AP Research: Going, Going, Gone: Betting Culture in e-Sports
Advisor: Bryan Meyerowitz
We’ve all seen the flashy plays on TV—the 4 move checkmates and the bold poker bluffs. But
does the common saying of “no risk no reward” truly hold any competitive advantage? This
study aims to assess this claim through the lens of the popular esports game Counter-Strike (CS).
In CS, two teams of 5 players each are pitted against each other. One team defends two
objectives while the other must capture either one within a given amount of time. Either team
can choose to risk the game early on with an unexpected attack or slowly and cautiously whittle
down their opponents. By measuring the impact of these decisions on the “aggressiveness” of a
team’s strategy, I will model how different teams may adapt their risk tolerance to maximize
their chances of winning. Do aggressive strategies that entail more risk net more successful
outcomes or more “reward” in CS? Could how well a team changes or maintains their
aggressive, risk-taking posture as the game progresses reflect deeper psychological workings?
After all, how well strategies work may vary from sport to sport or game to game, but how
different players respond to risk in order to achieve their goals applies to much more than online
games.
Project Blog
Post 1: An Analysis of Aggressive Strategies in Online Gaming
Introduction!
Context
Have you ever played an "aggressive" opponent in chess or perhaps been bluffed in a friendly poker match (I definitely have)? These pushy strategies are often the subject of discussion for being flashy or fast-paced. Moves like the "scholar's mate" in chess are quite entertaining if done successfully. In fact, we watch all these games and sports for moments like these where one side makes an offensive in hopes of achieving their goal. But, do these strategies actually have an impact on the outcomes of these games? For games of probability like poker, many professionals agree that an aggressive play-style is generally superior to a passive one. But, can this carry over to more complex games that don't necessarily depend on probability. Games like chess or even the popular online game Counter-Strike fit this description as an aggressive strategy isn't merely a calculated statistical advantage, but rather a conscious decision of personal preference. My project delves into the world of online gaming, but can carry broader implications in areas of game theory and even psychology.
As of now, I'm only planning on looking into chess and Counter-Strike. Chess is obviously a timeless classic while CS is one of my personal favorites. However, these games might be able to be generalized into a broader category of strategy (Can it really?). Does a football team need to decide whether an aggressive play has statistically better odds of success? I understand that the better way of answering that question would to just do a study on football, but I want to generalize it as much as possible.
Personally, I'm choosing this as my research topic because I love playing games. However, opponents that play in a fast-paced or risky manner have managed to absolutely flatten me too many times.
Inquiry
Do aggressive play-styles in popular strategy games have a statistical correlation on a game's overall success rate?
Please help me phrase this better.
Hypothesis
As much as I love seeing a flashy strategy work, I believe it offers no statistical advantage. The underlying psychology behind playing a certain way is too difficult to properly account for. Sometimes it will work and sometimes it won't. The best I can do is observe if there is a trend in the specific games I analyze. Perhaps looking into more games can help me generalize my results. But, overall I believe that while we tend to remember these strategies more, they will have little to no impact on the outcome of a game.
Oral Defense: "How did your initial exploration of the scholarly conversation lead to your final
research question/project goal?"
I'm not sure which oral defense questions I can be answering as of now, but I'll try my best (keep in mind this isn't my final research question yet). There isn't much scholarly work on this topic in general. There are plenty of articles on how poker plays aggressive, but none that can be generalized. And that's fine as poker has its own set of rules that reward risky actions anyways. There are a few articles that analyze certain aggressive strategies in their respective fields, but they're too specific. Is this a major gap because it's a dumb question?
Post 2: Preliminary Sources and Connections
It’s that time of the week again! This time, I’ll be covering some basic literature review for my project. Since no one seems to be writing about Counter-Strike in particular, most of my sources will relate to strategy games or strategies in sports in general.
Source 1: Tuathail, Gearóid Ó., and Gerard Toal. “Problematizing Geopolitics: Survey, Statesmanship and Strategy.”
You may be asking yourself how geopolitics has anything remotely to do with online gaming? Well, this source introduces three different perspectives or analogies that reflect the topic of geopolitics: geological surveying, statesmanship, and grand strategy. The second comparison is particularly important for my paper as this source compares one definition of geopolitics to chess: “To survey the geopolitical realities of the globe is to read a playing field or rather a series of different arenas of gaming.” If I can definitively map out the similarities between these two subjects, I can expand my paper’s implications to more than just esports betting.
Source 2: Plakias S, Tsatalas T, Armatas V, Tsaopoulos D, Giakas G. Tactical Situations and Playing Styles as Key Performance Indicators in Soccer.
This source focuses more on the quantitative side of strategy. This study aimed to address its own gap in the field of soccer by introducing new metrics to measure performance. By looking at tactical situations and the corresponding play-styles adopted by teams, they could gain insights into the game and predict winners. Since I’ll be doing something similar but for online gaming, I believe this serves as a decent framework.
Source 3: Dresher, Melvin. “Games of Strategy.”
This source also uses a mathematical approach to analyze certain strategy games from chess and poker to “military games as defense against attack”. Counter-Strike broadly fits this description as the objective of the game is to defend 2 objectives from an attacking force. This source claims that in an ideal and optimized situation where the stronger forces win, the solution for the attacker is to all-in on a single objective randomly. Could this mean CS favors this type of aggressive attacking strategy? Also such situations of optimally allocating defensive and offensive units can be compared to certain geopolitical situations in the real world too.
Now that a few, seemingly random sources are out of the way, let’s discuss my gap. Obviously no one is writing about Counter-Strike strategies in a professional environment. However, it can be used to hint at certain geopolitical situations (strictly speaking only ones related to attacking and defending certain objectives but). There are studies that compare chess strategy to geopolitics, but they seem surface-level at best and have little quantitative analysis. Though, the question of “Why CS?” arises. It’s a fun game. My data can really only answer one question: what is a favorable amount of aggression in certain strategies in CS and real life situations comparable to CS? However, that itself isn't too bad as the Counter Strike franchise is one of the worlds largest esports. Professional teams draft players to compete in frequent tournaments with up to $250,000 prize pools.
How did your research question evolve as you moved through the research process?
My research question wasn’t very clear cut from the beginning. Though, after seeing how many people write about chess, I decided it would probably be more favorable to only study Counter-Strike in my methodology. As such, I looked into broader implications of grand strategy. I’m currently deciding whether I should also conduct an analysis of geopolitical issues and compare those findings to my CS findings. Who knows?
Post 3: Methods For Recording And Analyzing Matches
Sources of Data
Some of you may be wondering how I can even find data on these matches. After all, who records random matches from a random video game? Could I analyze my own matches? However, since Counter-Strike is considered a competitive "esport", there are various third party sites that record official matches between professional teams (hltv.org for example). With this, I can filter through specific games that satisfy certain criteria like map type or skill level and download the associated "demo" file.
DEMO Files
Instead of just watching a tv cast or youtube recording of every tournament match and recording every small detail that occurs, it might be better to automate data collection. I turn the the game's demo system for this. Demo files (.dem files) are a native feature to the Counter-Strike that basically document every action from every player throughout an entire match and neatly package it into a exportable file. There are a couple things I can do with these files. I can either open the file up from within the Counter-Strike game itself or upload it to a third party website. However, these options are both slightly limiting. I am looking into "aggressive" strategies which I will need to define. Third party websites will certainly have no way to measure that. To fix this, I will have to write my own code to analyze the demos which is a whole other problem...
Classifying Aggressive Strategies
Firstly, I will only be looking at matches on a specific map from the game. Think of this as a unique chessboard in which each match takes place. The rationale behind only looking at a single map is because different maps influence the strategies used on them. It wouldn't be fair to compare what is or is not aggressive between a chessboard and checkers board using the same criteria.
Secondly, I will only be looking at matches played between D1 professional teams. Apparent skill gaps may influence how aggressive a player is. As such, in professional games every player will ideally have the same mechanical skill. This should mean that more weight is placed on strategy and thus the "aggressiveness" of it.
To start, I will record certain actions from each individual player throughout the game and classify them. Each classification will be assigned an arbitrarily defined value of "aggressiveness". I will also record the end result of this action as successful or unsuccessful.
Limitations
Only looking at a single map further limits the scope of my implications. Can the measure of aggressiveness be applicable to other areas other than Counter-Strike specifically? Can this measurement even apply to other maps within Counter-Strike itself? However, this does save the amount of work I would need to do.
Also, only looking at professional matches is a double-edged sword. Not all play is limited to professionals. Would similar strategies work in general or only in official matches?
Oral Defense
How generalizable is your research?
Again, this is a considerable obstacle for my project. While the competitive esports scene is already large enough, I believe there should be something more concrete than implications for a video game industry. A few areas which aggressive strategy could potentially generalize to include: game psychology, economics, geopolitics, etc. However, it still seems like a vague relationship. Even then, it could be argued that my study of aggressive strategy are only applicable to a specific map within Counter-Strike.