Ranked Changes

• Ranked badges have been collected for all Deadlock players.
• There is 106,917 players who have received a rank
• NekoScore leaderboard has been replaced by the real ranked leaderboard. The leaderboard is currently unordered, everyone who is Eternus VI is tied as rank 1 within their region.
• Tracklock now shows the average Rank on all games tracked and the ranks of the other players you are playing against (even in unranked games).
• You can view the rank distribution here . You can filter the data on this page by region and week. It will automatically update every week as new ranks are released.
• There is has been some confusion on reddit/discord that this not all of the player rank data. I am extremely confident that this is 99.99% of players with a ranked badge.

Valve Match Score

• Valve has changed the match scores that Tracklock used to calculate NekoScore (again), we have temporarily removed NekoScore.
• The new 50th percentile of the new Valve Match Score is around 1250
• For most players you can expect your match scores to be ~250 lower than they were before
• NekoScore may come back in the future once the scores adjust to the new lower levels. Come have a discussion on discord if you would like NekoScore back or have any suggested changes.

Current Match Scores

Previous Match Scores

Tracklock Live

How Accurate Is Tracklock?

There was a interesting reddit thread this morning where people were asking about how accurate NekoScore is https://www.reddit.com/r/DeadlockTheGame/comments/1g9s8si/ranks_are_out_lets_see_how_accurate_tracklock_is/

The general concenus from that thread was that its not very correlated so Mark & I decided to do some analysis on the data.

Tracklock has been internally calculating a seperated Ranked NekoScore for all players this week that has not been visible on the website. This is the data we will use for the analysis rather than players unranked NekoScores.

• There is a clear directionality between NekoScore and the players rank.
• This correlation is particularly evident in the median scores for each rank tier.
• While the correlation is strong, there exists significant overlap in NekoScore distributions between adjacent ranks. For example, a player with a NekoScore of 2,000 may fall into any rank between Emissary 1 and Phantom 6.
• The substantial overlap between ranks indicates that NekoScore alone cannot serve as a precise rank predictor.
• However, it effectively establishes broad skill ranges and can identify significant rank mismatches.
• The system appears more effective at identifying skill extremes than distinguishing between adjacent ranks.
• The predictive accuracy of NekoScore improves notably at higher skill levels.
• NekoScore had much less variability before the recent match making changes.

To sum it up, I would say NekoScore is not that accurate anymore.

The 1200 NekoScore Eternus 6 Player

• We have identified an extreme outlier where a player who finished the week with 1200 NekoScore was placed into Eternus 6.
• This player finished all games with scores of ~40 kills 0 deaths and was using hacks to fly under the map and kill the other team.
• This player did not play a large amount of games.
• This data point is interesting because it means other factors than win/loss and starting mmr are taken into account to determine your rank. This may only be the case for placement games but currently I am unsure.

Have any questions or feedback? Join our Discord and someone will be happy to help you out.