- 6,764
- Ontario
- SvennoJ
- SvenZ
Right now SR is very easily manipulated and has become a roller coaster ride for many.
The point difference between tracks, sprint and endurance races, and high and low SR is quite extreme.
At high SR the points you lose for contact are much higher than at low SR, while at low SR the points you gain are much higher than at high SR. The idea behind this is that you'll settle around a certain score with a certain kind of driving style. However since the point scale is capped at both ends and the differences between tracks, sprint and endurance races is vast, this simply isn't happening.
What players should be ranked by is their average performance compared to everyone else, per track, per race. This is not that hard to achieve. The server knows after each race how many points were gained and lost across the board. What it needs to do is compare your score to the average to determine what rank you should be.
For example, Tokyo East outer loop had an overall average of -2.886. My average on that track on Sunday was -1.235. I did a lot of races and went down from 99 to 65 while beating the odds. Likely by a lot more as that suggests as that overall average is made from the boosted points at lower levels and more severe deductions at higher levels.
(Averages taken from https://www.jasonguernsey.net/gts/tracks/by-category)
Then I raced Tsukaba to restore SR and made 50 points in 3 races, 16.7 points per race. The overall average was 10.038 for that day. While I'm beating the odds, the massive point boost on the daily C vs the massive downgrade on the daily B is what keeps the yoyo effect going. I should not be going up two levels from 3 races, nor should I be going down 2 levels while beating the odds.
Instead of the simple plus and minus on a 100 point scale, PD should first make the points gained and lost the same no matter what SR or DR level. Then after each race calculate the global average and compare each players score to that average to determine what bracket you belong in for that race. Take a running average over the past 10 races to determine your actual SR rating. That way one race can't instantly kill your rating and the yoyo effect will be severely lessened. Nor would you have to pick what race to do to preserve or restore your rating.
Thoughts?
The point difference between tracks, sprint and endurance races, and high and low SR is quite extreme.
At high SR the points you lose for contact are much higher than at low SR, while at low SR the points you gain are much higher than at high SR. The idea behind this is that you'll settle around a certain score with a certain kind of driving style. However since the point scale is capped at both ends and the differences between tracks, sprint and endurance races is vast, this simply isn't happening.
What players should be ranked by is their average performance compared to everyone else, per track, per race. This is not that hard to achieve. The server knows after each race how many points were gained and lost across the board. What it needs to do is compare your score to the average to determine what rank you should be.
For example, Tokyo East outer loop had an overall average of -2.886. My average on that track on Sunday was -1.235. I did a lot of races and went down from 99 to 65 while beating the odds. Likely by a lot more as that suggests as that overall average is made from the boosted points at lower levels and more severe deductions at higher levels.
(Averages taken from https://www.jasonguernsey.net/gts/tracks/by-category)
Then I raced Tsukaba to restore SR and made 50 points in 3 races, 16.7 points per race. The overall average was 10.038 for that day. While I'm beating the odds, the massive point boost on the daily C vs the massive downgrade on the daily B is what keeps the yoyo effect going. I should not be going up two levels from 3 races, nor should I be going down 2 levels while beating the odds.
Instead of the simple plus and minus on a 100 point scale, PD should first make the points gained and lost the same no matter what SR or DR level. Then after each race calculate the global average and compare each players score to that average to determine what bracket you belong in for that race. Take a running average over the past 10 races to determine your actual SR rating. That way one race can't instantly kill your rating and the yoyo effect will be severely lessened. Nor would you have to pick what race to do to preserve or restore your rating.
Thoughts?
Goodluck!
