
A4 = 4.812...
A5 = 5.10...
A6 = 6.12...
A7 = 7.14...
Now, to add up the first seven terms, you should be strategic about how many decimal
places to keep. You can drop the hundredths place and get a good approximation with
the tenths placeand if you find the sum too close to a boundary between choices,
then you can refine your answer if necessary.
1.2 + 2.4 + 3.6 + 4.8 + 5.1 + 6.1 + 7.1 = 30.3
Including more decimal places would only add a couple of tenths to the sumnot
enough to tip the sum over 31.
The correct answer is C.
3. If x is the reciprocal of a positive integer, then the maximum value of xy, where y = x2,
is achieved when x is the reciprocal of
A. 1
B. 2
C. 3
D. 4
E. 5
ANSWER:

As in so many of these problems, the first task is careful reading. Note that x is not itself
the integer (except in the case of 1/1); x is the reciprocal of an integer. You might even
make a quick table corresponding to the values in the answer choices:
Integer
x
1/1 =
1
1
2
1/2
3
1/3
4
1/4
5
1/5
Now, just keep adding columns. Add one for y, defined as x2:
y=
Integer
x
x2
1/1 =
1
1
1
2
1/2
1/4
3
1/3
1/9
4
1/4
1/16
5
1/5
1/25
Notice that according to PEMDAS, you apply the squaring operation (Exponent) before
the negative sign (Subtraction).
Add one more column for xy. For clarity, well use the caret symbol (^) to indicate
exponents. Notice that the negative sign undoes the reciprocal in x, leaving you with an
integer base.
y=
Integer
x
xy
x2
1/1 =
1
1
11 = 1
1
(1/2)^(1/4) =
2
1/2
1/4
2^(1/4)
(1/3)^(1/9) =
3
1/3
1/9
3^(1/9)
(1/4)^(1/16) =
4
1/4
1/16
4^(1/16)
(1/5)^(1/25) =
5
1/5
1/25
5^(1/25)
Finally, you are left with the task: which of those numbers in the last column is largest?
Remember, fractional exponents are roots, so the second number is the fourth root of
2, and so on.

You can quickly eliminate 1, since the positive roots (to any degree) of any integer
greater than 1 are always greater than 1. For instance, what is the 25 th root of 5? It must
be a number bigger than 1, even if only slightly, because that number times itself 25
times in all produces 5. So you can eliminate A.
Lets compare the second and fourth numbers, because the base of 4 can be rewritten
as 22. The fourth answer becomes then (2^2)^(1/16) = 2^(1/8), which is less than
2^(1/4). The number that solves z8 = 2 is smaller than the one that solves z4 = 2. So the
answer cant be D.
What about B versus C? Raise both numbers to the 36th power, so that we eliminate
fractional exponents.
2^(1/4)^36 = 2^9
3^(1/9)^36 = 3^4
Which of the results is bigger? 2^9 = 512, while 3^4 = only 81. So the original numbers
must be in that same order of size; B is bigger than C. C is out.
A similar argument takes out E (raise both B and E to the 100th power):
2^(1/4)^100 = 2^25
5^(1/25)^100 = 5^4 = 125 = much smaller than 2^25.
The correct answer is B.
4. For any numbers a and b, min(a, b) and max(a, b) represent the minimum and the
maximum of a and b, respectively. If c > a, is a < b < c?
(1) min(max(a, b), c) = max(min(b, c), a)
(2) max(max(a, b), c) min(min(b, c), a) > c a
ANSWER:
The question stem isnt so badit just defines the min and max functions, and then asks
whether b is between a and c, given that c is bigger than a. The statements, however,
are another matter, because the functions are nested. That is, you have mins and
maxs inside each other.
Heres the general principle. Always work your way from the inside out, as you simplify
complicated nested expressions.
Statement 1: NOT SUFFICIENT. Lets take the Yes case first, assuming that b is in fact
between a and c, so that we have a < b < c.
Take the left side first, and work your way from the inside out.
min(max(a, b), c)
= min(b, c) because b is bigger than a, so you can replace max(a, b) with b
= b because b is smaller than c
Right side:
max(min(b, c), a)
= max(b, a) because b is smaller than c, so you can replace min(b, c) with b
= b because b is bigger than a

However, what if the order is b < a < c, so that the answer to the question would be
No?
Left side:
min(max(a, b), c)
= min(a, c) because in this case, a is bigger than b, so you can replace max(a, b) with a
= a because a is smaller than c
Right side:
max(min(b, c), a)
= max(b, a) because b is smaller than c, so you can replace min(b, c) with b
= a because a is bigger than b in this scenario.
So again, the right side equals the left side.
Trying to figure out this statement in the abstract is tough. A better way to deal with it is
to assume an order of the variables (either Yes or No to the question), simplify the
statement and determine whether its true. Normally, you want to assume that the
statement is true and work to the question, but thats hard to do in this situation.
Also notice that any way you slice it, min(max(x, y), z) equals the middle number. So
does max(min(y, z), x), even with any arrangement of variables inside. So statement (1)
is always true, no matter whether b is the middle number, so the statement cant be
sufficient.
Statement 2: SUFFICIENT
The max of the max of the three variables will be the largest variable of the three. The
min of the min of the three variables will be the smallest variable of the three.
So if the largest variable minus the smallest variable is greater than c a, then it cant
be true that c is the largest and that a is the smallest. The variable b must be either
greater than c or smaller than a.
If you need to, think through cases. The Yes answer to the question is a < b < c. So
then the left side of the inequality would reduce to c a, and the statement would be
saying that c a > c a. That is impossible.
So the answer to the question is a definite Nob is NOT between a and c. It must be
either larger than c or smaller than a. A definite No answer to a Yes/No question on
Data Sufficiency is sufficient.
The correct answer is B.
5. If a, b, and c are nonzero integers and z = bc, is az negative?
(1) abc is an odd positive number.
(2) | b + c | < | b | + | c |

ANSWER:
For the expression az to be negative, you need the base a to be negative, and you need
the exponent to be an odd integer (or a fraction that, when fully reduced, has an odd
numerator, such as 1/3 or 5/3, and for that matter an odd denominator; otherwise, you
dont get a real number out).
Since Statement (2) doesnt contain any reference to a, start with this statement (its
easier to deal with).
Statement (2): NOT SUFFICIENT. You dont know whether a is negative or not, so you
can be pretty sure at this point that this statement is not sufficient. However, you have
to be a little careful if this constraint on b and c led to a situation in which az is never
negative, then this statement would be sufficient (answering the question with a no).
This statement is really telling you that b and c have different signs.
If theyre both positive, then | b + c | = | b | + | c |. Likewise, if both b and c are
negative, then the two expressions are equal. (Try numbers to see.) The only way for the
statement to hold true is for b and c to have different signs.
However, simply knowing that b and c have different signs isnt enough, as it turns out.
If you knew that z (= bc) would always be an even integer when b and c have different
signs, then youd have sufficiency (because then azwould never be negative), but thats
not always true.
For instance, if b = -2 and c = 3, then z = -8, but if b = -3 and c = 2, then z = 9.
Statement (1): NOT SUFFICIENT. This statement tells you that a, b, and c are all odd
integers, and either all three are positive or exactly one is positive (and the other two
are negative). If all three are positive, then the expressionaz is positive. But if a is
negative, then you can easily get a negative result: for instance, if a = -1 and z = -27
(fromb = -3 and c = 3), then az = -1.
Statements (1) and (2) together: SUFFICIENT. First, consider the signs of the variables.
You know from (2) that band c have different signs, so one is positive and one is
negative. Since (1) tells you that either all three (a, b, c) are positive or exactly one is
positive, then you know that a has to be negative. Moreover, all three are odd. The two
possibilities at this point are these:
EITHER a neg odd, b pos odd, c neg odd
OR a neg odd, b neg odd, c pos odd
In the Either case, what does z become? A positive odd number raised to a negative
odd power becomes a fraction: 1/(odd). For example, 3-5 = 1/35 = 1/243. So a (which is

negative) raised to that power (of, say, 1/243) will stay negative. (By the way, its legal
to take an odd denominator power of a negative number. You cant raise a negative
number to the power of , because thats taking the square root you dont get a realnumber result but you can raise a negative number to the power of 1/3, 1/5, or
1/243.)
How about the Or case? A negative odd number raised to a positive odd power
becomes a negative odd number. For instance, (-3)5 = -243. So a (which, as before, is
definitely negative) raised to that power (of, say, -243) stays negative, since a negative
number raised to the power of any odd number (positive or negative) is essentially
multiplied by itself an odd number of times, yielding a negative (and then, if the power
is negative, you take the reciprocal, which is also negative).
Either way (Either or Or), the result is negative. The answer to the question is
definitely Yes.
The correct answer is C.
6. In the country of Celebria, the Q-score of a politician is computed from the following
formula:
Q = 41ab2c3/d2, in which the variables a, b, c, and d represent various perceived
attributes of the politician, all of which are measured with positive numbers. Mayor
Flowers Q-score is 150% higher than that of Councilor Plant; moreover, the values of a,
b, and c are 60% higher, 40% higher, and 20% lower, respectively, for Mayor Flower
than for Councilor Plant. By approximately what percent higher or lower than the value
of d for Councilor Plant is the corresponding value for Mayor Flower?
A. 56% higher
B. 25% higher
C. 8% lower
D. 20% lower
E. 36% lower
ANSWER:
The form of the function we are comparing involves only multiplication, division, and
exponentsthere is no addition or subtraction. This is why a particular answer will be
generated from percent changes in the variables (otherwise, youd need to know more
than percent changes).

All the comparisons are to Councilor Plants numbers, so we can pick values for his or
her numbers: say, 1 throughout for Q, a, b, and c. Thus the numbers for Mayor Flower
will be as follows, converting to fractions on the fly (well fully reduce later):
Q = 25/10
a = 16/10
b = 14/10
c = 8/10
Since the problem asks for d, rearrange the formula to isolate d, or d2 (we can take the
square root at the end):
d2 = 41ab2c3/Q
We can ignore the 41, so that the starting value (Plants value) of d would be 1 (all of
Plants other values are 1). Then Flowers value of d2 would be this:
d2 = ab2c3/Q = (16/10)(14/10)2(8/10)3/(25/10)
Turn 16/10 into 8/5. 14/10 squared becomes 196/100, or with a little rounding, 2.
Meanwhile, 8/10 cubed becomes 512/1,000, or with a little more rounding, 0.5. This
rounded 0.5 and the rounded 2 multiply to 1, leaving this:
d2 = ab2c3/Q = (8/5) /(25/10) = 80/125.
Multiply top and bottom by 8 to get 640/1,000, or 64/100. Finally, take the square root
to get d = 8/10, or 0.8. Since d started off at an assumed 1 (for Councilor Plant), we can
see that Mayor Flowers 0.8 is 20% less.
The correct answer is D.
7. Rounded to four decimal places, the square root of the square root of 0.9984 is
approximately
A. 0.9990
B. 0.9992
C. 0.9994
D. 0.9996
E. 0.9998
ANSWER:
First, notice that the number you have to take the first square root of, 0.9984, is just a
little less than 1, meaning that you could represent it as 1 (something small).
Now, its hard to deal with square roots algebraically. But we can deal with their
opposites that is, squares. What would the square of 1 (something small) be? Lets
write that as 1 x, where we know that x is a small number, much less than 1.
(1 x)2 = 1 2x + x2
Now, since x is much less than 1, the x2 term is much much less than 1. (To see why,
imagine that x = 1/1,000. Then x2 = 1/1,000,000.) Since we are rounding in this problem,
we can make an approximation, dropping the x2term:

(1 x)2 = 1 2x + x2 1 2x
Now we have the insight we need. Since the square of 1 x is approximately 1 2x
(doubling the gap between the number and 1) if x is very small, then we can go in the
opposite direction: the square root of 1 2x is approximately 1 x. In other words, you
cut the gap between the number and 1 in half.
Write 0.9984 as 1 0.0016. In this case, 2x = 0.0016, so x = 0.0008.
The square root of 1 0.0016 is approximately 1 0.0008, or 0.9992.
Take the final step. The square root of 1 0.0008 is approximately 1 0.0004, or 0.9996.
You could also get to the answer by working backwards from the answer choices: the
square of the square of the right answer must be approximately 0.9984. It will take
longer, but brute force will get you there, eventually.
The correct answer is D.
8. The symbol $ is defined by the formula a$b = a2 + b2. If 35$x = 372 and ((3$4)$x)$n =
852, then |n| =
A. 76
B. 78
C. 80
D. 82
E. 84
ANSWER:
First, solve the first equation for x by translating the special symbol $ to the standard
algebraic form, using the definition a$b = a2 + b2.
35$x = 352 + x2 = 372
x2 = 372 352
Now use the difference of squares: a2 b2 = (a b)(a + b)
x2 = (37 35)(37 + 35)
= (2)(72) = 144
So x = 12 or 12.
Next, work your way from the inside out in the second equation.
3$4 = 32 + 42 = 9 + 16 = 25
(3$4) = 25 = 5 (remember that anything to the power is square-rooted). You might
notice that you just generated the third side of a 3-4-5 right triangle.
Since x will be squared, it wont matter whether you use 12 or 12; use 12 for simplicity.
(3$4)$x = 5$12 = 52 + 122 = 25 + 144 = 169
((3$4)$x) = 169 = 13
Again, you might have noticed that you just generated the third side of a 5-12-13
triangle, so you could have spared a little work.
Now write the whole equation and solve for n:

((3$4)$x)$n = 13$n = 132 + n2 = 852


n2 = 852 132
Again, use the difference of squares:
n2 = (85 13)(85 + 13)
= (72)(98)
Rather than multiply, try to factor out squares:
n2 = (72)(98) = (2 36)(2 49) = 4 36 49
So n (taking the positive square root) will be 2 6 7 = 84. Since the question asks for
|n|, it doesnt matter whether you choose positive or negative square roots; the result
in the end is 84.
The correct answer is E.
9. In quadrilateral ABCD, sides AB and BC each have length 2, while side CD has length 2.
What is the area of quadrilateral ABCD? (For this problem, quadrilateral means any
closed figure with four straight sides in a plane, with each side touching exactly two
other sides at their endpoints.)
(1) The length of side AD is 2.
(2) The angle between side AB and side BC is 90.
ANSWER:
You are told that quadrilateral ABCD has these side lengths: 2 (= AB), 2 (= BC), and 2
(= CD). You dont know any angles, so imagine that you have hooked together three
poles representing the three sides that you know. The poles are connected by flexible
hinges at B and C:

You dont know how long the fourth side (AD) is, so imagine that the fourth side is a
rubber band that can stretch and shrink as you adjust the angles. ABCD forms a
quadrilateral, defined in this problem as any closed figure with four straight sides in a
plane, with each side touching exactly two other sides. In other words, the sides cant
cross over each other, but you can have an indentation, such as the one shown below
at point B:

So you know, the term quadrilateral is actually defined on the GMAT more narrowly
than it is in typical geometry classes (and in this problem) to include
only convex quadrilaterals, in which every internal angle is less than 180 degrees. So
when the GMAT uses the term quadrilateral, you can ignore concave possibilities such
as the one shown above.
Statement (1): NOT SUFFICIENT. The fourth side (AD) is fixed in length, but you can
collapse or open up the kite to make different areas. In many cases, knowing the 4
sides of a quadrilateral does not determine the area of the quadrilateral (even
restricting yourself to convex quadrilaterals), since you can often change the angles and
therefore the area without changing the sides. Consider that a square and a typical
rhombus can both have the same side lengths, but the rhombus has less area because
you have collapsed the square somewhat.
Statement (2): NOT SUFFICIENT. Fixing the angle between two of the sides still leaves
you freedom to swing the third side (and therefore the fourth side) in various directions,
leading to different areas.
Statements (1) and (2) TOGETHER: STILL NOT SUFFICIENT! Here are the two possible
pictures of the quadrilateral, pictures that satisfy both constraints but that have
different areas:

To be fair, if you used the GMATs normal definition of quadrilateral, the concave
possibility on the right would be outlawed, and C would be the right answer. So if you

picked C, dont feel bad!


That said, does that mean you should never think about concave shapes such as the one
encountered here? Well, the GMAT is very clever, and the writers might avoid using the
term quadrilateral while making you think about an equivalent possibility (e.g., four
line segments intersecting in a plane at exactly four distinct points, end to end,
enclosing a planar figure) that has absolutely no such restriction against concavity.
Always pay close attention to the restrictions as given; if terms are defined or redefined
in the problem, those definitions will be critical.
The correct answer is E.
10. Given that a, b, c, and d all lie between 0 and 1 on the number line, and |a d| >
|a c| > |a b|, does c lie between b and d on the number line?
(1) ab < ad < cd
(2) ac < bc < bd
ANSWER:
From the stem, you know that the four variables are all negative (specifically, between 0
and 1). Interpreting an absolute difference expression such as |a d| as
a distance on the number line (in this case, the distance between a and d), you also
know that d is further from a than c is, and that c is further from a than b is. So, using an
analogy with the solar system, you could say that a is the Sun, b is Mercury (the closest
planet), c is Venus (next closest), and d is the Earth (further away than the other two
are). You dont know which body is wherethere are lots of possibilities for the leftright ordering, including having the Sun (a) be to the left or to the right of all the
planets.
The question you are asked is whether c lies between b and d on the number line. This is
a Yes/No question, of course, so be ready for sufficiency without knowing what the
numbers are.
Statement 1: SUFFICIENT. Split this three-way inequality into two standard inequalities:
ab < ad and ad < cd
Now, you can simplify each inequality by dividing by the common variable. Since you
know that every variable isless than zero, you must flip the direction of the inequality.
ab < ad becomes b > d
ad < cd becomes a > c
Now lay out these two inequalities on two number lines.
1..d..b0
and
1..c...a0

How can you combine these two lines into one? Use the solar system analogy: a = Sun,
then b is the closest planet (on either side of a), then c is further away, then d is the
furthest. There are only two possibilities for the relative ordering:
1..dc..ba0
OR
1..dc..ab0 (as long as c is further away from a than b is)
In either case, c is between b and d on the number line.
Statement 2: SUFFICIENT. Again, split into two standard inequalities:
ac < bc and bc < bd
Divide each inequality by the common variable, flipping the inequality:
a > b and c > d
Lay out the inequalities on two number lines, and combine:
1..d..c0
and
1..b...a0
only combine up one way, given the solar system constraint:
1..dcb....a0
Again, c is between b and d.
The correct answer is D.
11. On January 1, 2010, Dave invests 70% of his retirement savings in Antarctic largecap
stocks, 20% in Antarctic midcaps, and 10% in Antarctic smallcaps. In 2010, largecaps rise
5%, midcaps rise 10%, and smallcaps rise 15% in the Antarctic stock market; however, in
2011, largecaps fall 10% and midcaps fall 20%, while smallcaps rise x% in Antarctica. If,
on January 1, 2012, Dave has the same total amount of retirement savings as he did two
years before, then x is between
A. 10 and 20
B. 20 and 30
C. 30 and 40
D. 40 and 50
E. 50 and 60
ANSWER:
The difficulty of this percent change problem is not so much conceptual as it is
executionalyou want to be able to solve it quickly, easily, and of course accurately.
Lets get to the workout!
Youve got three investments (at various percent allocations) changing by various other
percents over two time periods. The numbers dont look too ugly, but you might suspect

that the result will be hard to compute exactly, because the problem only asks for a
range. Thus, you should be ready to switch to estimation at some point.
Pick a smart number for the total retirement savings Dave starts withsay, $10,000. (If
you pick $100, youll wind up needing to track decimals, so give yourself a couple more
zeros to start with.)
Here are the starting values:
L = $7,000
M = $2,000
S = $1,000
Apply the first years changes, so that you have these numbers on 1/1/2011:
Newer L = $7,000 + 5% = $7,350
Newer M = $2,000 + 10% = $2,200
Newer S = $1,000 + 15% = $1,150
Now apply the second years changes to L and M:
Newest L = $7,350 10% = $7,350 - $735 = $6,615
Newest M = $2,200 20% = $2,200 - $440 = $1,760
Add these to get $8,375. So the newest S must be $10,000 (the target final total of
Daves retirement savings) minus $8,375, or $1,625.
The dollar change in S from 1/1/11 to 1/1/12 is $1,625 - $1,150 = $475. So the question
is this: what percent change does $475 represent, from a starting point of $1,150? Since
$1,150 is a nasty divisor, switch to benchmarks:
10% of $1,150 = $115.
So 20% is just double that, or $230.
And 40% is double that, or $460.
Since $475 is just slightly larger than $460 (but not enough to get you to 50%, which
would be $460 + $115), xmust be between 40 and 50.
Intuitively, it should make sense that youd need a much bigger positive percent change
in the smallest investment (S) to make up for even a moderate downturn in the larger
investments (L and M), so if you were completely stuck for time and needed to guess in
a hurry, you should favor C/D/E over A or B.
The correct answer is D.
12. Jean puts N identical cubes, the sides of which are 1 inch long, inside a rectangular box,
each side of which is longer than 1 inch, such that the box is completely filled with no
gaps and no cubes left over. What is N?
(1) 56 < N < 63
(2) N is a multiple of 3.

ANSWER:
To be able to put N cubes into a rectangular box with no gaps and no left-over cubes,
you must have the following equality: N = length width height. Moreover, if the
length, width, and height are all greater than 1, then it must be true that N is the
product of at least 3 primes. If N is itself prime or the product of just 2 primes (unique or
not), then the condition fails.
So you can rephrase the question this way: is N the product of at least 3 primes?
Statement 1: SUFFICIENT. Since N must be an integer (it counts a set of objects), you can
rephrase the statement: N is either 57, 58, 59, 60, 61, or 62. Now test each of these
numbers.
57 = 3 19 = product of just 2 primes, so 57 is out.
58 = 2 29 = product of just 2 primes, so 58 is out.
59 is itself prime, so 59 is out.
60 = 3 5 = product of 4 primes, so 60 can work. For instance, the dimensions of the
box could be 4 by 3 by 5, or various other combinations of the primes.
61 is prime, so 61 is out.
62 = 2 31 = product of just 2 primes, so 62 is out.
Thus, the statement tells us that N must be 60.
Statement 2: NOT SUFFICIENT. Its easy to come up with multiple examples that fit the
conditions. For instance, we can reuse N = 60 (just be careful whenever you reuse a case
from one statement in another statement), but then try N = 27 (= 3 3 3), which also
fits. Many multiples of 3 can be legal values of N, so this statement is not enough.
The correct answer is A.
13. X is a three-digit positive integer in which each digit is either 1 or 2. Y has the same digits
as X, but in reverse order. What is the remainder when X is divided by 3?
(1) The hundreds digit of XY is 6.
(2) The tens digit of XY is 4.

ANSWER:
You are told that X is a three-digit integer in which each digit is either 1 or 2. There are 8
possibilities for X (8 = 222), but rather than list out this many possibilities, you might
just write something like this: X = abcproduct of a, b, and c, but rather the three-digit
number formed from the digits in that order (a = hundreds, b= tens, c = units). Likewise,
you can now write Y = cba.
You need to find the remainder when X is divided by 3. Since divisibility by 3 depends on
the sum of the digits, you really just need to find a + b + c, or more precisely, whether
this sum is itself a multiple of 3, one more than a multiple of 3, or two more than a
multiple of 3.
Statement 1: SUFFICIENT. To express the hundreds digit of XY in terms of a, b, and c, you
need to write Xand Y in a more formal algebraic way. A three-digit number is 100 times
its hundreds digit, plus 10 times its tens digit, plus its units digit. So X = 100a + 10b + c.
Likewise, Y = 100c + 10b + a. We now have to multiply these two expressions together.
The efficient way to do so is to think of the possible coefficients (10,000; 1,000; 100; 10;
and 1) and then match the terms that will create those coefficients.
(100a+ 10b + c)(100c + 10b + a) =10,000ac + 1,000(ab + bc) + 100(
10(ab + bc) + ac.

)+

So you are told that the hundreds digit is 6. Before you match that to the expression
above, you have to think about carried digitscould the expressions in the units or the
tens place cause a digit to be carried? The answer is no: even if all the digits were 2s
(the maximum), the tens product would only be 8, with no carrying. So we can now say
that 6 =
+
+ . Since each variable can only be 1 or 2, what are the possible
values of the digits? By testing numbers, you can quickly see that exactly one of the
digits must be 2; the other two digits must be 1. Thus, the sum of the digits of X is 2 + 1
+ 1 = 4, so the remainder after division by 3 is 1.
Statement 2: INSUFFICIENT. Using the same work from Statement 1, and checking that
you dont have to worry about carried digits, you get ab + bc = 4. Factor the left side: b(a
+ c) = 4. Given the possible digit values of 1 and 2, there are two possible solutions. One
solution is b = 2 and a + c = 2, or a = 1 and b = 1. The other solution is b = 1 and a + c = 4,
or a = 2 and c = 2. That means that the number is either 121 or 212. If the number is
121, the remainder is 1. If the number is 212, the remainder is 2.
The correct answer is A.

14. A decade is defined as a complete set of consecutive nonnegative integers that have
identical digits in identical places, except for their units digits, with the first decade
consisting of the smallest integers that meet the criteria, the second decade consisting
of the next smallest integers, etc. A decade in which the prime numbers contain the
same set of units digits as do the prime numbers in the second decade is the
A. Fifth
B. Seventh
C. Eighth
D. Ninth
E. Eleventh
ANSWERS:
This problem is all about reading. A complete set of consecutive nonnegative
integers should conjure in your mind the numbers {0, 1, 2, 3, } and youre going to
take a consecutive subset of those numbers. Here are the key words: that have
identical digits in identical places, except for their units digits So the units digits of the
numbers in a decade can differ, but all the other digits are the same. For example, the
integers 50 through 59 would form a decade; 150 through 159 would form a different
decade.
The first decade, we are told, consists of the smallest integers that meet the criteria. The
smallest nonnegative integers are 0, 1, 2, 3, and in fact, the integers 0 through 9 meet
the criteria (they have identical digits in identical places, except for their units digits
and since these numbers only consist of units digits, they have no digits in common, but
theyre still part of the same decade). So the first decade is {0, 1, 2, 3, , 9}. The second
decade is {10, 11, 12, 13, , 19}, and so on.
The prime numbers in the second decade are 11, 13, 17, and 19. So you must find a
decade in which the primes are xxx1, xxx3, xxx7, and xxx9 (where xxx represents the
unknown identical digits in the decade). Search backwards from the answer choices,
noting that since the first decade has no tens digit and thesecond decade has 1 as the
tens digit, the ordinal number (first, second, third, etc.) of the decade is one more
than the tens digit. That is to say, the fifth decade is {40, 41, 49}.
(A) cannot be right, because 49 is not prime.
(B) cannot be right, because 63 and 69 are not prime.
(C) cannot be right, because 77 is not prime.
(D) cannot be right, because 81 is not prime.

Hence, the answer must be (E): 101, 103, 107, and 109 are all prime. If you really want
to check, look for divisibility by primes up to the square root of the number in
questionand since theyre all less than 121, which is
, you can just check
divisibility by 2, 3, 5, and 7. Numbers ending in 1, 3, 7, and 9 are not divisible by either 2
or 5, so you only really have to check 3 and 7.
The correct answer is E.
15. Ax(y) is an operation that adds 1 to y and then multiplies the result by x. If x = 2/3, then
Ax(Ax(Ax(Ax(Ax(x)))))
A. 0 and
B. and 1
C. 1 and 1
D. 1 and 2
E. 2 and 2
ANSWER
Translate the operation from words into math. You add 1 to y, then multiply the result
(y + 1) by x.
So
(y) = (y + 1)x
Now work on the big expression from the inside out:
(x) = (x + 1)x =

+x

(
(x)) = (( + x) + 1)x =
+
+x
At this point, you should spot the pattern. By adding 1 to the result and then multiplying
by x, you add another power of x, so
(
(
(x))) =
+
+
+ x.
Notice that three iterations of the Ax operator give you the fourth power as the highest
power in the sum. Since there are five iterations of the operator, the whole sum must
be
+
+
+
+
+ x.
Now the question is this: if x = 2/3, what does this sum equal, approximately? We dont
need an exact number, because the answer choices express ranges.
Lets go in stages, starting with the smallest power, namely x. Since all the higher
powers yield positive numbers, the sum we want cant be less than 2/3, so we can
already knock out (A) as a possible answer choice.
What about ? If x = 2/3, then
= 4/9. Putting that into the sum, we get 2/3 + 4/9 =
6/9 + 4/9 = 10/9 > 1, so (B) cant be the answer either.
Now to
, which equals 8/27. Putting that into the sum, we get 10/9 + 8/27 = 30/27 +
8/27 = 38/27.

Keep going. Since


= 16/81, we get 38/27 + 16/81 = 114/81 + 16/81 = 130/81. Were
past the 1 point, we can see, because 130 is more than half bigger than 81 (half of
81 is about 40, so half bigger would be about 121). So (C) is out, and were between
(D) and (E).
How about the last two powers can we estimate, to avoid taking larger & larger
powers in the denominator, then finding a common denominator, adding, etc.? (Notice
the switch in tactics here.) Lets see we only have to be worried about crossing up past
2, which is 162/81. Thats 32/81 more than where we are now, or somewhere between
1/3 and 1/2, as we can quickly estimate (32/96 would be 1/3, so 32/81 with its smaller
denominator must be bigger than 1/3. On the other hand, 32/64 would be 1/2, so 32/81
with its bigger denominator must be smaller than 1/2).
= 32/243, or a little less than 0.15 (estimating exactly 0.1 would be 24.3/243, and
exactly 0.15 would be 36.something/243). Thus
(which is 2/3 of the previous
number) would be 2/3 of a little less than 0.15, or a little less than 0.1. Adding, you get
the leftover powers to add up to less than 0.25, so you definitely dont get all the way to
2. The sum we want is between 1 and 2.
The correct answer is D .
16. The 4 sticks in a complete bag of Pick-Up Sticks are all straight-line segments of
negligible width, but each has a different length: 1 inch, 2 inches, 3 inches, and 4 inches,
respectively. If Tommy picks a stick at random from each of 3 different complete bags of
Pick-Up Sticks, what is the probability that Tommy CANNOT form a triangle from the 3
sticks?
A. 11/32
B. 13/32
C. 15/32
D. 17/32
E. 19/32
ANSWER:
First, make sure that you understand the problem. Essentially, Tommy picks three line
segments at random. Each of the line segments could be 1, 2, 3, or 4 inches long. Then
he is going to try to form a triangle. Some of the time, evidently, he will not be able to
do so. The question is this: what is the probability that he cannot form a triangle from
the three segments?
Recall, from your knowledge of geometry, the so-called Triangle Inequality: in any
triangle, each side length must be less than the sum of the other two side lengths. This
is simply another way of saying that the shortest path between X and Y is a straight line.
If you have a triangle linking points X, Y, and Z, then the shortest way to get
from X to Y is to go straight there, rather than take the detour through Z. You can also

express the Triangle Inequality this way: each side length must be more than the
absolute difference of the other two side lengths.
Since there arent tons of options for the side lengths, lets go ahead and start
constructing cases that would fail the test.
1-1-2: These three lengths would not form a triangle, because the third side (2) should
be less than the sum of the other two sides (1 + 1). Now we can count the
rearrangements: there are 3 ways to rearrange 1-1-2 (in other words, Tommy could pick
the 2-side first, second, or third). You can do this count manually (1-1-2, 1-2-1, or 2-1-1),
or you can divide 3! by 2! (the repeats) to get 3 options.
1-1-3: Another 3 options that fail the test.
1-1-4: Another 3 options.
1-2-4: Another 6 options, because you can rearrange 3 distinct sides in 6 (= 3!) different
ways.
1-3-4: Another 6 options.
2-2-4: Another 3 options.
These are all the possibilities for triples that dont form triangles (make sure you dont
double-count). Adding up all the options, you get 3 + 3 + 3 + 6 + 6 + 3 = 30.
Finally, you have to divide by all the possible outcomes. Tommy has 4 outcomes in each
bag, and he picks from 3 different bags. So he has 4 4 4 = 64 possible outcomes.
30/64 = 15/32.
The correct answer is C.
17. Z is the set of the first n positive odd numbers, where n is a positive integer. Given that n
> k, where k is also a positive integer, x is the maximum value of the sum of k distinct
members of Z, and y is the minimum value of the sum of k distinct members of Z, what
is x + y?
A. kn
B. kn + k2
C. kn + 2k2
D. 2kn k2
E. 2kn
ANSWER:
Get concrete right away. Say n = 5. Then Z = {1, 3, 5, 7, 9}. Since n > k, pick a k thats less
than nsay, k = 3.

x = max value of the sum of k distinct members of Z. Thus, you must pick the 3 largest
values (5, 7, and 9). Their sum is 21, so x = 21.
y = min value of the same sort of sum. So you must pick the 3 smallest values (1, 3, and
5). Their sum is 9, so y = 9.
x + y = 21 + 9 = 30. Now, given that the target number is 30, while n = 5 and k = 3, check
the answer choices.
(A) kn = 3 5 = 15
(B) kn +

=35+

(C) kn + 2

= 15 + 9 = 24

=35+2

(D) 2kn
=235
(E) 2kn = 2 3 5 = 30

= 15 + 18 = 33
= 30 9 = 21

Only (E) fits.


The proof that (E) always works is substantially harder on a conceptual level. The sum of
the smallest kpositive odds equals

, while the sum of the k largest odds in this set

(where the largest odd is 2n 1) winds up being 2kn


2kn.

. The

terms cancel, leaving

However, you should recognize that the plug-numbers approach is faster and easier.
The correct answer is E.
18. The expression x[n]y is defined for positive values of x and y and for positive integer
values of n as follows:
x[1]y =

If n is odd, x[n+1]y =

If n is even, x[n+1]y =
If y = and x[4]y = 2, then x =
A.
B.
C. 1
D. 2
E. 4

ANSWER:
Construct x[4]y in stages, working up from n = 1:
x[1]y =
To figure out x[2]y, apply the second part of the definition, using n = 1 (odd):
x[2]y =
Be careful to apply the right part of the definition! Track carefully where n + 1 is and
where n is.
Now keep going, using the third part of the definition (with n = 2, even):

x[3]y =
(Well avoid even tinier exponents by simply writing xx for

.)

Finally, the x[4]y expression will add another y to the exponent:

x[4]y =
Now, we are told that y is and that this crazy expression

in y = , you get

equals 2. Substituting

You should recognize at this point that its unlikely to be most efficient to try to solve
directly for xin fact, it may not even be possible within GMAT math. Rather, you
should glance through the answer choices and see whether you can backsolve,
eliminating obviously wrong answers.
(A) x = produces an expression in which youre raising (a proper positive fraction) to
a fractional power (i.e., taking a root). This will result in a number closer to 1 than is,
but it cant be larger than 1, and its certainly not 2.
(B) Same problem as in (A).

(C) x = 1 produces 1^ = 1, not 2.

(D) x = 2 produces exactly what we want:

. This is the answer.

We can stop there, but we can check 4 quickly: the expression becomes
2.

, which is not

The correct answer is D.


19. From 2008 to 2009, the number of Easter eggs in a certain egg hunt rose 20%, then fell
17% from 2009 to 2010. From 2008 to 2009, the ratio of Easter eggs to Easter bunnies in
the same hunt fell 20%, then rose 22% from 2009 to 2010. By what approximate percent
did the number of bunnies change over the full period from 2008 to 2010?
A. 5% increase
B. 2% increase
C. 1% increase
D. 3% decrease
E. 6% decrease
ANSWER:
This problem will take some efficient calculation, knowledge of FDP connections &
benchmarks, and smart use of Smart Numbers.
Lets start by picking the number of Easter eggs in 2008 as 100. That number rose 20%
over the next year, so in 2009 it was 120. Finally, that number fell 17% in the second
year. You can compute this effect exactly by multiplying 120 by 0.83 to get 99.6 (not too
hard), or you can recognize that 17% is very slightly more than 16.666%, which is 1/6.
For something to fall by that exact percent (16 and 2/3%) is the same as multiplying it by
5/6. Notice also that increasing something by 20% (1/5) is the same as multiplying it by
6/5. These two effects would perfectly undo each other:
100 + 20% = 120
120 16.66% = 100.
Since you are taking away 17%, you must wind up under 100, but only by a very little bit
say, 0.5, so you could estimate 99.5 as your outcome (notice how close this is to the
right answer). On top of that, be sure to recognize how the percent decrease (-17%) is
smaller numerically, but its off of the larger base, so it actually has more impact than
the percent increase of 20%.

Now lets deal with the E/B ratio. Pick that as 1 to start (so that our number of bunnies
starts at 100 as well). Take away 20%, so you have 0.8 as the new ratio. (No need to
figure out the intermediate number of bunnies.) Finally, add back 22%. 20% of 0.8
would be 0.16, and 2% would be 0.016, so just add both of those to get 0.8 + 0.16 +
0.016 = 0.976.
Almost done! If E in 2010 is now just under 100 (= 99.6), and E/B fell from 1 to 0.976,
whats the new B? E (E/B) = B, so you have this:
99.6 / 0.976 = new B
Now you can do the long division, or you can estimate. This result will definitely be
bigger than 100, so we can rule out all the decreases. (The fraction would have to be
97.6 / 0.076 or 99.6 / 0.996 to get 100.)
Now, is the increase approximately 2% or 5%? Well, if x is a very small positive number,
then this relationship holds: 1/(1 x) can be approximated by 1 + x. (It turns out that
1/(1 x) is always bigger, but only by a slight amount when x is small.
So lets approximate 99.6 as 100 and 0.976 as 0.98:
99.6 / 0.976 100/0.98 = 100(1/0.98). Now pretend that x = 0.02. Approximate:
1/(1 0.02) 1 + 0.02 = 1.02
Now, substituting back in, the result is 102. (To more decimal places, you get 102.0492
only just slightly bigger.) This corresponds to a 2% increase.
The correct answer is B.
20. For positive integer m, the m-th heptagonal number is given by the formula (5

3m)/2. For positive integer n, the n-th triangular number is the sum of the first n positive
integers. Which of the following is true for k, the smallest triangular number that is also
heptagonal?
A. 33 k 40
B. 41 k 48
C. 49 k 56
D. 57 k 64
E. 65 k 72
ANSWER:
This problem isnt conceptually too difficult it just takes quick, accurate computation.
First, work out the first several heptagonal numbers.

If m = 1, then the heptagonal number is (5

31)/2 = (5 3)/2 = 1.

If m = 2, then the heptagonal number is (5

32)/2 = (20 6)/2 = 14/2 = 7.

If m = 3, then the heptagonal number is (5

33)/2 = (45 9)/2 = 36/2 = 18.

If m = 4, then the heptagonal number is (5

34)/2 = (80 12)/2 = 68/2 = 34.

If m = 5, then the heptagonal number is (5

35)/2 = (125 15)/2 = 110/2 = 55.

If m = 6, then the heptagonal number is (5

36)/2 = (180 18)/2 = 162/2 = 81.

We can stop here, because were above the range in the answer choices. In fact, the
target number must be 34 or 55, so we could narrow down to A or C now.
Next, which of these two numbers is triangular: the sum of the first n positive numbers?
Work these triangular numbers out:
1
1+2=3
1+2+3=6
plus 4 gives 10
plus 5 gives 15
plus 6 gives 21
plus 7 gives 28
plus 8 gives 36 we can stop here, because weve skipped over 34. The answer must be
55 (as you can show by adding 9 and 10).
The correct answer is C.
21. The positive value of x that satisfies the equation (1+2x)5 = (1+3x)4 is between
A. 0 and 0.5
B. 0.5 and 1
C. 1 and 1.5

D. 1.5 and 2
E. 2 and 2.5
ANSWER:
The two expressions in question,
and
, could in theory be
expanded, but youll wind up doing a lot of algebra, only to find an equation
involving
, ,
, , and x. Solving for the positive value of x that makes the
equation true is nearly impossible.
So how on Earth can you answer the question? Notice that the question does not
require you to find a precise value of x; you just need a range. So plug in good
benchmarks and track the value of each side of the equation.
Start with x = 1. The equation becomes
Compute the two sides:

= 243, while

=?=
=

.
= 256. So the right side is bigger.

Now, we need another benchmark. Try x = 2. The equation becomes


Compute or estimate the two sides.

=?=

= 625, and multiplying in another 5 gives

you something larger than 3,000 (3,125, to be precise). Meanwhile, 74 =


<
=
2,500, so the left side is now bigger. Somewhere between x = 1 and x = 2, then, the
equation must be true.

The only benchmark left to try is 1.5, or 3/2. Plugging in, you get
=

=?=

= 1,024 (its good to know your powers of 2 this high)

For the other side, first compute the denominator:


= 16. Now the numerator:
1211111. 12111 = 1,331 (this is quick to do longhand)
1,33111 = 14,641 (also quick to do longhand)

So

= 14,641/16 < 1,000. The right side is bigger.

Lets recap:
=

at what value of x?

<

when x = 1

>

when x = 1.5

>

when x = 2

So the value at which the two sides are equal must be between 1 and 1.5.
The correct answer is C.
Extra points:
As x grows past 2, the larger power (5) on the left takes over, so you can see that the left
side will always be bigger.
What about when x is between 0 and 1? Well, first notice that at x = 0, the two sides are
again equal. If x is a tiny positive number (say, 0.001 or something), then you can ignore
higher powers of x (

, etc.), and this simplifies the algebraic expansion:

For tiny positive x,


= 1 + 5(2x) + higher powers of x 1 + 10x
= 1 + 4(3x) + higher powers of x 1 + 12x
So right away, the right side is bigger than the left side. You can also check x = :
=

= 32

> 36 (=

) > 32

Again, the right side is bigger than the left side. For every x between 0 and 1, in fact, the
right side is larger than the left side. This fact isnt particularly easy to prove, but you
dont need to do so.

You might also like