First step analysis in the context of hitting times

113 Views Asked by At

I've recently started reading through Stochastic Calculus and Financial Applications by J. Michael Steele and I'm not sure I understand how first step analysis works regarding hitting times. Exercise 1.1 of the book gives the following probability distribution for the up and down movements of a stock:

$$\mathbb{P}(Y_{n+1}=k+1|Y_n=k)=1/3\quad \mathrm{and}\quad \mathbb{P}(Y_{n+1}=k-1|Y_n=k)=2/3 \;\;\mathrm{for} \;k>20$$ $$\mathbb{P}(Y_{n+1}=k+1|Y_n=k)=2/3\quad \mathrm{and}\quad \mathbb{P}(Y_{n+1}=k-1|Y_n=k)=1/3 \;\;\mathrm{for} \;k<20$$ $$\mathbb{P}(Y_{n+1}=21|Y_n=20)=0.9\quad \mathrm{and}\quad \mathbb{P}(Y_{n+1}=19|Y_n=20)=0.1$$ The exercise asks for the expected value of the hitting time: $$\tau = \min (n:Y_n=18|Y_0=25)$$ The answer in: Finding the expected time for a stock to go from $\$25$ to $\$18$ given there is a support level at $20 with upward and downward biases. suggests that we can partition the expected time into three distinct parts. They derive an answer for the 21->20 case, but I would like some clarification.

Suppose that $\tau_{1} = \min(n:Y_{n+1}<Y_n)$, then by applying the Law of Total Expectation, conditioning on the first step, we have $$\mathbb{E}[\tau_{1}|Y_0=25]=\mathbb{P}(Y_1=26|Y_0=25)\cdot\mathbb{E}[\tau_{1}|Y_0=25,Y_1=26]+\mathbb{P}(Y_1=24|Y_0=25)\cdot\mathbb{E}[\tau_{1}|Y_0=25,Y_1=24]+1$$ The last '$+1$' accounts for the fact that we have 'moved forward' by adding the next $Y_1$ conditions. Then, applying the known probability distribution of $Y$, we have $$\mathbb{E}[\tau_{1}|Y_0=25]=\frac 13\cdot\mathbb{E}[\tau_{1}|Y_0=25,Y_1=26]+\frac 23 \cdot\mathbb{E}[\tau_{1}|Y_0=25,Y_1=24]+1$$ I know that $\{Y_0=25,Y_1=24\} \Rightarrow \tau_1=1$, which means $$\mathbb{E}[\tau_{1}|Y_0=25]=\frac 13\cdot\mathbb{E}[\tau_{1}|Y_0=25,Y_1=26]+\frac 23+1,$$ but this is where I get stuck. The book and the above answer imply that $$\mathbb{E}[\tau_{1}|Y_0=25,Y_1=26] = \mathbb{E}[\tau_1+1|Y_0=25].$$ Why exactly are we adding the '$+1$'? Is this true in general, i.e. $\mathbb{E}[\tau_{1}|Y_0=k,Y_1=k+1] = \mathbb{E}[\tau_1+1|Y_0=k]$? The book gives this rationale to a similar problem:

When we take our first step, two things happen. First, there is the passage of one unit of time; and, second, we will have moved from zero to either -1 or 1.

Doesn't the first '$+1$' in the beginning application of the Law of Total Expectation account for this passage of time? I realise that this question may be a little trivial, but any help is appreciated!