aalto1 untyped-item.component.html
Hero win rate and pick rate correlation in DOTA2
Loading...
URL
Journal Title
Journal ISSN
Volume Title
School of Business |
Bachelor's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
Department
Major/Subject
Mcode
Degree programme
Language
en
Pages
30
Series
Abstract
This thesis investigates the relationship between win rate and pick rate correlation, as well as the elasticity of pick rate in the game Dota 2, across various player ranks. The dataset utilized includes both ranked and unranked game modes, covering a span of patches from 7.08 to the most recent updates at the time of analysis. The primary aim of the research is to understand how players at different skill levels change their hero selections based on win rate changes, and whether there are observable patterns in their behaviour across time.
The study reveals that in the full dataset, the win rate - pick rate correlation is relatively weak to moderate, with a low elasticity, suggesting that players do not strongly favour statistically stronger heroes, and their selection patterns are not significantly influenced by win rate changes. As the analysis moves to lower ranks, players show slightly higher responsiveness to win rate changes, with Herald players exhibiting the highest win rate - pick rate correlation at 0.43, although their elasticity is still modest. In contrast, higher ranks such as Crusader, Archon, and Legend show decreasing win rate - pick rate correlations, alongside a steady increase in elasticity, which may reflect a growing ability to adapt hero selections more strategically in line with win rate shifts.
For ranks above Legend, particularly in the Ancient, Divine, and Immortal categories, the elasticity continues to be relatively high, but the win rate - pick rate correlation continues to decline. These trends suggest that higher-ranked players, with their deeper game knowledge, are better at picking heroes that suit different game scenarios, regardless of their win rate. The study also highlights that for most ranks, there are no clear, consistent trends in hero selection behaviour over time, as evidenced by low R values for most trendlines, which suggests that player hero picks do not follow a strong linear relationship with win rates.
The findings reinforce the complexity of player behaviour in Dota 2, with different strategies and tendencies at various levels of skill. However, the overall low R values indicate that there is significant variability in how individual players adapt their strategies, and these behaviours may be influenced by factors beyond just win rates, such as player preferences, meta shifts, or hero knowledge.
Tässä tutkielmassa tarkastellaan sankarien voittoprosentin ja valintaprosentin välistä korrelaatiota sekä elastisiteettia DOTA2 videopelissä. Tutkielman tavoiteena oli selvittää, onko voittoprosentin ja valintaprosentin välillä korrelaatiota sekä kuinka vahva tämä suhde on eri taitotasoilla ja pelin eri versioissa.
Analyysin perusteella havaittiin, että voittoprosentilla on heikko positiivinen korrelaatio sankarien valintaprosentin kanssa. Eri taitotasoilla on eroja. Esimerkiksi alemmilla taitotasoilla havaittiin korkeampi voittoprosentin ja valintaprosentin korrelaatio kuin korkeammilla taitotasoilla. Ylemmillä taitotasoilla havaittiinn korkeampi valintasuhteen ja voittosuhteen elastisiteetti, tarkoittaen että ylemmillä taitotasoilla voittoprosentin yhden prosentin muutos johti suurempiin valintaprosentin muutoksiin kuin alemmilla taitotasoilla. Vaikka voittoprosentin ja valintaprosentin välinen korrelaatio oli havaittavissa, sen selittävyys sankarien valintoihin oli alhainen, koska R^2 arvot olivat alhaisia.
Tämän lisäksi korrelaation ja elastisiteetin trendikuvaajat eivät osoittaneet selvää lineaarista muutosta pelin eri versioiden välillä. Pelaajien käyttäytyminen ei muuttunut lineaarisesti, vaan enimmäkseen vaihteli laajasti riippuen peliversiosta. Havainnot vahvistavat käsitystä että sankarien valinta DOTA2:ssa on monimutkainen prosessi, missä pelaajien preferenssit ja sankarituntemus ovat todennäköisesti merkittävämpiä tekijöitä kuin sankarien voittoprosentti.