Maintenance for the week of December 23:
• NA megaservers for maintenance – December 23, 4:00AM EST (9:00 UTC) - 9:00AM EST (14:00 UTC)
• EU megaservers for maintenance – December 23, 9:00 UTC (4:00AM EST) - 14:00 UTC (9:00AM EST)
• NA megaservers for maintenance – December 23, 4:00AM EST (9:00 UTC) - 9:00AM EST (14:00 UTC)
• EU megaservers for maintenance – December 23, 9:00 UTC (4:00AM EST) - 14:00 UTC (9:00AM EST)
July Results
RCubed1967
✭✭
Hi everyone,
Results first, then the discussion.
This month I played 140 games in Rubedite and finished 11th with 1758 points, as shown in the picture and the simple table of results.
The next two graphs show the score and rank changes over the month. The variability of both are why we need a change in our ranking system. Some players who are better than me should have been ranked higher (and vice versa).
The third graph shows my win and loss streaks.
Discussion
I felt the most interesting item was on items related to the starting pick coin flips. The probability of winning with the first pick (~80%) continues to be much higher than with the second pick (~45%). More interesting was the streaks of the coin flip. I started the month with 10 second picks in a row and then had another set with 11 second picks in a row. The probability of that occurring is extremely low and points to a programming issue which I discussed in a different post.
Next, let’s consider series over the months (multiple games against one player). This month I played fewer games, with a maximum of 6 games against one person.
The data becomes more interesting with the score change per series. At least this month, the series I won were all positive gains, and the series I lost were negative. Part of the issue lies is in the score change themselves. Note that my single-game wins only gained 52.3 points per game, but my single-game losses cost me 73.4 per game. Overall, I won more single-game matches than I lost, yet lost points.
The game should not be based on gains per win because a win-loss ratio can be manipulated by games played. Instead, we need to consider the larger matches. When I win 6/6, my score gains per game should be more than if I win 3/5 because the person who won 2/5 against me is likely more skilled than the person who won 0/6 against me. Further, any matches I tied (yellow) should be zero because the RNG suggests we are similarly skilled.
Results first, then the discussion.
This month I played 140 games in Rubedite and finished 11th with 1758 points, as shown in the picture and the simple table of results.
The next two graphs show the score and rank changes over the month. The variability of both are why we need a change in our ranking system. Some players who are better than me should have been ranked higher (and vice versa).
The third graph shows my win and loss streaks.
Discussion
I felt the most interesting item was on items related to the starting pick coin flips. The probability of winning with the first pick (~80%) continues to be much higher than with the second pick (~45%). More interesting was the streaks of the coin flip. I started the month with 10 second picks in a row and then had another set with 11 second picks in a row. The probability of that occurring is extremely low and points to a programming issue which I discussed in a different post.
Next, let’s consider series over the months (multiple games against one player). This month I played fewer games, with a maximum of 6 games against one person.
The data becomes more interesting with the score change per series. At least this month, the series I won were all positive gains, and the series I lost were negative. Part of the issue lies is in the score change themselves. Note that my single-game wins only gained 52.3 points per game, but my single-game losses cost me 73.4 per game. Overall, I won more single-game matches than I lost, yet lost points.
The game should not be based on gains per win because a win-loss ratio can be manipulated by games played. Instead, we need to consider the larger matches. When I win 6/6, my score gains per game should be more than if I win 3/5 because the person who won 2/5 against me is likely more skilled than the person who won 0/6 against me. Further, any matches I tied (yellow) should be zero because the RNG suggests we are similarly skilled.
2
-
rbfrgsp✭✭✭✭✭I don't think you're allowed to display screenshots of player names in the forum, fwiw1
-
RCubed1967✭✭If ZOS needs to remove that, it is fine. The results are not meant to harm or call out anyone; they were publicly available in the leaderboards.
I felt it is important to show that information so that people realize that I am not faking results. It is not about the leaderboard itself.0