# Earth-likeness ranking methods

Data retrieved 2026-09-23 from the NASA Exoplanet Archive TAP service.

## TAP query

```sql
SELECT pl_name, hostname, disc_year, discoverymethod,
       pl_rade, pl_radeerr1, pl_radeerr2, pl_radelim,
       pl_eqt, pl_eqterr1, pl_eqterr2, pl_eqtlim,
       sy_dist, sy_disterr1, sy_disterr2,
       st_spectype, pl_insol
FROM pscomppars
ORDER BY pl_name
```

The `pscomppars` table supplies one composite row per confirmed planet. Its fields are more complete than a single-reference solution but may combine values from different publications and therefore need not be physically self-consistent.

## Benchmarks and metrics

- Size deviation: `abs(ln(pl_rade / 1 R_earth))`
- Temperature deviation: `abs(ln(pl_eqt / 255 K))`
- Distance: `sy_dist` in parsecs, minimized directly
- Non-positive values, missing values, and explicit radius/temperature limits were excluded for that metric.
- Combined rankings used only the 5,763 planets with all three usable values.
- Each metric was converted to a percentile rank across the complete-case sample so size, temperature, and distance contributed on comparable scales. Lower is better.

## Weighting profiles

- Balanced: size 1/3, temperature 1/3, distance 1/3
- Earth analogue: size 45%, temperature 45%, distance 10%
- Exploration/distance: size 15%, temperature 15%, distance 70%
- Size first: size 70%, temperature 15%, distance 15%
- Temperature first: size 15%, temperature 70%, distance 15%

These rankings are descriptive, not estimates of habitability. Equilibrium temperature is a black-body model and does not measure surface temperature; radius values may be measured or model-derived; proximity says nothing directly about life. Atmospheres, stellar activity, composition, water, and observational uncertainty are not included.
