Not sure if this is much better, at least to me it is much more line-noise and harder to parse.

map{sub __;$/=$_;sub _;map{$_[$/][$_]{_}=rand(time)%2;push@{$_[$/][$_]
+{map}},(
\$_[_$/][$_]{_}=>\$_[__$/][$_]{_}=>\$_[$/][_$_]{_}=>\$_[$/][__$_]{_},\
+$_[_$/][
_$_]{_}=>\$_[_$/][__$_]{_}=>\$_[__$/][_$_]{_}=>\$_[__$/][__$_]{_},)}(0
+..$__)}(
0..($__=(shift||24)-1));sub __{(($-=shift)+1)<=$__?$-+1:0}sub _{(($-=s
+hift)-1)
>=0?$--1:$__;}__;__;while(){map{$/=$_;$\="\n";print map{$_[$/][$_]{sub
+}=$_[$/]
[$_]{_}?'*':'.';$.+=$$_ for(@{$_[$/][$_]{map}});$![$/][$_]{_}=$_[$/][$
+_]{_};$!
[$/][$_]{_}=($.==3?1:($.!=2?0:($![$/][$_]{_}==1?1:0)));__;$.=~s _.__;$
+_[$/][$_
]{sub};}(0..$__);}(0 .. $__);map{$/=$_;map{$_[$/][$_]{_}=$![$/][$_]{_}
+}(0..$__
)}(0..$__);_;for(1..4){_;__;}_;sleep 1;}for(1..3){__;_;__;_;_;}__;_;__
+;_;__;
[download]

--
perl -p -e "s/(?:\w);([st])/'\$1/mg"