About this site and me


  • Everything is collected with Riot’s API. It’s still a development key so the amount of data I can scrap is quite limited.
  • All data is assumed to be almost always from players at Master ranks. When a new player is added to the leaderboard I start searching their match-history, since the Riot API return the last 20 matches without any other metadata of a match, initially I may recover some matches from Diamond too, but from that point onward those games are in Master ranks.
  • All results are aimed at Master rank: play rate and win rate are relative to the reference population and results at lower ranks may differ. For Plat+ Kozmic and Dr.Lor provide better (and more consolidated) data.
  • At the start of a season I may recover all games played at Master but once my rate-limit is less than the average number of games played, my approach is to extract a random sample of games played in the current patch. I won’t give priority to the most recent games or I may easily create bias in the sample.

About this site and me


  • Everything is collected with Riot’s API. It’s still a development key so the amount of data I can scrap is quite limited.
  • All data is assumed to be almost always from players at Master ranks. When a new player is added to the leaderboard I start searching their match-history, since the Riot API return the last 20 matches without any other metadata of a match, initially I may recover some matches from Diamond too, but from that point onward those games are in Master ranks.
  • All results are aimed at Master rank: play rate and win rate are relative to the reference population and results at lower ranks may differ. For Plat+ Kozmic and Dr.Lor provide better (and more consolidated) data.
  • At the start of a season I may recover all games played at Master but once my rate-limit is less than the average number of games played, my approach is to extract a random sample of games played in the current patch. I won’t give priority to the most recent games or I may easily create bias in the sample.