Show group of order $4563=3^3\cdot13^2$ is not simple. 
Show a group of order $4563=3^3\cdot13^2$ is not simple.

I am confused when dealing with these Sylow's game. Especially when the order becomes large. It seems there is no common rules to solve this kind problems.
My try: $n_{13}=1+13t$ and $n_{13}$ divides $3^3=27$, hence $t=0,2$. If $t=0$ then we are done. If $t=2$, There are $27$ 13-subgroups of order 169. Then consider the conjugation action of G on the set of 13-subgroups. Prove it induced a homomorphism between $G$ and $S_{27}$. And want to claim a contradiction by  $|G|$ does not divide $27!$. But sadly it does divide. Any other method can get this figured out?
 A: The next things to try are often:


*

*to consider the intersections of Sylow groups, and

*to use the fact that groups of order $p^2$, $p$ a prime, are always abelian.


Let's see. Assume that we have two Sylow $13$-subgroups that intersect non-trivially,
say $P_1\cap P_2$ and $x\neq 1_G, x\in P_1\cap P_2$ with $|P_1|=|P_2|=13^2$. Then $x$ is centralized by both $P_1$ and $P_2$. This means that $C_G(x)$ has order that is a proper multiple of $13^2$. The size of the conjugacy class of $x$, $[G:C_G(x)]$, is thus a proper factor of $[G:P_1]=27$, and therefore at most nine. If $x$ is in the center, $G$ cannot be simple. Otherwise we get a non-trivial homomorphism from $G$ to $S_9$ by the conjugation action on the conjugacy class of $x$. Again, we are done.
So let's assume that all those twenty-seven Sylow $13$-subgroups intersect trivially. In that case their union contains $(27\cdot 168)+1=|G|-26$ elements. Now we can look at Sylow $3$-subgroups. Their non-identity elements must all be among those $26$ elements not belonging to any Sylow $13$. This means that there is room for only a single Sylow $3$-subgroup, which is then normal concluding the proof.
A: The Burnside Transfer Theorem does this even more briefly.  A 13-Sylow cannot be in the center of its normalizer, so, as every group of order $13^2$ is abelian, it must be properly contained in its normalizer, which thus has index at most $9$.  And we again conclude that either the 13-Sylow is normal or there is a non-trivial homomorphism to $\mathfrak{S}_9$.
Of course, the Burnside $p^aq^b$ theorem is faster still....
