problem with low infinite number

 Ediff(8,35)
 ans =
   7.4525e+03
   
 profit(8,35)
ans =
   7.4525e+03  

 bu=minus(profit,Ediff);
 
 bu(8,35)

ans =

   1.8190e-12
  

Why does subtracting 2 equal numbers produce a very small number?

[Trop Trader]

ok thank

Answer 1

[the cyclist]

Because their floating-point representations are not exactly equal.

You can see that they are not quite equal:

load("matlab_Ediff.mat","Ediff");
load("matlab_profit.mat","profit");

fprintf("Ediff = %16.14f\n",Ediff(8,35))

Ediff = 7452.50000000000182

fprintf("profit = %16.14f",profit(8,35))

profit = 7452.50000000000364

Answer 2

[John D'Errico]

They only look the same, to 4 decimal places.

format short
x = 1.23456789;
y = 1.23456987;
[x,y]

ans = 1×2

1.2346 1.2346

They certainly look the same to me. But only if I don't look at their actual values, down to the last digits. If I subtract them of course, they are seen to be different.

x - y

ans = -1.9800e-06

Surely you would expect the difference to be non-zero, even though, when displayed as only 5 dignificant digits, they seem identical?

format long
[x,y]

ans = 1×2

1.234567890000000 1.234569870000000