Which of the following figures has the greatest area shaded in red?
Assumptions:
Note: If you get it wrong, don't worry! An Italian mathematician named Malfatti also got it wrong in the 1 9 th century.
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
How do you find side lengths of triangle in terms of r?
Triangle C has r= 1 2 3 . Area is calculated correctly for this r.
Log in to reply
Thanks. I worked this on paper and entered the wrong number here.
Can anyone share how this mathematician got it wrong?
Log in to reply
Log in to reply
Nice. This is a special case of a more general problem. Malfatti assumed a construction of 3 circles like A, with each circle tangent to two sides and both other circles. The funny thing is this is never the best. It's always better to do something like B where the first circle is as large as possible then make the second as large as possible. Proved by Zalgaller and Los' (1994)
How did you find the values of r sqrt(3) in A, r sqrt(3)/6 in B, and 2r*sqrt(3) in C?
Log in to reply
There are 30-60-90 right triangles everywhere. I just used the relevant ratios.
Law of sines
Figures A and B relate to a very famous conjecture that turned out to be 100% in error - not just provably wrong for some triangles but ultimately proved to be wrong for all triangles, so I knew that B had a larger red shaded area than A, and the large amounts of blue space visible in C made it fairly obvious that that had the smallest shaded area, hence the answer had to be B.
Log in to reply
What was the conjecture?
Log in to reply
That the maximum area of 3 circles in a triangle would come from each being tangent to two sides of the triangle and the other two circles.
In truth, the maximum area occurs when the first circle is as large as possible. The the second is the largest possible in the spaces remaining. Then the same for the third. (The greedy algorithm.)
How did you know that the smaller triangle was one-third the height of the original one in figure B?
Log in to reply
The ratio of the distance from the center of mass to the vertex to the midpoint of the opposite side is 2:1
Log in to reply
sorry i got this wrong XD
Log in to reply
@Zhu Xiaoye – Actually, this is sort of right. A 2:1 ratio implies r is one-third of the height. Take away 2r from the height and you're left with one-third of the height.
(You can also split the big triangle into 9 congruent ones.)
Only a side remark to the approximation of 3 : 1 5 2 6 has the smallest denominator better than 8 1 1 4 0 , and 5 6 9 7 is the best approximation with a denominator smaller than 81 :-)
I did this in exactly the same way and so got the right answer. But I cannot say that I was so confident that I had got it right. And so I think that I must complain at this question having being put in the 'basic' category of difficulty. I assumed that there must be a very simply way of seeing the answer by some cleaver use of geometry. And I think that there was to eliminate the filling of the triangle as done in triangle C - because I think that you could think of the triangle as cut into 4 parts, all with the maximum size of circle within them, but then within the triangle B there are also the 2 smaller circles with additional area. But I cannot see any 'cleaver way' of seeing that the circles in triangle B are larger in area that the circles in triangle A, which I'm sure is true because of the similarity in areas. And so I think that this question should have been put in the 'intermediate' category of difficulty. And I think that the person or people who determine the level of difficulty should think about their judgement more carefully!! Yours sincerely, David Fairer
Log in to reply
The difficulty level automatically drops as more people solve a problem.
The question is equivalent to finding out which packing of circles fill the largest proportion of the triangle.
Packing C can be broken up into 4 equilateral triangles with an inscribed circle in each.
Packing B is one inscribed circle plus some extra little circles. These little extra circles make packing B is better than C.
Packing A can be broken up into nine equilateral triangles. Some of these are completely red. For each fully coloured in triangle there are 3 extra circle segments.
Packing B can be broken up into 4 equilateral triangles. Some of these are completely red. For each fully coloured in triangle there are 3 extra circle segments. Again, packing B has extra little circles making it better than packing A.
Okay, this is not a super precise proof but good enough.
Edit: the description of Packing A is not right. From the Jeremy's answer, it is clear a more precise comparison is needed before the answer is obvious.
This is interesting, but hard to follow your reasoning for why B>A without pictures.
Log in to reply
I get this bit. Divide C into 4 equal triangles. Each will be identical and have an inscribed circle. These are similar (scaled down). In this respect they have the same proportion of red as B with the 2 small circles removed. Add them and so B>C.
I don't get the 9 triangles statement though.
Log in to reply
Okay so the 9 triangles bit wasn't thought through and doesn't fit. The argument I'm trying to make for B>A is that you could make inscribed triangles in the circles and have three segments left over. It falls down because this covering of A has bits that would overlap and therefore be more efficient at packing. Then the question becomes "is the efficiency gained more than the efficiency of the little circles in B?". From seeing the other answers, the efficiency in B is only slightly better so I would need to come up with a more precise argument before I claimed it was obvious.
The third statement "Packing A can be broken up into nine triangles..." is incorrect.
Log in to reply
True. I need something more precise to fix this argument
This is not a proof. A proof, by definition, is precise. Nothing less is “good enough”.
Log in to reply
Incorrect. It is perfectly valid and extremely common to use inequalities and approximations in proofs, if the proposition is fit for proof this way. It is also extremely common to leave out some details in proofs.
But yeah, you're right in that I should be more precise before I call this a proof.
Ordered by most to the least red area: B, A, C. The B and A is the tricky part, because they are seemingly the same in the red area, where B seems to mislead you that it has a little bit less area. The way you differentiate between is by comparing corners of blue area fused with the tangent blue area if needed to come close to equal, we find out that if we fuse A's upper corner and the two tangent blue sides we will get area approximately equal to that of B's upper corner without the tangents. if we compare the lower left and right corners of A with the lower fused corners with tangent sides of B we will see that there will be an extra centrepiece of blue area and a little bit of leftover of the blue area of A, that's how only by reasoning and sight we can conclude that the correct answer is B, since it has the least amount of blue area, that it has the highest amount of red area.
@Andrei Li , what did the Italian mathematician get?
The way I solved it was finding which triangle has the least blue. "c" has lots of blue so it's red must have the least red. "a" has three blue triangles at the top that looks to be equal to "b"s first triangle, ”b"s six other triangles looks to be equal to "a"s bottom left and right triangles, so now "a" still has a triangle in the middle and bottom, so "b"has a bigger red area.
Problem Loading...
Note Loading...
Set Loading...
Let the equilateral triangles have side length 1.
In triangle A , the circle radius can be found by finding the length of a side of the triangle in terms of r. r = 2 + 2 3 1 . And so the area 3 π r 2 = 8 6 − 3 3 π ≈ 0 . 1 0 0 4 8 π
In triangle B the center of the large circle is the centroid so r 1 = 6 3 and the smaller circles are inscribed in a triangle one-third as large so r 2 = 1 8 3 . And so the area ( r 1 2 + 2 r 2 2 ) π = 1 0 8 1 1 π ≈ 0 . 1 0 1 8 5 π
In triangle C the circle radius can be found in a manner similar to the first. r = 1 2 3 . [In retrospect this is easier to see as half of r 1 from triangle B ] And so the area 4 π r 2 = 1 2 1 π ≈ 0 . 0 8 3 3 3 3 3 π
So the greatest shaded area is in figure B , but just barely.
Note: A and B are so close that setting an approximate equality gives the close approximation 3 ≈ 8 1 1 4 0