Creating a matrix based on information in a vector and another matrix

Suppose I have a vector of success probabilities, for example:

 p = [0.5,0.25,0.75]

I also have a matrix of outcomes, o, which has the same number of columns as p, containing "0's" (failure) and "1's" (success):

 o = [1,0,1
      0,0,0]

What I wish to obtain is the following matrix:

m = [p(1,1), 1-p(1,2),   p(1,3)
   1-p(1,1), 1-p(1,2), 1-p(1,3)]

So, to clarify, 1's in o result in a value equal to the corresponding value in p, while 0's in o lead to 1 minus the corresponding value in p.

What is the most efficient way to obtain m?

[alice]

If p is a row vector, what do you mean by p(2,3)?

Do you mean:

m = [p(1), 1-p(2),   p(3)
   1-p(1), 1-p(2), 1-p(3)]

?

[Ulrik William Nash]

Oh, my mistake. Yes. That is, of course, what I meant. Sorry about this.

Answer 1

[Andrei Bobrov]

For MARLAB >= R2016b

out = ~o - p;
out(o>0) = -out(o>0);

MATLAB <= R2016a

out = bsxfun(@minus,~o,p);
out = out.*(1-o*2);

[alice]

I think this formula gives wrong signs.

[Andrei Bobrov]

Thank you Alice! My mistakes.

[Ulrik William Nash]

Thank you, Andrei. The following is just what you wrote, but in line with the wording of my question:

m = bsxfun(@minus,~o,p).*(1 - o*2)

I don't have the latest version of MATLAB.

[Andrei Bobrov]

As you please. Accept this answer or.. ?