Decision tree – perfect & Imperfect info

I really would not waste time on this partly because you cannot be asked to draw a tree, and also because any questions on perfect knowledge are most likely to be as I explain in the lecture on the previous knowledge (and calculations on imperfect knowledge are unlikely).

For perfect knowledge:

Without perfect knowledge we would be on the top part of the tree and they would choose the expensive refurbishment which would give 11.17 – 4 = 7.17

With perfect knowledge, if the knowledge said ‘good’ they would choose expensive refurbishment and get 9.5 (13.5 – 4) and if it said ‘bad’ they would shut down and get 5.00
Therefore the expected return would be (9.5 x 2/3) + (5 x 1/3) = 8

So the most to pay for perfect knowledge would be 8 – 7.17 = 0.83.

For imperfect knowledge, without it gives 7.17. With it (using the printed answer) gives (point F) 7.63. So the most to pay for imperfect knowledge is 7.63 – 7.17 = 0.46

(A minor point – why do you use 68% and 32%? The probabilities are 2/3 and 1/3. By all means use %’s if you want, but if so they are 67% and 33% 🙂 )

You are welcome 🙂