With the Battlescroll release, we’ve taken the opportunity to update the Elo calculation method. Don’t worry, this won’t affect your previous Elo values and they’ll be used as the basis for the calculations going forwards.

Due to political changes Aida understandably stepped away for the foreseeable future in protest to the current events that are affecting transgender people. This has raised a few challenges and opportunities for us in regards to both our stats collection and Elo calculations going forward. We wish Aida the best for the future and hope to see her back soon! She has our support!

Stats Collection

With the loss of Aida’s data collection system, we’re going back to our old method. This however will mean that I won’t have the time to collect data for RTTs as well. It may also mean a few days delay in posting the stats to the website. We will however collect data from all of the sources we used before, but we will now also include:

• Tourneykeeper.net
• Longshanks
• Milarki
• New Recuit

We have already started gathering data and have all the GT events from these as well as BCP, Ecksen and Stats and Ladders for the last couple of weeks using the new points update. There is very little data at present, so we won’t be releasing these stats publicly for another week or so.

We also won’t be able to have the time to record each individual game result, therefore we won’t be able to exclude mirror matches from the data going forwards. The loss of individual match results takes us to the next subject…

Elo System

The Elo system is being updated for the loss of the individual matches. It will now instead be based on the overall tournament result of an event and will also taken into account the faction win rates and player elo rates that are present at the event.

The calculation will now be as follows:

Players Updated Elo = (K Factor*(Players Tournament Result-(1/(10^(((Average Elo of Players at Event)+(Average of the Faction Bias at the Event))-(Players Pre-Event Elo+Faction Bias))/400)+1)*Number of Games the Player Took Part in)))

K Factor is set at 40 for players with less than 30 tournament games, and then at 20 after that.

Players Tournament Result will be there wins and draws from the events presented as for example 3.5, where 0.5 represents a draw.

Faction Bias

Faction Bias this represents the strength of the faction globally during this battlescroll and is calculated as:

Faction Bias = 400*LOG10(Faction Win Rate/(1-Faction Win Rate))

For Example Disciples of Tzeentch may have a win rate of 68.75%, the calculation would be:

400*LOG10(0.6875/(1-0.6875)) = 136.97

At the other end of the scale Cities of Sigmar may have a 24.29% win rate, the calculation would be:

400*LOG10(0.2429/(1-0.2429)) = -197.53

What are these Faction Bias numbers used for?

The faction bias number gets added to the players Elo rating in an attempt to reflect their tournament strength. Where a strong player takes a strong faction, their Elo will be inflated and they’ll earn less rating points from the others at a tournament than if they used a weaker faction. This Faction Bias, has a larger effect on the Elo calculations when the ratings are lower than once players get into the higher ratings.

For example:

Player A who has a rating of 840.9 attends a 33 player tournament with Disciples of Tzeentch who have a win rate of 69%. They go on to win 4 games and draw their final one.

Their rating change would be:

Starting Elo 840.9+(20*(Their result of 4.5-(1/(10^(((Average Elo of all the players at the event 446.0)+Average Faction Bias of all Armies at the event 8.06))-(Their Starting Elo 840.9+Their Armies Faction Bias 136.97))/400)+1)*The number of games they played at the event 5)))

This would result in a new Elo rating of 835.5. This shows that they underperformed with the faction against a weaker field in terms of both opponents and the armies they bought. If this player had won all five games then their elo would have increased to 845.5, which would still only be a slight increase.

If the same result had applied to the player using the Cities if Sigmar rate we mentioned earlier then the same calculation would be:

Starting Elo 840.9+(20*(Their result of 4.5-(1/(10^(((Average Elo of all the players at the event 446.0)+Average Faction Bias of all Armies at the event 8.06))-(Their Starting Elo 840.9+Their Armies Faction Bias+-197.53))/400)+1)*The number of games they played at the event 5)))

This would result in a new Elo rating of 856.1 an increase on their old score rather than a loss from using a stronger faction and underperforming against a weaker field in the tournament.

That’s probably a massive amount of info to take in! So post your questions below and I’ll do my best and giving a clear answer!