the SCORE() function effectively uses the low-level SOLR / Lucene scoring algorithm. The CMIS spec does not define how a Repository is supposed to score a result except that it should limit the score to between 0 and 1. Documentation for <a href="https://wiki.apache.org/solr/SolrRelevancyFAQ">SOLR relevance</a> and <a href="https://lucene.apache.org/core/3_6_0/scoring.html">Lucene scoring</a> may help gain an understanding of the general algorithm, but don't expect a formula that you can calculate on your own in most instances.