Mastering Runeterra - EU edition #04 - 2021-10-02

Mastering Runeterra EU04 | ||
---|---|---|

Partecipants |
Matches |
Games |

92/94 | 218 | 476 |

Top8 Players | ||||
---|---|---|---|---|

from Mastering Runeterra - EU04 | ||||

Player |
Result |
Games Won |
Matches Won |
Line Up |

Goudaddy | 1 | 14 | 7 | Akshan / Sivir (DE/SH) - Gangplank / Sejuani - Poppy / Ziggs (BC/NX) |

Boky95 | 2 | 13 | 6 | Draven / Sion (DE/NX) - Fiora / Jarvan IV / Shen - Lulu / Poppy (BC/DE) |

lorgi | 3 | 11 | 5 | Akshan / Sivir (DE/SH) - Caitlyn / Draven / Sion - Jarvan IV / Poppy (BC/DE) |

SmoothSwoleoist | 4 | 8 | 5 | Lee Sin / Zoe - Lissandra / Taliyah - Trundle / Tryndamere (FR/SI) |

arren2398 | 5/8 | 8 | 4 | Draven / Sion (NX/PZ) - Fizz / Nami (BW/MT) - Lee Sin / Zoe |

Ghosterdriver | 5/8 | 8 | 4 | Draven / Sion (NX/PZ) - Lee Sin / Zoe - Poppy / Ziggs (BC/NX) |

SonikHolik | 5/8 | 8 | 4 | Ezreal / Vi (BC/PZ) - Fizz / Poppy (BC/NX) - Lee Sin / Zoe |

Sorry | 5/8 | 6 | 4 | Akshan / Sivir (DE/SH) - Caitlyn / Draven - Lulu / Poppy (BC/DE) |

Metadata of games collected with RiotGames API. |

As some decks have not being played I don’t have complete information

Decks for each player who took part at the tournament. Missing values are from the lack of games with other decks. Metadata of games collected with RiotGames API.

PingCity (TF/GP Bandle) and Nami/Zoe both performed way worse compared to the ladder. In the case of Fizz/Nami it may be because the results are split with Fizz/Nami (BW/MT) which is pretty much the same deck. (Simposon Paradox?)

Other notable results are BandleCity-Poppy/Demacia decks that performed better better than the Noxus version and even if the data is small and sparse among several decks the direction remain the same.

Akshan/Sivir (Demacia) and Lee/Zoe are confirmed as usual staple tournament decks with the first still being tier1 choice

BanRate from matches whose I can deduce the banned deck. Playrates from all lineUps data so including also incomplete lineUps. Metadata of games collected with RiotGames API.

**Ban Rate**: ratio between the number of bans and the number of matches of a deck.

\[\begin{equation} BanRate = \frac{\#ban}{\#match} \end{equation}\]

Example: 2 Line-Ups contained a Teemo/Ezreal deck, both played all 9 matches and Teemo/Ezreal was banned respectively 3 and 6 times; the ban rate would be \(\frac{(3+6)}{(9+9)} = 50\%\)

**PlayRate**: ratio between the number of times a deck appears among all lineUps (including incomplete information cases) and the number of all decks in all lineUps.

How the results change is I consider only lineUps with complete information or not.

Relative frequencies from all data or only line-ups with full information. Metadata of games collected with RiotGames API.

Usually the play-rates and win-rates on the ladder are highly predictable of performances on a tournament, yet this time the correlation among tournament play-rate and ladder play-rates. There is a change part of it is because of the smaller sample pool of decks or also high likely that people aren’t confident if abusing the most popular deck this time as they could be more easily handled with bans and counter-lineUps.

The correlation of play-rates is also interesting as there is almost none of it, while all the win-rates are indeed more unstable the data seems to suggest that certain decks are to be re-evaluated for a Bo3 setting meaning there may be more dark horses and some “overhyped” decks (like Ping City)

Please be aware that I use several variable to compute the LMI and not just the win-rate, this is because the LMI wants to evaluate the “overall” performance and while J4/Poppy for example has the highest win-rate the other values like ban-rate and play-rate are among the worst.

**Tier0**with LMI >= 97.5**Tier1**with LMI \(\in\) [85,97.5)**Tier2**with LMI \(\in\) [60,85)**Tier3 or lower**with LMI < 60

Note:Hovering over a circle will display a deck values.

The LMI

^{1}^{2}is an Index I developed to measure the performance of decks in the metagame. For those who are familiar with basic statistical concept I wrote a document to explain the theory behind it: , it’s very similar to vicioussyndicate (vS) Meta Score from their data reaper report. The score of each deckjust their “strength”, it takes in consideration both play rates and win rates that’s why I prefer to say it measure the “performance”. The values range from 0 to 100 and the higher the value, the higher is the performance.is not

The Gini Index is a measure of heterogeneity so, in this case and in simpler terms, how much the play rates are similar. The Index goes (when normalized like here) \(in\) [0,1] and it’s equal to 1 when there’s a single value with 100% play rate or 0 when all play rates are equal. Of course a Gini Index of 1 needs to be avoided but it’s not like the aim should be 0. As said, it’s just to add some additional tools.

Region Play Rate | ||
---|---|---|

Relative Frequencies by Inclusion Rate of a Region | ||

Region |
N |
Freq |

BandleCity | 196 | 22.48% |

Noxus | 141 | 16.17% |

Bilgewater | 112 | 12.84% |

Piltover | 98 | 11.24% |

MtTargon | 78 | 8.94% |

Demacia | 75 | 8.60% |

Ionia | 47 | 5.39% |

ShadowIsles | 47 | 5.39% |

Freljord | 39 | 4.47% |

Shurima | 39 | 4.47% |

This content was created under Riot Games’ “Legal Jibber Jabber” policy using assets owned by Riot Games. Riot Games does not endorse or sponsor this project.