Based on the scatter mentioned in you question I combined the two loops and the resulting points in a single grid. This allows to plot the results with a single call to the scatter command. Is this what you looking for? I made a minor modification to the routing so that it returns lines of equal height.
samples = 1000;
S1 = zeros(1000,1000);
S2 = zeros(1000,1000);
maxVal = zeros(1000,1);
minVal = zeros(1000,1);
for z = 1:1000
maxVal(z) = round(rand*195)+5;
minVal(z) = round(rand*3)+2;
repeat = round(rand*8)+2;
xr = 0.01/repeat;
x = linspace(0, 10, samples);
y = mod(x,xr)*(maxVal(z) - minVal(z))/xr + minVal(z);
S1(z,:) = y;
x = linspace(10, 0, samples);
y = mod(x,xr)*(maxVal(z) - minVal(z))/xr + minVal(z);
S2(z,:) = y;
end
% gather the points in a single grid
Grid = [ (1:length(S1(1,:)))' S1(1,:)';
(1:length(S2(1,:)))' S2(1,:)'];
figure
scatter(Grid(:,1), Grid(:,2))
grid on
Edit, extra images after comment by the OP. Note that there are some odd points, that appear not to be connected. These are artefacts from the mod operation that generates the points. Have a good look at the algortihm that generates the points.
%% now gather all sides for the seperate 'triangles'
% since the lines are generated by random comands, we need a bit of logic to seperate them
repIdx = [0 find(diff(S1(1,:)) < 0) samples];
F1 = [];
F2 = [];
for i = 1:2:(numel(repIdx)-2)
F1 = [F1 S1(1,repIdx(i)+1:repIdx(i+1)) S2(1,(repIdx(i+1)+1):repIdx(i+2))];
F2 = [F2 S2(1,repIdx(i)+1:repIdx(i+1)) S1(1,(repIdx(i+1)+1):repIdx(i+2))];
end
F1 = [(1:numel(F1))' F1'];
% this gahters the points for triangle 2
F2 = [(1:numel(F2))' F2'];
figure
subplot(1,2,1)
scatter(F1(:,1), F1(:,2))
grid on
subplot(1,2,2)
scatter(F2(:,1), F2(:,2))
grid on
%% only gather points for the first 'triangle'
T1 = [(1:2*repIdx(2))' [S1(1,1:repIdx(2)) S2(1,(repIdx(2)+1:repIdx(3)))]'];
T2 = [(1:2*repIdx(2))' [S2(1,1:repIdx(2)) S1(1,(repIdx(2)+1:repIdx(3)))]'];
figure
subplot(1,2,1)
scatter(T1(:,1), T1(:,2))
grid on
subplot(1,2,2)
scatter(T2(:,1), T2(:,2))
grid on
