Provider selection factors, in one clear definition
Provider selection factors are a set of criteria you use to decide whether a provider is suitable for copy-style or automated-following arrangements. The key idea is that you evaluate factors (inputs and constraints) rather than treat past results as a prediction of future outcomes.
A worked example helps because it forces you to state assumptions: what numbers you use, how you measure them, and what you assume stays constant.
Worked example: ranking two hypothetical providers with the same inputs
Scenario and assumptions (all stated)
Assume you are evaluating two hypothetical providers, A and B, using the same allocation model.
Assumptions
- You allocate the same base amount to copy: $1,000.
- Each provider delivers returns to a strategy performance stream; you assume your copy allocation results in proportional changes before any additional frictions.
- You model a simple cost for the provider relationship: 0.50% of positive profit (performance fee) for both providers.
- You ignore currency conversion costs and any tax effects because you want a clean mechanics example.
- The evaluation period is one month.
- You assume the only numeric outputs available for this month are gross return, gross drawdown, and average monthly trade frequency.
Inputs for the month
- Provider A
- Gross return: +4.0%
- Gross maximum drawdown (peak-to-trough): -2.0%
- Trade frequency: 30 trades
- Provider B
- Gross return: +3.5%
- Gross maximum drawdown: -1.0%
- Trade frequency: 10 trades
Step 1: Convert gross return to profit
- Provider A gross profit = $1,000 × 4.0% = $40
- Provider B gross profit = $1,000 × 3.5% = $35
Step 2: Apply the performance fee assumption
Performance fee = 0.50% of profit (only on positive profit, per the assumption).
- Provider A fee = $40 × 0.50% = $0.20
- Provider B fee = $35 × 0.50% = $0.175
Step 3: Net profit (under the assumptions)
- Provider A net profit = $40 − $0.20 = $39.80
- Provider B net profit = $35 − $0.175 = $34.825
Step 4: Compare using two example factors (one return, one risk)
We use a simple scoring rule just to illustrate mechanics; your own scoring could differ.
Assume:
- Return factor weight: 70%
- Drawdown factor weight: 30%
Normalize each metric to a 0–1 scale for the month (illustrative method):
- Higher gross return gets higher score.
- Lower (less negative) drawdown magnitude gets higher score.
Return score:
- Provider A: 4.0 / (4.0 + (0.0 to compare)) is ambiguous without a chosen normalization. To keep it explicit, choose this direct relative rule: score = gross_return / 4.0.
- A score = 4.0 / 4.0 = 1.00
- B score = 3.5 / 4.0 = 0.875
Drawdown score (less negative is better): choose relative to the worst drawdown among the two: score = drawdown / (-2.0).
- Provider A drawdown = -2.0 → score = (-2.0)/(-2.0) = 1.00 (interpreting “1” as “highest risk level” is confusing)
Important adjustment to keep direction correct: Instead, use “risk improvement” score = (-drawdown) / (max_risk), where max_risk = 2.0, then invert so that lower risk gets higher score.
- Risk amount = -drawdown
- A risk amount = 2.0 → risk_score = 2.0/2.0 = 1.0
- B risk amount = 1.0 → risk_score = 1.0/2.0 = 0.5
- Drawdown factor score = 1 − risk_score
- A drawdown score = 1 − 1.0 = 0.0
- B drawdown score = 1 − 0.5 = 0.5
Weighted final score:
- Provider A = 0.70×1.00 + 0.30×0.0 = 0.70
- Provider B = 0.70×0.875 + 0.30×0.5 = 0.6125 + 0.15 = 0.7625
Result under these explicit assumptions and this explicit scoring rule: Provider B wins the score despite lower net profit in this single month.
How this “worked example” reflects the mechanics and where it can fail
Stable mechanics vs variable conditions
What the example does show well:
- How costs can reduce net profit.
- How a scoring model can trade off return against drawdown.
What the example cannot guarantee:
- Future performance, because returns and drawdowns vary with market conditions. - Whether actual copy outcomes match proportional assumptions, because real copy execution can be affected by matching, latency, order handling, and the specific provider/platform mechanics.