MathIsimple

Probability Theory – Problem 61: Prove that is also a stopping time with respect to

Question

Let {Fn}\{\mathcal{F}_{n}\} be a filtration (an increasing sequence of σ\sigma-fields). Let TT be a stopping time with respect to {Fn}\{\mathcal{F}_{n}\}, and let τ\tau be an integer-valued random variable that is FT\mathcal{F}_{T}-measurable with τT.\tau\geq T. Prove that τ\tau is also a stopping time with respect to {Fn}\{\mathcal{F}_{n}\}.

Step-by-step solution

Proof: Step 1. First decompose {τn}\{\tau \le n\} according to the values of TT:

{τn}=k=0n({τn}{T=k}).\{\tau \le n\} = \bigcup_{k=0}^n \left( \{\tau \le n\} \cap \{T = k\} \right). Since τT\tau \ge T, when T=kT = k, the condition τn\tau \le n implies kτnk \le \tau \le n, so knk \le n holds automatically. Thus k=0,1,,nk = 0,1,\dots,n.

Step 2. Since {τm}FT\{\tau \le m\} \in \mathcal{F}_T, taking m=nm = n: {τn}{T=k}Fk,k.\{\tau \le n\} \cap \{T = k\} \in \mathcal{F}_k, \quad \forall k.

Moreover, since knk \le n and FkFn\mathcal{F}_k \subset \mathcal{F}_n, we have {τn}{T=k}Fn,k=0,,n.\{\tau \le n\} \cap \{T = k\} \in \mathcal{F}_n, \quad \forall k = 0,\dots, n.

Therefore {τn}=k=0n({τn}{T=k})Fn.\{\tau \le n\} = \bigcup_{k=0}^n \left( \{\tau \le n\} \cap \{T = k\} \right) \in \mathcal{F}_n.

Step 3. In summary, for every nn, {τn}Fn\{\tau \le n\} \in \mathcal{F}_n, so τ\tau is a stopping time. QED.

Final answer

QED.

Marking scheme

The following is the complete marking scheme for this probability theory problem.

1. Checkpoints (max 7 pts total)

Core idea: Use the definition of FT\mathcal{F}_T to decompose {τn}\{\tau \le n\} and convert it into events in Fk\mathcal{F}_k.

*The following steps form a single logical chain; points are awarded cumulatively:*

  • Event decomposition [additive] (1 pt)
  • Decompose the event {τn}\{\tau \le n\} by the values kk of TT, writing a union of the form:

{τn}=k({τn}{T=k})\{\tau \le n\} = \bigcup_{k} \left( \{\tau \le n\} \cap \{T = k\} \right)

  • *Note: The upper limit of summation may be nn or \infty (as long as subsequent handling is correct). If the union formula is not explicitly written but each term for each kk is correctly discussed subsequently, credit may still be awarded.*
  • Applying the FT\mathcal{F}_T-measurability definition [additive] (3 pts)
  • Identifying the measurable set (1 pt): State that {τn}\{\tau \le n\} (as the preimage of a Borel set under the random variable τ\tau) belongs to FT\mathcal{F}_T.
  • Applying the definition (2 pts): Using the definition AFT    A{T=k}FkA \in \mathcal{F}_T \iff A \cap \{T=k\} \in \mathcal{F}_k for all kk, conclude:

{τn}{T=k}Fk\{\tau \le n\} \cap \{T = k\} \in \mathcal{F}_k

  • *Scoring note: This is the most critical technical step. If the student only vaguely writes "since it is measurable, it belongs to Fk\mathcal{F}_k" without explicitly invoking the structural definition of FT\mathcal{F}_T, deduct 1 point.*
  • Monotonicity of the filtration and index restriction [additive] (2 pts)
  • Index control (1 pt): State that since τT\tau \ge T, when {τn}\{\tau \le n\} occurs we necessarily have TnT \le n (i.e., knk \le n), or state that when k>nk > n the intersection is the empty set (which belongs to Fn\mathcal{F}_n).
  • Subset property (1 pt): Use kn    FkFnk \le n \implies \mathcal{F}_k \subseteq \mathcal{F}_n to argue that the above non-empty intersections belong to Fn\mathcal{F}_n.
  • Conclusion [additive] (1 pt)
  • State that σ\sigma-algebras are closed under finite unions, therefore {τn}Fn\{\tau \le n\} \in \mathcal{F}_n, hence τ\tau is a stopping time.
  • *Note: This must be based on the preceding proof; stating only the conclusion without justification receives no credit.*

Total (max 7)


2. Zero-credit items

  • Merely copying the problem conditions (e.g., listing "τT\tau \ge T", "τ\tau is FT\mathcal{F}_T-measurable").
  • Merely listing the general definition of stopping time or σ\sigma-algebra without applying it to the specific variables τ\tau and TT.
  • Citing a theorem: Directly citing "if τT\tau \ge T and is FT\mathcal{F}_T-measurable then τ\tau is a stopping time" as a known result without providing a proof.
  • Circular reasoning: Using the conclusion "τ\tau is a stopping time" within the proof (e.g., using properties of Fτ\mathcal{F}_\tau).

3. Deductions

*Only deduct for the most significant single error; deductions should not reduce the total below 0.*

  • Index logic error (flat -2):
  • Summing over kk up to \infty in the decomposition and claiming FkFn\mathcal{F}_k \subseteq \mathcal{F}_n for all kk (ignoring the case k>nk > n).
  • *Exception: If the student correctly identifies that the set is empty when k>nk > n, no deduction applies.*
  • Symbol confusion or unclear definitions (flat -1):
  • Confusing events (sets) with random variables, or confusing the relationship between FT\mathcal{F}_T and Fn\mathcal{F}_n.
  • Failing to specify the range of kk or handling T=kT=k ambiguously.
  • Logical gap (flat -1):
  • After obtaining {τn}{T=k}Fk\{\tau \le n\} \cap \{T=k\} \in \mathcal{F}_k, jumping directly to the conclusion {τn}Fn\{\tau \le n\} \in \mathcal{F}_n without mentioning FkFn\mathcal{F}_k \subset \mathcal{F}_n.
Ask AI ✨