The resource rank: what it is, and what it isn't

We rank every Otherdeed by a single measure of their resource deposits — species scarcity weighted by deposit tier — and nothing else. Because the future importance of resource scarcity and deposit tiers is not yet known, we use a defined baseline methodology and show how every rank changes under alternative assumptions.

What we actually measure

Every Otherdeed has up to four resource slots; each filled slot holds one of 74 resources at tier 1, 2 or 3. That is the entire input. For each resource we count every deposit in the collection and weight it by tier: effective supply = n₁ + 3·n₂ + 9·n₃. A deposit is worth more the scarcer its resource, and more again at a higher tier. Add up a deed's four slots and you get its resource index; sort every deed by that score and you get its resource rank, 1 to 100,000. The counts are exact — derived from every deed and checked against a second, independent copy of the trait data, which agrees on every deed.

What this ranking deliberately ignores

Environment. Environment Tier. Sediment. Artifact. Koda. Location. All of it — a Chaos plot and a Steppes plot with the same four deposits get the same resource rank. That is the point: this number answers "how scarce are this deed's resources?" and nothing else. It is not a "which deed is better" score, it is not a price, and it does not feed the model estimate anywhere on this site.

The two numbers we chose, not measured

α = 1 controls how much scarcer resources are worth — at α = 1 a resource half as common is worth twice as much per deposit. We cannot prove 1 is right. Our published ranges run α from 0.5 to 2.0 and stop there; they exclude α = 0 (the case where resource identity carries no value) because at α = 0 all 74 resources become identical and the collection collapses from ~26,270 distinct positions to 27 — that is a different measurement, not a harsher version of this one, and sales data gives strong evidence that resource identity carries price information beyond deposit count alone. y = 3 is our baseline yield step between tiers. Yuga has not published final tier mechanics, so we test y from 2 to 4 rather than treating 3 as known. The scale constant 34,100 is not a third assumption — it is simply the effective supply of Psychosilk, so the unit price is 1.00 there; it cancels out of every rank and is verified to do so at every build. There is, however, a third structural choice we should name as plainly as the first two: the four slots are combined by summation. Addition says four modest deposits can outrank one ultra-scarce tier-3 — a max-weighted or concave aggregate would say otherwise and would produce a materially different top-100. We publish the sum because it matches the tier-weighted yield reading the rest of this method rests on, but it is a chosen aggregator, not a game mechanic. An earlier version of this page said current sales testing supported the sum over max, root-sum-square and concave alternatives; the saved diagnostic does not reproduce that order-based test (see below), so the sum stands as a disclosed design choice, not a tested result. Treat cross-deed gaps that are smaller than a deed's own published range as unresolved by this ranking.

Why every deed shows a range

Because the rank moves when α and y change, we recompute the entire ranking across a grid of values and publish the best and worst rank each deed reaches at the tested settings. The near range (12 cells: α 0.75, 1, 1.25 × y 2, 2.5, 3, 4) is a defined perturbation — not a confidence interval, and no coverage claim is made; most of its width comes from y. The stress range (24 cells: the near grid plus α 0.5, 1.5, 2) is deliberately aggressive and always contains the near range. Both are the best and worst rank over those specific cells, not bounds over every value between them: at settings inside the limits but off the grid a deed can land outside its published range (an independent recount on a finer grid found this for about a quarter of the collection, by up to ~1,400 places, with no top-250 badge lost). Of the top 100 deeds by baseline rank, 40 stay in the top 100 across the near range and 18 across stress. Use the rank as the baseline estimate, and the ranges to judge how robust it is — a deed at #40 with a narrow range and one at #55 with a wide range may be closer, or farther apart, than their headline ranks suggest. Deeds marked ◆ hold the top 250 under every stress assumption; sixty deeds do — 0.06% of the collection — and 228 hold the top 500, reported as a statistic rather than a badge.

Ties are the normal case

Two deeds with the same deposits at the same tiers are the same package and share a rank — 79.8% of deeds sit in a tie group, the median group holds 40 deeds, and the 12,739 deeds with no resources are all tied at rank 87,262. The panel says so when it applies.

What we will not tell you

We do not publish an ETH figure for the index — we cannot yet identify what a unit of resource scarcity is worth in the market with enough confidence to put a number on your screen, and two independent reviews said withhold. We do not price same-resource concentration: 299 deeds carry the same resource at tier 3 twice (one carries it three times), and only 40 of those repeat a rare resource; in the current diagnostic window (March–September 2026) the first group has 12 sales on 8 deeds and the rare group 2 sales on 2 deeds — not evidence of a premium either way. And we do not claim your rank is the true market rank — the ranges exist so you can see for yourself how firmly a position holds across the tested settings, and adjacent ranks inside overlapping ranges should never be used to argue that #12 beats #19. Supply is gross — deposits on deeds held by wallets that have never sold still count.

How stable is this rank? — what the card says

Each deed's rank card opens a "How stable is this rank?" line that states, in words, how far the rank moves across the whole tested grid: at most 2% of the collection ("moves at most #a–#b under every tested assumption"), at most 8% ("stays within…"), at most 20% ("ranges…"), or wider ("read the range, not the rank"), together with the near range (α 0.75–1.25, y 2–4) and the stress range (α 0.5–2.0, y 2–4). Deeds that hold the top 250 under every tested setting carry the ◆ badge. A tie is named as a tie, with the number of deeds sharing the rank and the positions they occupy together (competition ranking: tied deeds share the best position and the next rank skips the span). The sentence describes the rank's dependence on α and y — it says nothing about the deed's quality, and a narrow range near the bottom of the ranking mostly means there are few positions left to move to. Deeds with no deposits share a single rank and cannot move. The ranges are across the tested settings, not bounds. (Until 5 September 2026 the card carried a one-word chip — "Assumption sensitivity: low / low–moderate / moderate / high" — with the same thresholds; every top-100 deed earned "low", so the chip told the reader nothing and was retired after the Resource Rank review of 5 September. The thresholds behind the sentence did not change.)

What current sales say — a diagnostic, not an input

We check the ranking against recorded sales, but the result does not feed it. A note on what is actually on file (corrected 5 September 2026 after an independent review): the saved diagnostic is a fitted-sample regression of sale price on log(1 + index) over 3,564 qualifying sales on 2,165 deeds (7 March – 2 September 2026), with controls for environment and tier, sediment and tier, category, artifact, currency and week of sale, on a grid of y = 2, 2.5, 3, 4. It is not the order-only test an earlier version of this page described, and because the log transform changes shape with the index's arbitrary scale constant, the α it prefers moves with that constant (0.75, 1.5 or 2 depending on the scale) while the ranking itself does not. That makes it unusable as evidence for a particular α or for the "ridge" previously reported here, and we have withdrawn those claims. What the saved test does support: resource identity and scarcity matter to buyers (the tier-count-only version of the index, with no scarcity in it, fits sales far worse). An order-based, scale-invariant diagnostic will be restored, versioned and published with its sample definition and outputs; until then, nothing here changes the baseline or the bands. The sample's upper tail is thin — 4 sales on 3 deeds in the baseline top 100 — so it cannot validate fine distinctions among the leading deeds either way.

Automated data output. Not investment advice and not an appraisal. Every parameter is an assumption. The operator of this site actively trades this collection — see disclosures.