Tensions in Practice: First-arrival capture in tension with continued exploration¶
One-dimensional walk · a lower boundary
A walker starts at position 1 and attempts fair, independent steps of −1 or +1. The displayed realization is −1, −1, +1, +1. An absorbing boundary at 0 stops the walker at its first contact. A reflecting boundary keeps it active: from 0, an attempted −1 is redirected to +1. The attempted sequence is the same, but the boundary supplies a different transition rule.
Capture first arrival
End a trajectory when it first reaches the collection boundary.
Retain an active walker
Keep the walker inside the allowed half-line so motion can continue.
Why these aims pull against each other
Absorption makes first contact terminal; reflection preserves participation by changing outward motion. These are different boundary mechanisms, not interchangeable approximations of one fixed mechanism.
Choose an arrangement to see what changes and what remains difficult.
Finite illustrative comparisons. Text states carry the meaning; color is not a measured score or universal preference.
What this choice protects
What it costs
When it fits
Compare the arrangements
Capture at zero
Stop at the first visit to zero and keep a terminal position record there. Later attempted steps have no effect.
| Attempt | Position | Status | |
|---|---|---|---|
| Step 1 | −1 | 0 | Stopped |
| Step 2 | −1 | 0 | Stopped |
| Step 3 | +1 | 0 | Stopped |
| Step 4 | +1 | 0 | Stopped |
- What it protects
- The captured record directly marks first arrival; no later motion can be mistaken for continued participation.
- What it costs
- The walker no longer explores and cannot contribute later positions.
- When it fits
- Fits a collection, removal or first-passage process whose boundary really ends the trajectory.
Illustration note: The finite setting and values are editorial assumptions, not measured effects or recommended operating settings. Subsequent draws are displayed solely to align the comparison, not as moves taken by a stopped walker.
Reflect at zero
Allow visits to zero, but redirect any outward attempted step there inward by one unit.
| Attempt | Position | Status | |
|---|---|---|---|
| Step 1 | −1 | 0 | Active |
| Step 2 | −1 | 1 | Active |
| Step 3 | +1 | 2 | Active |
| Step 4 | +1 | 3 | Active |
- What it protects
- The walker remains active and explores allowed positions after contact.
- What it costs
- The post-contact path is boundary-dependent and cannot serve as the trajectory of a captured or extinguished walker.
- When it fits
- Fits a reflecting wall or a deliberately constrained ongoing exploration process.
Illustration note: The finite setting and values are editorial assumptions, not measured effects or recommended operating settings. The convention is bounce-to-one, not remain-at-zero rejection; other reflection conventions would produce different paths.
What this illustration does—and does not—establish
Random Walk: Pure Walk versus Walk-with-Boundary (Reflection, Absorption, Bias) supplies boundary-dependent behavior. Both arrangements are correct under their own declared boundary contract; neither is a correction of a knowingly false calculation.
- This is one declared four-step realization, not a claim about average displacement, ruin probability or first-passage time.
- Attempted increments are independent fair draws; the actual bounded transitions are state-dependent after contact.
- No upper boundary or drift is included. Selecting the boundary requires evidence about the modeled process or an explicit design objective.
Source entries
Random Walk
Random Walk: Pure Walk versus Walk-with-Boundary (Reflection, Absorption, Bias) supplies the conflict examined here.
Pure Walk versus Walk-with-Boundary (Reflection, Absorption, Bias)
The free random walk explores an unbounded space, but most applied walks live in bounded or biased environments — absorbing barriers (gambler's ruin, default, extinction), reflecting walls, or a drift that biases the steps.