Reference
Route metrics
What Lowest cost, Deepest and Fewest hops actually sort by in Route Radar.
Route Radar's six sample routes can be sorted three different ways, and each sort answers a genuinely different question rather than three approximations of the same one.
Lowest cost
Sorts by the smallest absolute deviation from reference, in basis points — regardless of sign. R-03 (USDC, −2.2bp) currently ranks as the single cheapest route in the sample set, ahead of every positively-priced route including R-04 (USDG, +6.8bp) — what's compared is distance from reference, not direction, so a route priced below reference and one priced above it are judged on the same scale.
Deepest
Sorts by reported venue liquidity, highest first, completely independent of price. A route can be the deepest available and still not be the cheapest — R-01 (USDC, $890K liquidity, +18.71bp) happens to lead on both dimensions at once in the current sample set, which is closer to coincidence than the norm; most routes lead on exactly one metric, not both.
Fewest hops
Sorts by how many pool legs a route passes through to reach the asset, counted directly from its path. A one-hop route is structurally simpler than a two-hop route even when the two-hop route is cheaper or deeper — fewer legs means fewer places for something to go wrong between quote and settlement, which is a property worth seeing on its own rather than folded into a cost figure.
Why three sorts instead of one score
It would be straightforward to combine cost, depth and hop count into a single "best route" ranking. State.OS deliberately doesn't. Collapsing three independent properties into one number requires picking weights for how much each one matters, and that weighting is itself a judgment call — exactly the kind of implicit recommendation this product avoids making on your behalf. Showing all three, unweighted, keeps that judgment where it belongs.