• Our Minecraft servers are offline but we will keep this forum online for any community communication. Site permissions for posting could change at a later date but will remain online.

Top Players Statistics

ALL1DO1SW1N

District 13
Joined
Sep 22, 2012
Messages
1,688
Reaction score
2,217
Whoa, has it already been over a year? I guess it has.

I just updated the Top Players Statistics sheet, using the numbers taken from today. Check the link on the first post.

I've also fixed some of the algorithms and added two new statistics:
  • Median Win Probability: Finds the median Win Probability. It's like an Average, but less weighted towards extreme scores.
  • Median Games per Player. Finds the median Games per Player. It's like an Average, but less weighted towards extreme scores.
  • Rank's Additional Wins (Top Players 2014 Spreadsheet only). Compares the differences in number of wins a rank gained from 2013 to 2014's comparisons.
Some interesting things to note:
  • The top 100 players played a total of 184020 games between 2013 and 2014. That's an average of 5.04 games per player per day for one year's time.
  • The Median number of wins is usually consistently higher than the Average number of wins by a nearly-significant amount (depending on whether you consider a 5% or 10% difference significant). In addition, the median number of games played is consistently higher than the average number of games played. This suggests that the top players contain skilled individuals who have high win ratios, and the people who compensate for a high win ratio by playing a relatively large number of games.
  • Despite claims that hackers and boosters are throwing off the leaderboards rankings, the win probability between years have been quite consistent among all tiers sampled. If there truly were some illicit force manipulating the statistics, win probability should be significantly different between years.
  • In order to stay within their tier, the top 100 players had to play a median of 2000 games in one year, the top 250 had to play 1700 games in one year, and the top 500 had to play 1500 games in one year.
  • The top ranks are significantly affected by win inflation. Rather than having the ranks being held stable by difference with a constant number of wins across the board, the win difference between years is drastically different. The percentile 25% have a consistent difference of 350 additional wins since 2013. The majority percentile up to 80% have a difference of about 400 additional wins since 2014. However, the top 100 have a phenomenal increase in difference of wins at 650, the top 50 players have 800 additional wins. This exponential difference in wins suggests that the top 20% is subject to an immense amount of win inflation, making it more difficult for newer players not in the top 100 to actually enter into the top 100 tier, and the difficulty gaining in rank increases the closer you get to the top. This statistically confirms the concerns I've heard from high-ranked players that the competitive aspect of the leaderboards is skewed due to this inflation; the high ranked players sit at the top, and the lower ranked players don't have a chance. For anyone who thinks a global stats reset would increase competitiveness, this is a good statistic to take note of.
Personally, working with the stats on the current leaderboards, I really, really want a new Leaderboards system. A lot of these statistics can be automated, and we'd be able to track cool stats like K/D ratio, distances traveled, and regional stats. Plus it would update more often than once per year.
Well done.
 

jtmboy

Diamond
Joined
Jun 3, 2013
Messages
2,097
Reaction score
1,577
This is really cool! Good job on creating this! *Pats you on the back!
 

brynemannen8

Platinum
Joined
Jun 19, 2012
Messages
52
Reaction score
42

Members online

No members online now.

Forum statistics

Threads
242,193
Messages
2,449,633
Members
523,972
Latest member
Atasci