Liquipedia, a cornerstone resource for competitive esports statistics, has introduced an entirely new ranking system for Dota 2 teams. This isn`t just another list based purely on wins and losses; it aims for a more nuanced and dynamic assessment of team strength.
Developed in collaboration with data and artificial intelligence specialists Artefact, the new ranking departs from conventional esports methodologies. Instead of a simple points accumulation or Elo-style system based solely on match outcomes, it employs a Bayesian approach. This sophisticated statistical method allows the ranking to incorporate a wider array of information beyond who simply won or lost a given game.
The system draws data from recent Tier 1 and Tier 2 tournaments, beginning its analysis from roughly The International 2023 season data onwards, though results from the very latest events are still being integrated. Key factors influencing a team`s rank include the strength of their opponents, the overall stability of the roster, and critically, the recency of matches. Newer results hold significantly more weight, reflecting the rapid shifts inherent in the competitive Dota 2 landscape, often driven by frequent balance patches.
What truly sets this ranking apart is its stated intention to incorporate individual player performance metrics alongside team results. While traditional rankings might see a team drop precipitously after losing a star player or hitting a slump, a system that considers individual contribution could potentially offer a more stable and insightful view of the underlying skill level present within a roster.
The developers acknowledge the complexities of capturing every variable in Dota 2. While future iterations might aspire to factor in intricate details like creep stack efficiency, specific item build choices, or talent selections, current data availability and technical limitations mean the system`s initial focus is on the more readily available, high-impact variables mentioned above.
Currently, the ranking is limited to the top 20 teams. The reason provided for this limitation is practical: the accuracy of the calculated strength tends to decrease significantly when attempting to rank teams further down the competitive ladder using the current model. However, the authors are open to expanding the list in the future and emphasize their commitment to continuously refining the algorithm and methodology over time. It`s an evolving tool designed to grapple with the inherent volatility and strategic depth of competitive Dota 2.
Finley Holt