Early game AFK data analysis

  1. 1. Introduction
  2. 2. Data disclaimer
  3. 3. Game modes context
  4. 4. Early game AFK : the numbers
    1. 4.1. First Blood
    2. 4.2. Players
  5. 5. Conclusion

AFK_banner

Investigation on the trend of players being AFK in very early game.

Introduction

Have you ever been frustrated by someone AFK very early in games, meaning that they do not cover the jungle or that you cannot fight head on in these ARAM bushes, and ever had the feeling that it happens almost ever game?

It is not just a feeling and I will detail my findings on this matter in this article.

Data disclaimer

All the data used in this analysis comes from the Riot Games API. It covers all games from patch 13.18 on EUW.

For ease of use, I consider a player having bought an item being a “ready” player. Of course, it is possible to become AFK after buying an item, but it is the most granular and reliable data we can get in the API. Let just say the numbers given in this analysis will be on the optimistic side.

Game modes context

Before diving straight into the main subject, I’ll give you some context on shares of game modes :

Games mode shares

Figure 1 : Game mode share by number of games

Ranked Solo games represent almost 1/3 of the games and is the most popular mode. Together with the Ranked Flex mode (the team version), Ranked games make up more than 40 % of the games. These modes are the competitive aspect of League of Legends.
Next is the ARAM mode, a fun mode where all players are stacked in a single lane to fight. It represents more than 1/4 of the games.
Normal Draft follows the same rules as the Ranked games, but without the players being ranked. Normal Blind has a different champion selection process but follow the same rules afterward. Normal games represent 1/5 of the games and are the basic casual players versus players League of Legends.
Coop VS AI is simply the same without fighting other players, only bots. It represents 1/10 of the games.
Clash is a special mode happening only during week-ends aimed at being serious and focusing more on the team aspect. It is less than 1% of the games.

Early game AFK : the numbers

Reminder 1 : For this analysis, I consider a player being ready when this player buys an item.

Reminder 2 : LoL has a mechanic preventing players from going out of the spawning area before 15 seconds (14 in ARAM). I consider the game starting at this time.

In Patch 13.18 on EUW, all players were ready to play when the game started in only 269,897 games, which is only 1.75% of the games played during that time.

Some might argue that nothing happens very early, as minions spawn at 65 seconds and 22 to 32 seconds more to reach the lanes, so it is no big deal, and that is understandable, but also wrong.

First some more stats, it takes 36 seconds to have all players ready in 50% of the games, and at minions spawn there is still a player AFK in 20% of the games.

Besides, in ARAM, there is a portal to reach the lane faster, and it is a mode that incites to fight.

At last, before minions spawn, champion can still fight, the First Blood is a great indicator for that.

First Blood

First Blood is simply the first player kill in a game. And it can happen very fast. The earliest First Blood I found in Ranked is at 28 seconds, and 23 seconds in ARAM.

In Figure 2 I represented the distributions ofTeam Ready timing and also when the First Blood happens. We can see that there is an overlap between some of the earliest First Bloods and and majority of Team Ready timings. The overlap is even larger in ARAM, as the First Bloods are very early because players tends to be ready later

First Blood and Team Ready distribution

Figure 2 : Distributions of Team Ready and First Blood timings by game mode

To be more precise, in ARAM, 23% of the First Bloods happens with at least one player missing. In Ranked, this number is 4.6%.

Players

Now, it can be interesting to know how many players are consistently on time, or in the other way, consistently late.

In Figure 3, we have plots of readiness timing by player, with a break down by game mode and by percentage of game taken.

For instance, the top right plot shows that 98.18% of the players are ready at game start in at least 10% of their games. So we can say that there are at least 1.8% of ARAM players that are “consistently late”. The bottom right plot shows that only 4% of the player are ready at game start in 90% of their game, those are the “consistently on time” ones.

In Ranked and Normal games, the number of of “consistently on time” is almost 10% while the consistently late represent only 0.5% of the players.

In any way, the vast majority of the players are in the in-between, most of them ready by 40 seconds.

Player Readiness

Figure 3 : Player Readiness across game time

Conclusion

The huge majority of games start without all players being ready, even after the First Blood happened. This can negatively impact the game, at least player behavior from having missing teammates. There are already some solutions such as remake that are in place, but it happens in less than 2% of the games, while the problem I report happens in 98% of the games. Sure there is something strange to ask to be serious in a fun mode such as ARAM, but there is certainly a huge lack of incentive to be on time in League of Legends games.

Code is available here : https://github.com/HextechLab/EarlyGameAFK