The internals of approximate nearest neighbour search: distance metrics, graph and cluster indexes, quantization, and the recall-latency curve they all trade against.https://whatisvectordatabase.com/https://whatisvectordatabase.com/blogs/the-recall-latency-curve/https://whatisvectordatabase.com/blogs/the-recall-latency-curve/Every index parameter moves you along one curve with diminishing returns. How to measure recall honestly, and why distance concentration sets its ceiling.Sun, 26 Jul 2026 00:00:00 GMTrecalltheoryhttps://whatisvectordatabase.com/blogs/how-cluster-indexes-search/https://whatisvectordatabase.com/blogs/how-cluster-indexes-search/Partition the space in advance, then only search the partitions near your query. Why boundary effects cost recall and what probing more of them buys.Sat, 25 Jul 2026 00:00:00 GMTcluster-indextheoryhttps://whatisvectordatabase.com/blogs/quantization-trading-memory-for-recall/https://whatisvectordatabase.com/blogs/quantization-trading-memory-for-recall/Scalar, product and binary quantization compress vectors by discarding precision. What each one throws away, and why reranking recovers most of it.Sat, 25 Jul 2026 00:00:00 GMTquantizationrecallhttps://whatisvectordatabase.com/blogs/how-graph-indexes-search/https://whatisvectordatabase.com/blogs/how-graph-indexes-search/Greedy traversal over a proximity graph, why it gets stuck in local minima, and how layers and connectivity fix it — the mechanism behind HNSW.Fri, 24 Jul 2026 00:00:00 GMTgraph-indextheoryhttps://whatisvectordatabase.com/blogs/distance-metrics-and-when-each-is-right/https://whatisvectordatabase.com/blogs/distance-metrics-and-when-each-is-right/Cosine, dot product and Euclidean measure different things — except when they don't. What each compares, and why normalization collapses all three.Thu, 23 Jul 2026 00:00:00 GMTmetricstheory