Calculating AoS Player Rankings
Some of you may remember back in August I had a failed attempt at generating player rankings based on the ELO Chess rating system. While the theory was okay in itself it wasn’t without its flaws. Firstly, it required a lot of man hours to maintain, as I had to cross-reference each player score against their opponents in each round. Secondly, it didn’t take into account that claiming a win with a faction such as Seraphon or Nurgle would be perhaps easier than claiming a win with Gloomspite Gitz or Kruleboyz. So, the system was eventually dropped with few people noticing.
The idea hasn’t gone away, and I would like to revisit it, perhaps with your own input this time around.
First of all, I believe factions need to be taken into account when scoring. But how can you do that?
Faction win rates can vary at present from 60% down to 30%, so a faction that has a 30% win-rate is expected to end a tournament with 1 win and 1 draw while a faction with a 60% win-rate should get 3 wins. Let’s level this out by taking a base value of 150 points and dividing it by the expected results for each win-rate.
|Win Rate (%)||Expected Result (5 Game Event)||Win Value|
Now let’s put this into practice. Currently Beasts of Chaos have a 61% win-rate, meaning there’s an expectation for them to claim 3 wins and 2 losses from a five-game tournament. Looking at their win rate in the above table, each win is worth 50 points and a draw half that. If a player manages to achieve 3 wins, they’ll get 150 points. Now let’s look at a Kruleboyz player. They have a win rate of 37.8% so let’s round that up to 40% and look at our table above, they should achieve 2 wins and three loses. The player goes on to achieve this and scores 150 points, the same as our Beasts of Chaos player. This is because they have both matched their factions expected win rates.
Now let’s look at what happens when these two factions achieve four wins and a loss each at the event. The Beasts of Chaos player will score 50 points for each win giving them a total of 200 points. Our Kruleboyz player scores 75 points for each win and so scores a massive 300 points placing them above the Beasts of Chaos player in the rankings. This reflects the difficulty of achieving those four wins with what is viewed as a more difficult faction to win with.
Using this method, should place all the players on a relatively level playing field, a player who goes undefeated with Beasts of Chaos will score 250 points while a Kruleboyz player need only claim 3 wins and a draw to surpass that score.
The size of the event also needs to be taken into account; this is where we change the base value of the scoring. For example, in a five round event you need 32 players to get a single 5-0 winner. Weighting for a tournament would then be applied once the scores had been calculated.
To do this, take the number of players at the event and divide by 32. Thsi will give you a multiplier to use on the final score.
For example, a Beasts of Chaos player manages four wins at a 54-player event. They score 50 points per win giving them a total of 200 points. We then take the number of players and divide that by 32 giving 1.6875. To calculate their final score, we multiply 200 by 1.6875 giving that player a final score of 337.50. These multipliers should only be used for 5 round events. If the event is 6 rounds, then it should be based on 64 rather than 32.
This scoring then reflects the number of players at the event as well as that event being able to attract more ‘big name’ players.
Timing of Calculations
The win rates for factions fluctuate and so these can’t be calculated with accuracy at the time of the event. To resolve this, each event should be calculated once the next FAQ, Battlescroll or GHB has dropped, this will ‘lock’ the win rates for those factions between those periods. For example, an event takes place on 13th August between the GHB being released at the beginning of July and the FAQ for a new battletome being dropped on 24th September. Once the FAQ has dropped, win-rates for factions are locked in that period and can be applied to the calculations for all events that happened during that time.
So, what do you think? Is there any viability in scoring players using this method?