Direct answer
Market selection is the process of deciding which market conditions (for example, volatility regime, liquidity level, and whether spreads or execution quality are acceptable) you will include in your trading plan. A “worked example” shows, step by step, how you apply those criteria and how the numbers change when your assumptions change.
Mechanism or definition
Think of market selection as a gate in front of trading. It usually has three parts:
- Criteria: fixed rules for what “good enough” conditions mean (e.g., you only act when spreads are below a threshold, or when historical price movement has stayed within a chosen volatility band).
- Inputs: the data you use to evaluate the criteria (e.g., observed spread, a volatility measure, and your expected trading costs).
- Decision rule: if the criteria pass, you allow trading; if they fail, you do not.
A key point is separating stable mechanics (your rules and calculations) from variable conditions (market behavior, costs, and execution). Your selection process should aim to keep the decision rule consistent even when conditions change.
Evidence or example (hypothetical, fully specified)
Goal
Show how a simple market-selection filter changes which trades you would consider.
Assumptions (all hypothetical)
- You are testing a plan over 10 potential trading instances.
- For each instance, you observe a quoted spread (in price points) and you estimate that your execution cost is: execution cost points = quoted spread points × 1.0 (assume no extra slippage for simplicity).
- Your plan only allows trading if quoted spread ≤ 1.5 points.
- You measure outcome in “points” using a simplified model where: net result = (future move points) − (execution cost points).
- For illustration, the future move points for each instance are predetermined by the test data (no live pricing).
Data
For 10 instances, assume the following (quoted spread, future move):
- Instance 1: (1.0, +2.0)
- 2: (1.8, +3.0)
- 3: (1.2, +1.4)
- 4: (1.6, +2.5)
- 5: (1.4, +1.2)
- 6: (2.0, +4.0)
- 7: (1.1, +1.0)
- 8: (1.5, +2.2)
- 9: (1.3, +0.8)
- 10: (1.7, +3.5)
Selection decision
Allow trading only when spread ≤ 1.5:
- Allowed: 1, 3, 5, 7, 8, 9 (6 instances)
- Not allowed: 2, 4, 6, 10 (4 instances)
Net results calculation
Using net result = future move − spread:
- 1: +2.0 − 1.0 = +1.0
- 3: +1.4 − 1.2 = +0.2
- 5: +1.2 − 1.4 = −0.2
- 7: +1.0 − 1.1 = −0.1
- 8: +2.2 − 1.5 = +0.7
- 9: +0.8 − 1.3 = −0.5
Sum over allowed instances: +1.0 +0.2 −0.2 −0.1 +0.7 −0.5 = +1.1 points.
Comparison to “no selection” (same assumptions)
If you traded all 10 instances, net results would be:
- 2: +3.0 − 1.8 = +1.2
- 4: +2.5 − 1.6 = +0.9
- 6: +4.0 − 2.0 = +2.0
- 10: +3.5 − 1.7 = +1.8
Add these to the allowed sum: +1.1 + (1.2+0.9+2.0+1.8) = +1.1 +5.9 = +7.0 points.
What this illustrates
In this hypothetical dataset, filtering by spread reduced the number of trades and reduced total points. That does not mean market selection is wrong; it means the effect depends on what market-selection criteria are actually capturing. In other datasets, filtering could reduce losses more than it reduces gains.
Limitations and risks
- Assumption sensitivity: If execution includes slippage, then execution cost points may be higher than quoted spread, changing net results.
- Changing relationships: Historical price movement, spread behavior, or volatility regimes may not repeat, so past selection performance can fail.
- Hidden failure modes: Selection criteria can be met due to temporary conditions that later reverse (e.g., a brief spread improvement while volatility rises).
- Overfitting: If criteria are tuned to one period, they may not generalize to other periods.
Verification or next question
To independently verify market selection in practice, use a test design where you:
- Keep the selection criteria and decision rule fixed.
- Evaluate performance separately for allowed vs not-allowed instances.
- Stress assumptions that affect costs and execution (especially when you don’t observe real execution prices).