How it works
The match score, explained.
Every destination on Sunbound gets a single score from 0 to 100. It rewards trips where the weather actually fits the dates you asked about, and it stays honest about how much of that is a real forecast versus a long-run average.
The four ingredients
The score is a weighted sum of four ratios. Each ratio is between 0 and 1; the weights below add up to 100, so the score is always on a 0–100 scale.
- Temperature in range 55 pts
The share of your trip days whose daily high falls inside the temperature you asked for. This is the strongest signal — if a place is warm enough on most days, it stays in the running even when other components are weaker.
- Dryness 15 pts
The share of days with measurable precipitation that stay under 1 mm of rain. If precipitation is unknown for every day, dryness scores a neutral 1 so the destination isn’t penalised on a guess.
- Forecast confidence 20 pts
Real forecasts beat 30-year climate normals. High confidence (mostly real forecast) scores 1.0, medium 0.7, low 0.4 — so two destinations with the same temperatures can score differently if one is honestly an estimate.
- Proximity 10 pts
Closer destinations get more points: 1 − drive_km ÷ your radius cap. We use OpenRouteService driving distance when available, falling back to great-circle distance silently.
What the labels mean
Each card prints a one-word verdict next to the number, so you can skim a results page without doing arithmetic.
| Score | Label | What it means |
|---|---|---|
| 90–100 | Excellent match | The window is dominated by days in your temperature range, almost all dry, with high forecast confidence. |
| 80–89 | Great match | A strong fit — a handful of days drift outside your range, or it’s further than the top picks. |
| 70–79 | Solid match | The bulk of the trip works. Expect some compromises: a wet stretch, a cool morning or two, or a longer drive. |
| 60–69 | Borderline | Roughly fits your criteria but with significant trade-offs. Worth a look only if the top of the list doesn’t suit. |
| Below 60 | Below your minimum | Cards under 60 are dimmed. They don’t meet what you asked for, and we surface them only so you can see why they’re out. |
A worked example
Suppose 90% of your trip days fall inside your temperature range, 80% of days with known precipitation stay dry, the forecast is mostly real (high confidence), and the destination is 60% of the way from zero to your radius cap.
- 55 × 0.9 = 49.5 Temperature in range
- 15 × 0.8 = 12 Dryness
- 20 × 1 = 20 Forecast confidence
- 10 × 0.6 = 6 Proximity
- Total = 88 / 100 “great match”
What we filter out before scoring
- Out of radius. Destinations beyond your drive cap (km or hours) are removed. We use real road distance from OpenRouteService when possible, otherwise crow-flies with a 1.5× slack so nothing gets cut unfairly.
- Too much missing data. If fewer than half of the trip days have a known daily high, we drop the destination rather than pretend to score it.
- Ties. When two destinations score equally, the closer one ranks first.
Where the numbers come from
- Open-Meteo — daily forecast for the next 9–14 days, and 30-year climate normals for anything further out.
- OpenRouteService — real driving distance and duration, cached per (origin, destination) pair. If no API key is configured we fall back to great-circle distance and an assumed 100 km/h, and the card labels the figure “Drive (approx)” so you know.
Frequently asked
- How is the match score calculated?
- Each destination gets a score from 0 to 100, computed as 55 × (days in your temperature range) + 15 × (dry days) + 20 × (forecast confidence) + 10 × (proximity to you).
- What does “borderline” mean on a card?
- Scores of 60–69 are labelled borderline — the destination roughly fits your criteria but with significant compromises. Below 60 the card is dimmed because it falls short of what you asked for.
- Why does one card say “Forecast” and another “Climate likelihood”?
- Weather forecasts are only reliable about 9–14 days out. Trip days inside that window use a real forecast; later days fall back to 30-year climate normals, which describe typical conditions but cannot predict a specific day. The badge tells you which signal dominates, and the forecast-confidence component of the score follows suit.
- Why is a far-but-perfect destination ranked above a close-but-mediocre one?
- Temperature match is worth 55 points and proximity only 10, so weather dominates the ranking. If two destinations tie on score, the closer one wins.
- Where does the weather data come from?
- Forecast and climate normals come from Open-Meteo. Driving distances and times come from OpenRouteService when an API key is configured; otherwise we use great-circle (haversine) distance and an assumed 100 km/h average.