devtrackers.gg
The following details a design for how I believe the system for determining the Apex Predator ranks can be improved.
TLDR
Once you reach Masters (and complete the new trial) you cannot derank out of Masters.
Masters players earn Pred Points (PP). Pred points are the highest number of LP you earn in consecutive games up to a maximum of 20 games.
All Masters players have the same top level hidden MMR.
If one player in the team is masters, the MMR for the team will be masters level.
GOALS OF THE NEW SYSTEM
The current pred system rewards players who spend the most number of hours playing. While being an Apex Predator requires playing at a high skill level, it also requires a large number of hours, since the current system is a pure accumulation of points. Highly skilled players who are limited in the number of hours they can play have no chance of reaching Pred level.
This proposed system will still slightly favor players with a lot of hours since they will have a larger sample size to pull their best 20 games from and by playing more should in theory continue to get better at the game. However, players with less hours will still easily be able to put together 20+ games and if they play at an extremely high level, have a chance to score well and reach Pred status.
This system will also still allow people to grind if they so desire since they should always have the possibility of improving their score. The chase for top 10 should be very interesting as nobody will ever be completely out of the running, since no player can run away with the points through games played.
By having all players in masters have the same MMR and all teams with a masters player having the same MMR we ensure that any points gained will be at a similar skill level. Players of lower levels may still be pulled up into lobbies with high level players as they currently do, but a low MMR player who reaches Masters will not have the ability to get easier games through a low MMR or by bringing in teammates with low MMR.
To some degree this will also discourage Masters players from boosting lower level players since they will both be pulled into tougher lobbies and also potentially ruin a good streak of games.
DETAILED DESIGN
Upon reaching Masters a player will have zero pred points (PP). The first game played will add to their PP using the same system as diamond LP. Once the player has more than 20 games played the system will keep track of the best 20 consecutive games played in addition to the last 20 games played. If the last 20 games played is higher than the players previous best then those games will become their new pred score.
Preferably players would be able to view the points gained in both their best 20 and most recent 20 games so they can see what is needed to improve their score.
This system would give a max placement PP of 3000 points for winning every game at 150 points. Players will then gain additional bonus points for eliminations and skill as they do in the current system. Given that it seem unlikely that a team would win 20 games in a row, and that a 50 point bonus average would be pretty strong given the current system, I think we could expect the top preds to be in the 1500 to 3000 point range assuming an average of around 100 points per game for the top players.
The following example show PP earned in each of 40 games.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
100 20 55 220 90 -45 85 165 95 -75 55 45 115 168 135 25 100 80 125 85
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
65 176 125 55 35 195 135 255 10 0 -45 165 87 63 95 110 18 -75 187 55
After 20 games, the player has 1643 PP. After 22 games they increase to 1764 PP. They reach their highest PP for games 13 to 32 where they have 2004 PP. If the player was to play a 41st game they would need to beat 211 points to improve their points total.
In the case of a tie, the player who reaches that total first has the higher ranking. For example if player 1 has 1000 points in first 20 games and player 2 gets 1000 points from game 2 to 21 the player 1 is the higher ranked player.
CONCLUSION
I believe that this approach is the best way to determine who is the best player and not just those high skilled players who are willing to grind 12 hours a day seven days a week. I think that 20 games is probably a good number that provides a big enough sample size, but this number could be increased based on analysis of player data if necessary.
External link →