Dotabuff to Sponsor Professional Dota Team!
At Dotabuff we have received many inquiries to sponsor a professional Dota team, and we have always declined. Today, our decision has changed. Behind the scenes our developers have painstakingly worked on an evidence-based, sabermetric model to discover who are the best Dota players.
Sabermetrics started in baseball as an objective way to answer questions like "Who is the better offensive player?" Do we have the evidence to say that Dendi may be better than Arteezy (and it's not LAN wins)? At Dotabuff we have developed a propriety algorithm to suss out the best players in the Dota community. We are calling our new rating DBPR, or Dotabuff Pro Rating. Today, we are happy to announce that we are ready to sponsor the first professional team comprised of Dota's top pub players.
Everyone has the potential to become a professional player. We hope that our new statistics may set a precedent by showcasing hidden talent among the Dota pub playing community.
Through our algorithm we have assembled a professional team containing the top pub players with unique characteristics. We currently consider our roster private, pending Visa issues, but here are some statistics that guided our recruitment:
Creep Score Stolen (CSS): CSS is the measure of a teammate to steal another teammate's creep kills. This situation is also known as the Devil's Trilane, where a carry has to out compete his support and the enemy offlaner for last hits. We use this metric to rate winning support players. Higher scores represent supports with enough determination to frustrate their carries into the jungle.
Hero Diversity (HD): A smaller number represents players who specialize in few heroes. These players know every nuance and intricacy. For example, the following player was in the running to be our Wisp specialist.
With a 93% win rate on a hero that averages 39% win rate among pub players, he was strongly considered for one of the support players on team DB. Though the team may have had communication issues because he doesn't speak English, we felt it was fitting as Wisp itself can't speak at all. However, in the end, we were not able to successfully lobby IceFrog to remove the Ban phase from Captain's Mode.
Time Elapsed Before Picking (TEBP): In All Pick mode, we have found that players who pick late are often more cerebral, maximizing their time to calculate hero matchups and counters before choosing the one hero (Pudge) that will upend their opponent's synergies. An important stat for team captains.
Percentage of Nyx Impale Stuns (PNIS): A player's success with this skill shot, meaning a greater PNIS, is a determining factor in wins. However, nyx we are reluctant nyx to weigh this heavily, due tnyxo a fewnyx of Nyx's nyx faultsnyxnYxNyx.
Battlefury Bounty Hunters (BBH): Though Battlefury is often deemed a poor item on Bounty Hunter, we've found a winning pattern among a small percentage of players who buy Battlefury on BH. These players also tend to be early Rapier buyers, which has a 100% win rate in games they don't throw.
Wards Purchased (WP) The number of wards purchased in the game. Players who don't buy wards depend on high acute map awareness and game sense. We weigh this statistic heavily when recruiting our support players.
High Ground CS (HGCS): The number of creep kills from high ground, specifically from impassable terrain. This is a tactic used mostly by Nature Prophet players who cliff jungle. We find that these players exhibit resourcefulness: farming all lanes at the expense of teammates, using terrain to win versus the jungle, and ignoring calls for ganks to summon two more treants at the fountain.
Frequency of Phantom Lancer Picks (F.PL): Players who pick Phantom Lancer are generally humans of lesser character, we have found through our analysis. We consider this a "soft" statistic.
The Path to TI4
With this in mind, our staff is very excited about TI4, where we expect our team to promptly receive an invite upon this announcement. Our projections show that team DB has close to a 95% chance of winning TI4, provided that Valve creates a low priority division for the tournament.
DBPR For Everyone
We have painstakingly run our new skill algorithm on every player and are happy to share the results with you. As a way to give back to the community we have decided to add the DBPR (Dotabuff Pro Rating) statistic to every player profile:
DBPR ranges from -3000 up to 7000 and very accurately measures viability for the Dotabuff Professional Team. We expect DBPR will be the standard for measuring selfish pub play for years to come. We're also releasing the algorithm for DBPR to the public for close examination. Check out the source on Github.