Let $E$ be a compact metric space. Suppose that closure of every open ball $B(a,r)$ is the closed ball $B'(a,r)$. Must every open ball in $E$ be connected? I think it most probably is. But I don't know how to go about proving this.
put on hold as offtopic by user44191, Sean Lawton, Yemon Choi, Davide Giraudo, Wojowu Mar 15 at 17:15
This question appears to be offtopic. The users who voted to close gave this specific reason:
 "MathOverflow is for mathematicians to ask each other questions about their research. See Math.StackExchange to ask general questions in mathematics." – user44191, Sean Lawton, Yemon Choi, Davide Giraudo, Wojowu
Yes, every open ball is connected.
Suppose the open ball $B(a,r)$ is disconnected: $B(a,r) = U \cup V$ where $U$ and $V$ are nonempty, open and disjoint, and $a \in U$. Since $\overline{V}$ is compact, there is a point $v \in \overline{V}$ whose distance $s = d(a,v)$ to $a$ is minimal. Since $V \subset B(a,r)$, $s < r$ and $v \in B(a,r)$. Note that $U \cap \overline{V} = \overline{U} \cap V = \emptyset$, so $v \in V$ and $v \notin \overline{U}$. Thus we have $v \in B'(a,s)$, but $B(a,s) \subseteq U$ so $v \notin \overline{B(a,s)}$, contradicting the assumption $\overline{B(a,s)} = B'(a,s)$.

$\begingroup$ Can you explain how v does not belongs to closure of U. I don't think anything is stopping from this. Of course then s would be zero. $\endgroup$ – cherry Mar 16 at 6:37

$\begingroup$ @cherry $v$ is in $V$, which is open. Thus there is a neighbourhood of $v$ which is disjoint from $U$, so $v \notin \overline{U}$. $\endgroup$ – Robert Israel 2 days ago
Yes, it is true.
Assume that an open ball $B(a,R)$ is not connected. Let $S\not\ni a$ be a connected component of $B(a,R)$. Since the space is compact there is a point $s\in S$ that minimize the distance $as$. Note that $s$ does not lie in the closure of $B(a,r)$ for $r=as$  a contradiction.

$\begingroup$ Why is $S$ compact (from which you know distance from $S$ to $a$ has a minimum)? A connected component of an open subset is closed in the open subset but need not be closed (hence compact) in the whole space, in general. $\endgroup$ – KConrad Mar 15 at 10:01

1$\begingroup$ @KConrad: I don't think this is being claimed: I interpret "the space" as the whole space, not $S$, though then the claim should be "... there is a point $s\in\overline{S}$ ..." Robert just posted the same argument with more details. $\endgroup$ – Christian Remling Mar 15 at 14:57

1$\begingroup$ Oh, so "$s \in S$" should be "$s \in \overline{S}$". Then my objection no longer applies. $\endgroup$ – KConrad Mar 15 at 15:07

$\begingroup$ @KConrad, Note that $xa=R$ for any $x\in \bar S\backslash S$, therefore $s\in S$. $\endgroup$ – Anton Petrunin Mar 16 at 0:04