There is so much context and metadata missing (timespan, players, type of maps, how many of said type of maps, gamemodes, game, reasons for "round loss"), not to mention the difference in sample sizes. Unfortunately this data does not mean anything as much as it seems so convincing.