Leaderboard register

Three views, one score per task

Read the field from wide to specific: the overall register, six market editions, then ten task categories. A category or market figure is the arithmetic mean of the eligible individual task scores behind it, and raw accuracy, speed and cost stay published beside it. The structure is live now; verified names and figures appear only after publication.

Preview 0.1

6 markets · 10 task families · 3 separate outcome axes.

Global register

Overall leaderboard

A system enters the overall register only with eligible coverage across all six markets. Market and category figures are arithmetic means of the individual task scores inside them; a single cross-market, cross-category overall score is not defined yet, and MICA will not invent a weighting to produce one. Raw accuracy uses a market macro-average, raw speed remains a separate successful-run measure, and raw cost per success has no global rank because currencies cannot be combined.

Accuracy

Not yet published

Awaiting verified systems

Six-market macro-average · higher is better

Speed

Not yet published

Awaiting verified systems

Successful eligible runs · lower is better

Cost

Not yet published

No global cost rank

Market-only comparison · no global rank

Six local editions

Leaderboards by market

Each market has its own ranked field, local currency and evidence lineage. Open an edition to compare systems under the rules that actually exist there.

Ten capability fields

Leaderboards by category

Category views reveal specialization that an overall standing can hide. Each view keeps the three outcomes apart and can be narrowed to one market.

Score architecture

What a published ranking column will hold

Stated in advance so the structure can be argued with before any figure exists. Nothing below has been computed: no system has been measured and no task carries a final score today.

Per-task final score: One validated, executable task attempt earns 100 × accuracy × speed × cost. Each factor is a normalized component between 0 and 1, so the score is bounded by 0 and 100.
Category score: The arithmetic mean of the eligible individual task scores in that category. There is no weighting and no geometric mean at the aggregate level.
Country score: The arithmetic mean of the eligible individual task scores in that market, on exactly the same rule as the category score.
Overall register: A single cross-market, cross-category overall score is not defined by the methodology yet. Until it is, the overall register reports coverage and the separate axis measures, and no overall number is implied.
Excluded tasks: A task that was not attempted, was not eligible, or has no pre-registered reference is excluded from the mean with a stated reason. It never enters as a zero, and the shrinking denominator stays visible.
Raw axes: Raw accuracy, raw wall-clock latency and raw evaluation cost stay published in their own units beside every score.

Detailed slice

How results will be sliced

The form submits as a normal link, so any future view can be bookmarked or shared. Selecting a slice today returns no results, because none have been published.

Ordering by AccuracyDid the system reach the task's declared final state, inside the confirmation boundary, without an unrecoverable error?

Showing Independent rerun onlyMICA re-executed the submitted system snapshot on MICA-controlled accounts and infrastructure, and holds the full evidence trace.

Taiwan · All task families

Results

No verified result is published for the selected market and task family. The controls describe the requested slice, but there is no result table or evidence record yet.