In a class of implementations, generating an event-adaptive encoding ladder for a given category of live content includes obtaining video samples of past live content (302) within that category, for performing trial encodes to generate points for a data space. Referring to FIG. 1, media content 105 contains multiple video presentations. For clarity purposes, FIG. 1 depicts four video presentations, though it should be appreciated that media content 105 contains more than the four depicted video presentations. Furthermore, FIG. 1 depicts that the video presentation 125 is included in the samples of media content 105 that are encoded using trial encoding profiles for a particular category of live events. It should further be appreciated that the number of samples of media content 105 can include more than the four video presentations shown in FIG. 1 and that samples of multiple formerly-live events of a common type (category of live event) may be used in generating an event-adaptive encoding ladder for that type of live event.

The video presentations included in the samples (obtained in 302) may be selected from a subset of media content 105 representative of the category of live event for which an event-adaptive encoding ladder is being generated. For example, media content 105 may contain thousands of video presentations of a Premier League soccer games, whereas only a few dozen or a few hundred are selected as samples for performing trial encodings when creating a Premier-League-adapted encoding ladder due to computation time limitations.