What is a good portfolio score?

Any system that compresses a portfolio into one number is throwing information away. That is the point — a single score is a triage signal, not a diagnosis. This guide covers what a composite score can legitimately tell you, where it misleads, and why the components underneath it are the part worth reading.

What a composite score measures

A portfolio score is a weighted judgement about structure, not a prediction of returns. This distinction is the single most important thing to understand about it. No score can tell you whether a portfolio will make money, because that depends on future prices, which are unknown.

What a score can assess is whether a portfolio is built sensibly:

• Is capital spread across enough genuinely independent positions?
• Are the underlying holdings financially sound?
• Does the risk taken match the risk intended?
• Is the allocation aligned with, or fighting against, current market conditions?

A high score means no structural problems were found. It does not mean the portfolio will outperform, and a portfolio can score well and still lose money if the market falls.

Reading the ranges

Score bands are conventions rather than laws, but on a 0–100 scale the following interpretation is broadly reasonable:

The components underneath

A composite is only as honest as its decomposition. Separating the score into independent dimensions is what turns it from a verdict into something actionable:

These measure genuinely different things and routinely disagree. That disagreement is the most useful output of the whole exercise.

When components disagree

Consider a real evaluation of a three-position technology portfolio — Apple, Microsoft, Nvidia:

{
  "score": 45.9,
  "status": "Neutral",
  "score_components": {
    "risk":            99.2,
    "quality":         72.1,
    "diversification": 34.5,
    "signal":          33.9,
    "trend":           17.3
  },
  "max_position_weight": 45.45,
  "top_sector": "Technology",
  "top_sector_weight": 100
}

The composite of 45.9 reads as mediocre and tells you almost nothing. The components tell you precisely what is happening: these are excellent, stable businesses — risk scores 99.2 and quality 72.1 — but every one of them is the same bet. Diversification scores 34.5 because the portfolio is one hundred percent technology, with 45 percent in a single name.

The prescription follows directly from the decomposition, and it is not "buy better companies." The companies are fine. The structure is the problem, and the fix is exposure outside technology. A single number of 45.9 would have sent you looking in the wrong place.

Why the same portfolio scores differently

A portfolio is not good or bad in isolation — only relative to what it is meant to do. The same holdings evaluated against a conservative mandate and an aggressive one should produce different answers, because a 1.27 portfolio beta is a problem in the first case and unremarkable in the second.

This shows up most clearly in the recommended exposure. The same two-position portfolio evaluated at different risk levels returns materially different targets:

Identical holdings, nearly double the recommended market exposure. Any scoring system that ignores the mandate is answering a question nobody asked.

Using scores in practice

Three habits make scores genuinely useful rather than decorative:

Track the trajectory, not the level

A portfolio drifting from 78 to 62 over a quarter is a more meaningful signal than one sitting steadily at 62. Direction reveals whether a problem is developing or already priced in.

Read the weakest component first

The lowest component is where the marginal effort pays most. Raising diversification from 34 to 60 will move the composite far more than pushing quality from 72 to 80 — and it addresses a real risk rather than polishing a strength.

Separate structural from market-driven movement

Diversification and concentration only change when you change holdings. Trend and macro components move on their own. If a score drops without any trade having been made, the market moved, not the portfolio — and the appropriate response is usually different.

For the full response schema and every field returned, see the API reference. The mechanics of the concentration components are covered in measuring portfolio concentration risk.