1987 Specimen STEP solutions thread
Watch
Found this paper the other day and realised there wasn't a solutions thread up and running so I figured it should be done. Unfortunately, I've only been able to locate the STEP III paper so if anyone could locate the others it would be awesome.
EDIT: Got them. Courtesy of Professor Siklos himself!
STEP I
1.
2.Solution by bensmith
3.Solution by bensmith
4.Solution by bensmith
5.
6.Solution by bensmith
7.Solution by Farhan.Hanif93
8.
9.
10.
11.Solution by Farhan.Hanif93
12.Solution by Farhan.Hanif93
13.Solution by Farhan.Hanif93
14.
15.
16.
STEP II

STEP III
1.Solution by bensmith.
2.Solution by DFranklin
3.Solution by bensmith
4.Solution by bensmith
5.
6.Solution by Farhan.Hanif93
7.Solution by Farhan.Hanif93
8.Solution by DFranklin
9.Solution by Farhan.Hanif93
10.Solution by Kennethx
11.solution by bensmith
12.Solution by bensmith
13.Solution by bensmith
14.Solution by bensmith
15.Solution by DFranklin
16.Solution by Dfranklin
EDIT: Got them. Courtesy of Professor Siklos himself!
STEP I
1.
2.Solution by bensmith
3.Solution by bensmith
4.Solution by bensmith
5.
6.Solution by bensmith
7.Solution by Farhan.Hanif93
8.
9.
10.
11.Solution by Farhan.Hanif93
12.Solution by Farhan.Hanif93
13.Solution by Farhan.Hanif93
14.
15.
16.
STEP II

STEP III
1.Solution by bensmith.
2.Solution by DFranklin
3.Solution by bensmith
4.Solution by bensmith
5.
6.Solution by Farhan.Hanif93
7.Solution by Farhan.Hanif93
8.Solution by DFranklin
9.Solution by Farhan.Hanif93
10.Solution by Kennethx
11.solution by bensmith
12.Solution by bensmith
13.Solution by bensmith
14.Solution by bensmith
15.Solution by DFranklin
16.Solution by Dfranklin
0
reply
STEP III Q1
a)let denote the product up to n.
By inspection, these numbers seem to be of the form so let's guess that.
We have already proved this to be the case for n=2 so let us presume, for the purpose of induction, that . Now let n=k+1:
So our guess is thus true by mathematical induction.
b) Consider the binomial expansion of :
. Letting x=1 gives us:
Notice that the i+1 th element in the first summation cancels with the ith element in the second summation leaving the first element in the first summation so:
As required.
a)let denote the product up to n.
By inspection, these numbers seem to be of the form so let's guess that.
We have already proved this to be the case for n=2 so let us presume, for the purpose of induction, that . Now let n=k+1:
So our guess is thus true by mathematical induction.
b) Consider the binomial expansion of :
. Letting x=1 gives us:
Notice that the i+1 th element in the first summation cancels with the ith element in the second summation leaving the first element in the first summation so:
As required.
0
reply
Report
#3
At a brief glance, the specimen questions are quite a lot easier than the actual exam questions were that year... {Tut tut}.
0
reply
(Original post by DFranklin)
At a brief glance, the specimen questions are quite a lot easier than the actual exam questions were that year... {Tut tut}.
At a brief glance, the specimen questions are quite a lot easier than the actual exam questions were that year... {Tut tut}.
Shall I continue uploading solutions?
1
reply
Report
#5
(Original post by bensmith)
Aren't all STEP questions relatively easy for you though?
Aren't all STEP questions relatively easy for you though?
Shall I continue uploading solutions?
0
reply
Report
#6
Question 10 was quite easy, managing attachments is alas a bit more troublesome. Enjoy reading down to up instead of left to right.
0
reply
Report
#9
STEP III, Q15: It's easy to convince yourself that the villages all remain connected if less than 3 roads are down (draw diagrams if you wish to illustrate the possible cases).
If 3 roads are down and they all go to the same village, that village is cut off. Otherwise, the villages remain connected. Again, draw diagrams if you wish.
If 4 roads are down, then it's clearly impossible for 2 roads to connect all 4 villages. (The roads must connect at a village, and that only leaves 2 ends to connect 3 villages).
Even more obvious for 5 and 6 roads.
P(more than 3 roads down) = (binomial)
P(A cut off by having exactly 3 roads down) = . So P(any village cut off by exactly 3 roads down) =
So P(villages are cut off) =
So P(villages are connected) =
When p = 1/2, this equals
Wow  only 2 probablity questions out of 16. Hard times for probability specialists...
If 3 roads are down and they all go to the same village, that village is cut off. Otherwise, the villages remain connected. Again, draw diagrams if you wish.
If 4 roads are down, then it's clearly impossible for 2 roads to connect all 4 villages. (The roads must connect at a village, and that only leaves 2 ends to connect 3 villages).
Even more obvious for 5 and 6 roads.
P(more than 3 roads down) = (binomial)
P(A cut off by having exactly 3 roads down) = . So P(any village cut off by exactly 3 roads down) =
So P(villages are cut off) =
So P(villages are connected) =
When p = 1/2, this equals
Wow  only 2 probablity questions out of 16. Hard times for probability specialists...
0
reply
STEP III Q3
Consider the complex numbers which forms a triangle of side . By the triangle inequality, so the given inequality is just a restatement of the triangle inequality.
. Taking the modulus of both sides we get:
.
We can maximise the RHS summation by letting so:
which must hold if there are to be solutions.
notice that this does not hold for z=1/3 and therefore does not hold for values less than a third (this is obvious as the sum is just a power series with positive coefficients and so must have a +ve derivative). Furthermore since we chose a such that the sum was maximised, the result must hold for smaller values of it. Therefore there must be no solutions for z less than or equal to a third.
Consider the complex numbers which forms a triangle of side . By the triangle inequality, so the given inequality is just a restatement of the triangle inequality.
. Taking the modulus of both sides we get:
.
We can maximise the RHS summation by letting so:
which must hold if there are to be solutions.
notice that this does not hold for z=1/3 and therefore does not hold for values less than a third (this is obvious as the sum is just a power series with positive coefficients and so must have a +ve derivative). Furthermore since we chose a such that the sum was maximised, the result must hold for smaller values of it. Therefore there must be no solutions for z less than or equal to a third.
0
reply
(Original post by DFranklin)
STEP III, Q15: It's easy to convince yourself that the villages all remain connected if less than 3 roads are down (draw diagrams if you wish to illustrate the possible cases).
If 3 roads are down and they all go to the same village, that village is cut off. Otherwise, the villages remain connected. Again, draw diagrams if you wish.
If 4 roads are down, then it's clearly impossible for 2 roads to connect all 4 villages. (The roads must connect at a village, and that only leaves 2 ends to connect 3 villages).
Even more obvious for 5 and 6 roads.
P(more than 3 roads down) = (binomial)
P(A cut off by having exactly 3 roads down) = . So P(any village cut off by exactly 3 roads down) =
So P(villages are cut off) =
So P(villages are connected) =
When p = 1/2, this equals
Wow  only 2 probablity questions out of 16. Hard times for probability specialists...
STEP III, Q15: It's easy to convince yourself that the villages all remain connected if less than 3 roads are down (draw diagrams if you wish to illustrate the possible cases).
If 3 roads are down and they all go to the same village, that village is cut off. Otherwise, the villages remain connected. Again, draw diagrams if you wish.
If 4 roads are down, then it's clearly impossible for 2 roads to connect all 4 villages. (The roads must connect at a village, and that only leaves 2 ends to connect 3 villages).
Even more obvious for 5 and 6 roads.
P(more than 3 roads down) = (binomial)
P(A cut off by having exactly 3 roads down) = . So P(any village cut off by exactly 3 roads down) =
So P(villages are cut off) =
So P(villages are connected) =
When p = 1/2, this equals
Wow  only 2 probablity questions out of 16. Hard times for probability specialists...
0
reply
Report
#12
STEP III, Q16:
(i) If is the number of extra children needed for the ith male child (that is, the number of children since the i1th male child), then C = X_1 + ... + X_r, and so E[C] = E[X_1]+...+E[X_r]. The X_i are iid geometric distributions with p = 1/2, so E[X_i] = 2 and C = 2r.
(ii) Suppose the king ignores the rules and just has 2r1 children. If he has r or more boys, clearly C < 2r, and conversely, if he has < r boys, C>=2r.
So P(C < 2r) = P(there are at least r boys from 2r1 children). Since boys and girls are equally likely, P(there are at least r girls from 2r1 children) = P(there are at least r boys from 2r1 children).
But these two events are mutually exclusive, and 1 of them must occur. So P(there are at least r girls from 2r1 children) = 1  P(there are at least r boys from 2r1 children).
So P(there are at least r boys from 2r1 children) = 1  P(there are at least r boys from 2r1 children), and so P(there are at least r boys from 2r1 children) = 1/2. That is, P(C < 2r) = 1/2.
(i) If is the number of extra children needed for the ith male child (that is, the number of children since the i1th male child), then C = X_1 + ... + X_r, and so E[C] = E[X_1]+...+E[X_r]. The X_i are iid geometric distributions with p = 1/2, so E[X_i] = 2 and C = 2r.
(ii) Suppose the king ignores the rules and just has 2r1 children. If he has r or more boys, clearly C < 2r, and conversely, if he has < r boys, C>=2r.
So P(C < 2r) = P(there are at least r boys from 2r1 children). Since boys and girls are equally likely, P(there are at least r girls from 2r1 children) = P(there are at least r boys from 2r1 children).
But these two events are mutually exclusive, and 1 of them must occur. So P(there are at least r girls from 2r1 children) = 1  P(there are at least r boys from 2r1 children).
So P(there are at least r boys from 2r1 children) = 1  P(there are at least r boys from 2r1 children), and so P(there are at least r boys from 2r1 children) = 1/2. That is, P(C < 2r) = 1/2.
0
reply
Report
#13
STEP III, Q8.
(i) I'm going to be lazy and use capitals to denote vectors. Let the vertices of the tetrahedron be A, B, C, D. Sum of squares of edges = AB^2+AC^2+AD^2+BC^2+BD^2+CD^2.
Sum of squares of midpoints = [(A+BCD)^2 + (A+CBD)^2 + (A+DBC)^2] / 4.
So 4 x Sum of squares of midpoints = (A+BCD)^2 + (A+CBD)^2 + (A+DBC)^2 = ((AD)+(BC))^2 + ((AB)+(CD))^2 +((AC)+(DB))^2
= (AD)^2+(BC)^2 + 2(AD).(BC)+(AB)^2+(CD)^2+2(AB).(CD)+(AC)^2+(DB)^2 + (AC).(DB).
So, sufficient to show (AD).(BC) + (AB).(CD) + (AC).(DB) = 0. Some tedious expansion gives us A.BA.C+C.DB.D+A.CA.D+B.DB.C+A.DA.B+B.CC.D and everything cancels.
(ii) Take a = 3I + 2J + 6K, b = xI + yJ + K.
Then we have .
a.b = cos t a b, where cos t is the angle between the vectors.
So a.b = a b if and only if cos t = 1.
So we require 3I+2J+6K and xI+yJ+1 to be in the same direction. So x = 3/6, y=2/6, or x = 1/2, y=1/3.
(i) I'm going to be lazy and use capitals to denote vectors. Let the vertices of the tetrahedron be A, B, C, D. Sum of squares of edges = AB^2+AC^2+AD^2+BC^2+BD^2+CD^2.
Sum of squares of midpoints = [(A+BCD)^2 + (A+CBD)^2 + (A+DBC)^2] / 4.
So 4 x Sum of squares of midpoints = (A+BCD)^2 + (A+CBD)^2 + (A+DBC)^2 = ((AD)+(BC))^2 + ((AB)+(CD))^2 +((AC)+(DB))^2
= (AD)^2+(BC)^2 + 2(AD).(BC)+(AB)^2+(CD)^2+2(AB).(CD)+(AC)^2+(DB)^2 + (AC).(DB).
So, sufficient to show (AD).(BC) + (AB).(CD) + (AC).(DB) = 0. Some tedious expansion gives us A.BA.C+C.DB.D+A.CA.D+B.DB.C+A.DA.B+B.CC.D and everything cancels.
(ii) Take a = 3I + 2J + 6K, b = xI + yJ + K.
Then we have .
a.b = cos t a b, where cos t is the angle between the vectors.
So a.b = a b if and only if cos t = 1.
So we require 3I+2J+6K and xI+yJ+1 to be in the same direction. So x = 3/6, y=2/6, or x = 1/2, y=1/3.
0
reply
STEP III Q14
v is the velocity of the particle after the collision.
Conservation of KE:
I is the moment of inertia of the rod.
Conservation of angular momentum:
As required.
The rod forms a circle when it rotates. The particle and the rod can only collide when the particle is within it. By pythagoras, the particle will travel before leaving the circle and so the time it takes is:
The time for the rod rotate through radians such that the tip touches the point where the particle leaves the circle is
Note that which can easily be seen if we overestimate pi to 3.2 and underestimate root 3 to 1.5 where the inequality holds. If it holds for those values then it certainly must hold for their actual values. Therefore the tip of the rod reaches the leaving point before the particle so they must collide again.
v is the velocity of the particle after the collision.
Conservation of KE:
I is the moment of inertia of the rod.
Conservation of angular momentum:
As required.
The rod forms a circle when it rotates. The particle and the rod can only collide when the particle is within it. By pythagoras, the particle will travel before leaving the circle and so the time it takes is:
The time for the rod rotate through radians such that the tip touches the point where the particle leaves the circle is
Note that which can easily be seen if we overestimate pi to 3.2 and underestimate root 3 to 1.5 where the inequality holds. If it holds for those values then it certainly must hold for their actual values. Therefore the tip of the rod reaches the leaving point before the particle so they must collide again.
0
reply
STEP III Q11
as required.
For the second plough:
z=k(time it takes for 2nd plough to go y metrestime taken for first plough to reach y)
=
And since
As required.
Differentiating this with respect to time we get:
Which means it is a solution as it satisfies the equation.
They will collide when their times are equal and x=y. Notice how the time for second plough is the same as the one for the 1st
This should disappear when they collide so:
As required.
as required.
For the second plough:
z=k(time it takes for 2nd plough to go y metrestime taken for first plough to reach y)
=
And since
As required.
Differentiating this with respect to time we get:
Which means it is a solution as it satisfies the equation.
They will collide when their times are equal and x=y. Notice how the time for second plough is the same as the one for the 1st
This should disappear when they collide so:
As required.
0
reply
STEP III Q4
The parametric form of a parabola is .
So the given line touches the paabola, as required, at the point .
Lets assume the tangents intersect at the point (x_1,y_1) which would mean:
So the equation of the tangent that goes through (x_1,y_1) is:
Substituing in x=a and subtracting the y values we get:
As required.
Note: at the end I removed the suffixes, they were only there in the first place for notational purposes. They are just generalised coordinates.
The parametric form of a parabola is .
So the given line touches the paabola, as required, at the point .
Lets assume the tangents intersect at the point (x_1,y_1) which would mean:
So the equation of the tangent that goes through (x_1,y_1) is:
Substituing in x=a and subtracting the y values we get:
As required.
Note: at the end I removed the suffixes, they were only there in the first place for notational purposes. They are just generalised coordinates.
0
reply
Report
#18
STEP III, Q2. This is completely standard once you've done an analysis course...
(1+x)^n = 1 + nx + n(n1)/2 x^2 + ... + x^n. All terms are +ve, so discarding all but the first two terms on the RHS we have (1+x)^n >= 1+nx > nx. Now suppose y > 1, write y = 1 + x, where x > 0. y^n > nx and so if n > K/x, y^n > K. So we can take N = K/x = K/(y1).
Similarly, for n >=2, by discarding all but the n(n1)/2 term we have (1+x)^n >= n(n1)/2 x^2. If y > 1, again write y = 1+x, then y^n > n(n1)/2 x^2 and so y^n / n > (n1)/2 x^2. If n > (2K/x^2) + 1, then (n1)/2 > K / x^2 and so (n1)/2 x^2 > K. So we can take N = (2K/(y1)^2) + 1.
(1+x)^n = 1 + nx + n(n1)/2 x^2 + ... + x^n. All terms are +ve, so discarding all but the first two terms on the RHS we have (1+x)^n >= 1+nx > nx. Now suppose y > 1, write y = 1 + x, where x > 0. y^n > nx and so if n > K/x, y^n > K. So we can take N = K/x = K/(y1).
Similarly, for n >=2, by discarding all but the n(n1)/2 term we have (1+x)^n >= n(n1)/2 x^2. If y > 1, again write y = 1+x, then y^n > n(n1)/2 x^2 and so y^n / n > (n1)/2 x^2. If n > (2K/x^2) + 1, then (n1)/2 > K / x^2 and so (n1)/2 x^2 > K. So we can take N = (2K/(y1)^2) + 1.
1
reply
Q12
A, B and C form a triangle. Let the angle at A be theta and the angle a C be phi. Let us further denote D to be the point on CB that also lies on the line perpendicular to CB and goes through A and T is the tension in the string.
The rod is in equilibrium so the forces vertically and the moments about A are balanced:
To find AD, notice that the are of triangle is and also .
Equating the two, we get:
Substituting in from our first equations:
By the sine rule:
So:
The function is greater that or equal to 0 and less than or equal to 1 but, in this case, we don't want the equality case as that would mean A, B and C would be collinear so:
To find the tension:
So:
As required.
A, B and C form a triangle. Let the angle at A be theta and the angle a C be phi. Let us further denote D to be the point on CB that also lies on the line perpendicular to CB and goes through A and T is the tension in the string.
The rod is in equilibrium so the forces vertically and the moments about A are balanced:
To find AD, notice that the are of triangle is and also .
Equating the two, we get:
Substituting in from our first equations:
By the sine rule:
So:
The function is greater that or equal to 0 and less than or equal to 1 but, in this case, we don't want the equality case as that would mean A, B and C would be collinear so:
To find the tension:
So:
As required.
0
reply
STEP III Q13
By conservation of energy:
As required
When the ring has rotated 2theta radians, the particle has moved horizontally relative to the ring but the ring itself has moved so:
and similarly for y
And by our first result:
as required.
Differentiating w.r.t time:
Similarly:
As required.
At theta=pi/4, from our previous results, we can see that the acceleration vertically is g which would imply the reaction force is 0 but at that point the hoop ha left the table so it can't roll past theta=pi/4.
By conservation of energy:
As required
When the ring has rotated 2theta radians, the particle has moved horizontally relative to the ring but the ring itself has moved so:
and similarly for y
And by our first result:
as required.
Differentiating w.r.t time:
Similarly:
As required.
At theta=pi/4, from our previous results, we can see that the acceleration vertically is g which would imply the reaction force is 0 but at that point the hoop ha left the table so it can't roll past theta=pi/4.
0
reply
X
Quick Reply
Back
to top
to top