Statistics on the Duration of Games (May 2016) – Over 25,000 Games Analyzed

In April 2016, Swagblanket started a discussion about the length of time it takes for one to get to legend, and how deck choice affects the climb. Of course, the conclusion from the discussion is that the climb length is a function of the type of deck you play, and of the win rate you can achieve with it. One important parameter is the length of the game with each deck.

Following the article, we published our first set of statistics based on games played in the LOE meta between February 2016 and April 2016.

We are happy to share with the community new statistics played in the new Standard WotOG meta. These numbers are based on over 25,000 games played in May 2016. The numbers we present are about the duration of games (in minutes and seconds) and the number of turns. These numbers will help you with the table that Swagblanket posted here.

Figure 1: Game turns and duration data in minutes (sorted by game duration)

Screenshot 2016-06-06 07.55.47

Figure 2: Game turns and duration data in minutes Separated for Wins and Losses  

Screenshot 2016-06-06 06.23.10

Observations:

Comparing the length statistics of WoTog to the length information we published at the end of April, for the LOE metagame, we provide the following insights:

  • There is about a 6% increase in the length of the average game. Under WotoG, the average game lasts 435 seconds (7:15) compared to 410 seconds (6:50) under LOE.
  • The fastest deck now is Face Hunter, with an average length of 05:22. This is very similar to the length it had in LOE (5:19). However, Face Hunter is not that popular, and was played only in 300 games. A deck that is gaining popularity and is as fast is Pirate Warrior, averaging 5:23.
  • The priest archetypes saw their games shorten, compared to LOE, probably because they are not very successful right now. For control priest, the difference in length between Loss and Win is 1:35.
  • When you examine the wins vs. loss length, you can see how the numbers reflect the deck’s strategies. If an aggressive deck’s game starts to drag, it is likely to lose. Control deck’s strategy is to prolong the game, and it is born out by the numbers.  

We do not present matchup (deck vs. deck) duration because many matchups did not reach a sufficiently large sample to provide meaningful statistics. We can share some information that we feel is sufficiently reliable.

  • The longest matchup with at least 30 games is the Control Warrior Mirror, averaging 18:30. The next one is Control Priest vs. Control Warrior at 15:45.
  • The shortest matchup with a sufficiently large number of games is Pirate Warrior vs. Midrange Hunter, at 4:38. The next one is Aggro Shaman vs Pirate Warrior at 5:01.  

Comments:

    • To appear in the analysis a deck had to be played in at least 300 games.
    • In Figure 1, CV is the coefficient of variation – equal to the standard deviation divided by the average. It tells you how dispersed the variable of interest is, but also allows for comparison across decks with different averages.
    • In Figure 2, the stars to the right of numbers tell whether there is a statistically significant difference between the duration (or the number of turns) for decks between wins and losses. For example, for Zoo Warlock, we can see that the losses last 6:42 and they are significantly higher than the duration of wins (6:07).  
    • The games are based on professional Hearthstone players who are part of our organization as well as friends of vS and dozens of volunteer trackers from the community at large. If you are interested in contributing to the efforts please sign up here.
    • Your game duration may vary based on your natural tendency to play fast or slow. The numbers in the tables reflect averages across hundreds of players and their styles, both vS players and their opponents. So, in that sense, the numbers should reflect your average ladder experience.  




sign up - data projects