Indexing in nested loops x 3

My title may not make sense, but I do not know the technical term. In the code below, I want to use the elements in sv_10 to index into pairs, which can only be done by indexing via sv_gain/sv_loss. However, I have tried several ways to do this and none of them work. Please help

G = 75:5:110;
L = 15:5:50;
[g,l] = meshgrid(G,L); %returns 2-D grid; The grid represented by the coordinates g and l has length(L) rows and length(G) columns
pairs = [g(:) l(:)];

p =  0.5 % 0.1:0.1:1 % one p
delta = 0.6;
gamma = .5; % (0.2:0.1:0.8);

y = zeros(length(p),length(gamma));
sv_gain = zeros(length(pairs),length(p))
sv_loss = zeros(length(pairs),length(p))
sv_gamble = zeros(length(pairs),length(p))

for i = 1:length(p)
        y(i) = (delta .* p(i).^gamma/(delta .* p(i).^gamma + (1 - p(i)).^gamma))
        for c = 1:length(pairs)
            sv_gain(c,i) = pairs(c,1).* y(i)   % sv_gain            
            sv_loss(c,i) = pairs(c,2).* (y(i).* 2) %sv_loss
        
            sv_gamble(c,i) = (sv_gain(c))- (sv_loss(c))
            ev(c) = (pairs(c,1)* p(i)) + (pairs(c,2)* p(i))
        end 
end

ev = transpose(round(ev))
sv_gamble = round(sv_gamble)
upbound = 10
lowbound = -10
sv_10 = sv_gamble(sv_gamble >= lowbound & sv_gamble <= upbound)

g_10 = [];
l_10 = [];
pairs_10 = [];
for k = 1:length(sv_10)  
    idx = find(round((sv_gain-sv_loss) == sv_10(k)))
    for o= 1:length(idx)
        idx_gain = idx(o)
        idx_loss = find(sv_loss == sv_gain(idx_gain)-sv_10(k))
           if ~isempty(idx_loss)
            pairs_10 = [pairs_10; idx_gain, idx_loss];
            
           end
    end
end

This returns empty vectors which should contain the values that correspond to sv_gain/sv_loss, and then after I would be able to use those to index into pairs

Answer 1

[Matt J]

You are using '==' in several places in your code. When you do so, you must remember the limitations of floating point math,

0.7*0.4==0.28

ans = logical
   0

abs(0.7*0.4-0.28)

ans = 5.5511e-17

[Jade T]

Thanks, I defintely overlooked this!