seedlookout
English

OVERWORLD DISCOVERY INDEX

Minecraft seed experience calculator

See how much modeled biome and structure variety lies near spawn, with distance credit for nearby discoveries.

Model predictions · Default world generation · Selected versions from Java 1.18.2 and Bedrock 1.18.0 · Not checked in game

Java accepts seed text; Bedrock requires a number. Use the version that generates new chunks. Radius is horizontal distance from estimated spawn.

What our reference sample shows

We ran this calculator on 1,024 pseudorandom seeds in each edition, at a 2,048-block radius with default generation. The median index was 62 in Java 26.3 and 63 in Bedrock 26.50. At that radius in those two versions, your result shows how many sampled seeds scored lower, the same and higher than yours.

Discovery Index scores in our two 1,024-seed reference samplesJava 26.3: 1,024 seeds, scores 44 to 76, median 62, 10th percentile 56, 90th percentile 68. Bedrock 26.50: 1,024 seeds, scores 46 to 76, median 63, 10th percentile 56, 90th percentile 69.
Discovery Index scores in our two 1,024-seed reference samples. Number of sampled seeds at each Overworld Discovery Index score, at a 2,048-block radius around estimated spawn with default generation (rubric revision 1): Java 26.3 on top, Bedrock 26.50 below, on the same score axis. The dashed line marks each sample’s median: 62 in Java and 63 in Bedrock. Scores outside 44–76 did not occur in either sample; the index runs from 0 to 100. Model predictions from our generation model, not checked in game, and not a rarity ranking of all seeds. Chart by SeedLookout. Data: per-seed scores (CSV). © 2026 SeedLookout.

The sample is model research: it describes those seeds under rubric revision 1, not every possible world, and it is not a rarity ranking. Other versions and radii have no sample, so their results show no comparison.

How we designed the Discovery Index explains why the rubric works this way. The sample’s summary statistics and per-seed scores (CSV) are published with it.

Choose the area you want to explore

The radius is a circle around estimated world spawn, measured in horizontal blocks. Choose 512, 1,024, 2,048 or 4,096 blocks. The 2,048-block default is a starting choice for this tool; it is not a measured average for Minecraft players. A circle describes available variety, not a route you will walk.

A transparent score, from 0 to 100

Every eligible biome family and structure type earns credit once, using its nearest sample or prediction. A discovery at spawn earns 1; one at the radius boundary earns 0.5; one not found earns 0. Biomes and structures each make up half the final score.

Credit = 1 − 0.5 × distance / radius
Index = round(100 × (0.5 × biome average + 0.5 × structure average))

Biome families group related variants, such as plains with sunflower plains. Raw sampled biome IDs remain visible. Cave families use horizontal distance to a sampled biome region, with no estimate of digging effort. Ten villages still count as one structure type. The denominator includes only families in that version's model catalogue and the structure types in its reviewed criteria.

What each part of the score is worth

In Java 26.3 the structure half covers 17 types: villages, pillager outposts, swamp huts, igloos, woodland mansions, desert and jungle pyramids, ruined portals, ocean monuments, ocean ruins, shipwrecks, buried treasure, strongholds, ancient cities, trail ruins, trial chambers and abandoned camps. Older versions list fewer types, because those structures did not exist yet.

With 35 eligible biome families and 17 structure types in Java 26.3, each family is worth up to 1.43 index points and each structure type up to 2.94. A missing structure type therefore costs about twice as much as a missing biome family.

Credit falls in a straight line with distance. A village at spawn earns 1, a village halfway to the edge earns 0.75 and a village at the edge earns 0.5. Moving one structure type from the edge to spawn adds about 1.5 points to a Java 26.3 index.

A structure type that the search could not cover is unknown, not missing. Counting it as zero would understate the seed, so the calculator withholds the whole index instead.

Worked example: a published Java 26.3 seed

Dappled forest camp on a lake shore below an ice spikes field, seed 17108149081942125, is a published Java 26.3 seed. At the catalogue’s 2,048-block radius, the model found 26 of 35 biome families and 15 of 17 structure types near its estimated spawn.

The found biome families earned an average credit of 0.85, which places their nearest samples about 600 blocks from spawn on average. The found structure types averaged 0.81, about 760 blocks. The 9 missing families and 2 missing structure types earned 0.

Spread over all eligible entries, that gives a biome average of 0.633 and a structure average of 0.718:

round(100 × (0.5 × 0.633 + 0.5 × 0.718)) = round(67.6) = 68

In the Java 26.3 model sample at the same radius, 876 seeds scored lower than 68, 46 the same and 102 higher.

The index is a model prediction. The seed page’s game verification covers the features listed there, not this score.

How to compare seeds

  1. Keep the edition, generation version and radius the same. A different radius changes both the area searched and the distance credit, and versions differ in which biomes and structures exist.
  2. Read the tiles as well as the number. Two seeds with the same index can differ: one may have every structure type at a middle distance, the other most types close to spawn and two missing.
  3. Treat small gaps lightly. A difference of one or two points can come from a single structure type found nearer to or farther from spawn.

The seed catalogue can sort published seeds by their index at 2,048 blocks, and the seed map shows the same score for any seed you load.

What this calculation can tell you

Use the index to compare nearby variety under the same edition, generation version and radius. It does not measure beauty, survival difficulty, loot, structure instance counts or how much fun you will have. It has no calibrated rarity percentile. Mineshafts and other dimensions are excluded.

Surface height and spawn are estimates. Biomes are sampled on a grid, so small patches can be missed. Cave samples can lie inside solid terrain, and structure placement predictions can fail later generation checks. If a structure search cannot establish coverage, the tool withholds the overall score.

Use the version that generated the chunks. Existing chunks in an upgraded world, custom data packs, Large Biomes and superflat generation are outside these criteria. Bedrock uses the seed map's Bedrock model; its structure locations were compared with the game on nine seeds, and its terrain, spawn and caves were not.

Seed experience calculator FAQ

Is a higher Discovery Index a better seed?

Only if you want variety close to spawn. The index rewards many biome families and structure types nearby. If you want one large biome, or a single structure close by, look at those results directly instead.

Does the calculator read my world?

No. It uses only the seed, edition, generation version and radius you enter, and it models default generation. It cannot see chunks you have explored, built on or upgraded.

Has the index been checked in game?

No. It is calculated from the same generation models as the seed map. Published seed pages show game verification for their listed features; the index remains a model prediction.

Read how we designed the seed experience index, inspect destinations on the seed map, or use the seed finder to search for features you care about.

SeedLookout drafted this page with AI assistance, working from our own code and data and the facts cited here. See How SeedLookout is made for our methods and who runs the site.

← All world tools