Sample paths of B.M are not monotonic in any interval.

96 Views Asked by At

Is this an ok proof for a.s no interval of monotonicity for sample paths of B.M

EDIT : [In the proceeding proof I just prove the sample path is not strictly increasing, it is an equivalent proof to show it is not strictly decreasing ]

first we show for the interval [0,1]

$P(B(t)-B(s)\geq 0$ $ \forall s,t\in [0,1]) $

$\leq P(B(\frac{i+1}{n})-B(\frac{i}{n})\geq0 $ for $i =0,...n-1$ )= $\frac{1}{2^n}$ $\to 0 $

now we use scaling properties and stationary increments to prove for rational intervals [p,q]

consider for , p,q rational

$P(B(t)-B(s)\geq 0$ $ \forall s,t\in [p,q]) $

$=P(B(t-p)-B(s-p)\geq 0$ $ \forall s,t\in [p,q]) $ by stationary property

$=P(B(t)-B(s)\geq 0$ $ \forall s,t\in [0,q-p]) $

$=P((\frac{1}{q-p})^{-1/2}B(t\frac{1}{q-p})-(\frac{1}{q-p})^{-1/2}B(s\frac{1}{q-p})\geq 0$ $ \forall s,t\in [0,q-p])$ by scaling property

$=P(B(t)-B(s)\geq 0$ $ \forall s,t\in [0,1]) $

edit: If we have no monitancity in all rational intervals then we have no monotonicity in any arbitrary interval since all rational intervals contain such arbitrary intervals.