Player modeling using self-organization in Tomb Raider: Underworld

Abstract

Abstract-We present a study focused on constructing models of players for the major commercial title Tomb Raider: Underworld (TRU). Emergent self-organizing maps are trained on high-level playing behavior data obtained from 1365 players that completed the TRU game. The unsupervised learning approach utilized reveals four types of players which are analyzed within the context of the game. The proposed approach automates, in part, the traditional user and play testing procedures followed in the game industry since it can inform game developers, in detail, if the players play the game as intended by the game design. Subsequently, player models can assist the tailoring of game mechanics in real-time for the needs of the player type identified.

Give me a reason to dig Minecraft and psychology of motivation

“…For example, Canossa et al [11] used self-organizing maps to identify player types among tens of thousands of players of the modern action-adventure game Tomb Raider: Underworld; in a follow-up study, Mahlmann et al [16] managed to predict future behavior of players, including at what level players stop playing the game, based only on early-game player behavior.…”

Player Types: A Meta-synthesis

“…A study conducted by Drachen et al (2009) looked at how a set of players completed the popular adventure game Tomb Raider: Underworld. They identified four different styles each with different playing patterns and solutions to specific problems and also a certain level of performance.…”

Controlling the crucible

“…The performance of each model was described, and sets of 4-5 playstyles identified across each model. The authors concluded that Archetype Analysis (AA) [13], [12] performs best in terms of developing clearly separated and explainable profiles, the latter forming a key quality criteria in games-based behavioral profiling as argued by Drachen et al [10].…”

“…A substantial number of papers have been published on behavioral profiling in games. The first paper to specifically utilize behavioral profiling in commercial game titles was Drachen et al [10] who worked with Self-Organizing Networks. The majority of previous work is focused on employing cluster analysis or segmentation methods, but comparative analyses were provided by Bauckhage et al [11] and Drachen et al [12].…”

“…Players customize their characters by allocating points to each of these base stats to complement their class and game-play style. After analyzing the results from k-means for 3-5 clusters, the 4 cluster results were chosen to be the best balance between prediction and interpretability of clusters [22], [10] 2. Each cluster was assigned a profile to reflect playstyle.…”