Y Combinator vs Andreessen Horowitz: whose portfolio ships the better APIs?

{"a":{"kind":"vc","slug":"y-combinator","name":"Y Combinator","members":792,"stats":{"kind":"vc","slug":"y-combinator","name":"Y Combinator","tier":"report","coverage":{"declared_members":799,"resolved_members":792,"scored_members":706,"agent_scored_members":706,"scored_share_pct":89,"unexamined_members":83,"enrichment_depth":{"mean_catalog_gap":80.4,"median_artifact_dirs":7,"n":706,"thinly_covered":672,"thinly_covered_of":706,"thin_coverage_note":"672 of 706 scored members had under half their catalog-attributable points satisfied at all (catalog_earned / catalog_max < 0.5). They are INCLUDED in the mean above. A large share here means the cohort mean is partly a measurement of how far our enrichment reached, not only of how these providers publish.","note":"Rubric points API Evangelist could add to the average member with no action by the providers, and the median count of artifact directories we hold on them. A cohort with a higher gap and thinner artifacts has been enriched less, so part of any difference between its mean and another cohort's is our coverage rather than the market. Compare two cohorts only when this figure is comparable. It is an upper bound on our backlog: it says who CAN close a gap, not who did."},"unexamined_note":"Excluded from the distribution: these carry a catalog-owned needs_work flag (never enriched, or no resolvable host). They are listed as members and keep their own page and score; they do not vote in the averages, because their score measures our coverage rather than their posture."},"sufficient":true,"kin_score":{"mean":21.8,"median":14.6,"min":0,"max":81.3,"stdev":18.2},"agent_readiness":{"mean":12.7,"median":0,"min":0,"max":72.2,"stdev":17.6},"band_distribution":{"exemplar":13,"strong":29,"developing":97,"thin":105,"emerging":185,"minimal":277},"facet_averages":{"discoverability":{"mean":58.5,"n":706},"contract_quality":{"mean":14.5,"n":706},"contract_governance":{"mean":4.7,"n":706},"operational_transparency":{"mean":11.2,"n":706},"developer_ergonomics":{"mean":24,"n":706},"access_clarity":{"mean":26.6,"n":706}},"adoption":{"mcp_pct":24,"mcp_first_party_pct":23,"mcp_note":"mcp_pct counts any mcp/ artifact including ones API Evangelist derived from the provider OpenAPI; mcp_first_party_pct counts only servers the provider publishes. Prefer the latter.","skills_pct":0,"scopes_pct":11,"security_pct":99,"arazzo_pct":3,"rules_pct":7}}},"b":{"kind":"vc","slug":"a16z","name":"Andreessen Horowitz","members":1040,"stats":{"kind":"vc","slug":"a16z","name":"Andreessen Horowitz","tier":"report","coverage":{"declared_members":1099,"resolved_members":1040,"scored_members":761,"agent_scored_members":761,"scored_share_pct":73,"unexamined_members":258,"enrichment_depth":{"mean_catalog_gap":78.3,"median_artifact_dirs":8,"n":761,"thinly_covered":688,"thinly_covered_of":761,"thin_coverage_note":"688 of 761 scored members had under half their catalog-attributable points satisfied at all (catalog_earned / catalog_max < 0.5). They are INCLUDED in the mean above. A large share here means the cohort mean is partly a measurement of how far our enrichment reached, not only of how these providers publish.","note":"Rubric points API Evangelist could add to the average member with no action by the providers, and the median count of artifact directories we hold on them. A cohort with a higher gap and thinner artifacts has been enriched less, so part of any difference between its mean and another cohort's is our coverage rather than the market. Compare two cohorts only when this figure is comparable. It is an upper bound on our backlog: it says who CAN close a gap, not who did."},"unexamined_note":"Excluded from the distribution: these carry a catalog-owned needs_work flag (never enriched, or no resolvable host). They are listed as members and keep their own page and score; they do not vote in the averages, because their score measures our coverage rather than their posture."},"sufficient":true,"kin_score":{"mean":23.9,"median":18.3,"min":0,"max":91.6,"stdev":18.7},"agent_readiness":{"mean":12.9,"median":2.5,"min":0,"max":66.2,"stdev":16.8},"band_distribution":{"exemplar":24,"strong":37,"developing":107,"thin":127,"emerging":214,"minimal":252},"facet_averages":{"discoverability":{"mean":58.2,"n":761},"contract_quality":{"mean":18.1,"n":761},"contract_governance":{"mean":6.1,"n":761},"operational_transparency":{"mean":14.8,"n":761},"developer_ergonomics":{"mean":26.4,"n":761},"access_clarity":{"mean":25.9,"n":761}},"adoption":{"mcp_pct":17,"mcp_first_party_pct":16,"mcp_note":"mcp_pct counts any mcp/ artifact including ones API Evangelist derived from the provider OpenAPI; mcp_first_party_pct counts only servers the provider publishes. Prefer the latter.","skills_pct":0,"scopes_pct":9,"security_pct":97,"arazzo_pct":4,"rules_pct":14}}},"comparable":true,"delta":{"kin_score_mean":-2.1,"kin_score_median":-3.7,"agent_readiness_mean":-0.2,"facets":{"discoverability":0.3,"contract_quality":-3.6,"contract_governance":-1.4,"operational_transparency":-3.6,"developer_ergonomics":-2.4,"access_clarity":0.7},"adoption":{"mcp_pct":7,"mcp_first_party_pct":7,"mcp_note":null,"skills_pct":0,"scopes_pct":2,"security_pct":2,"arazzo_pct":-1,"rules_pct":-7},"enrichment_depth":{"mean_catalog_gap":2.1,"median_artifact_dirs":-1}},"depth_confounded":true,"depth_note":"Andreessen Horowitz scores 2.1 points higher than Y Combinator AND is the better-enriched cohort: 2.1 points less catalog_gap and 1 more artifact directories on the median member. The market difference and our coverage difference point the same way, so this delta cannot be attributed to the markets on its own. Close the enrichment gap, or report both figures together."}