Analysis of the Starladder Major Berlin 2019 Quarterfinals: Astralis vs. Liquid

By Joshua Goodman

Introduction

“It’s a shocking defeat for a team that has had an otherwise stellar 2019” were the words that came out of commentator Jason “Moses” O’Toole. The number 1 team in the world and clear favorites to win the Fall Major for 2019 in Counter-Strike:Global Offensive, the American side of Team Liquid, had just been stunned in a 2-0 loss to their long-time rivals from Denmark, Astralis. Before the remainder of the playoffs even continued most were calling this match the true final of the tournament, even though it occurred in the quarterfinals. Team Liquid were ranked #1 and were on a 6 event win streak where they had won 24 straight Best of 3/5 matches, the best any team had done in over 5 years. For them to do anything but win this Major would be a disappointment. Astralis, on the other hand, had suffered a dip in form during 2019. The 2-time reigning Major champions had started 2019 off with a bang winning the first Major of the year in convincing style, showing the world they were still at their peak. However, directly following their victory they would go on a cold streak, unable to repeat their unprecedented success from 2018 and took many months off for mental health and burnout reasons throughout the year. Going into this Major they were a large unknown to many experts. Which Astralis would show up?

This paper will examine an economic analysis of this quarterfinal between Astralis and Liquid. Topics that will be addressed include tournament theory, a look at league governance for Counter-Strike:Global offensive and how it compares and contrasts to “traditional” sports, and finally a deep dive into the game theory and strategy behind the map selection process before the match, specifically Astralis’s decision to pick Vertigo as the first map. As most are unfamiliar with the game of Counter-Strike:Global Offensive (CSGO) an executive summary is directly below for some basic insights into the scoring and outline of the match.

CSGO Summary

CSGO is a First Person Shooter game from Valve Software. It is a 5v5 game with no substitutions. The game is a Best-of-30 rounds with each team being either the Terrorists or Counter-Terrorists. After Round 15 the teams switch sides and prepare for the second half. A team becomes victorious if they reach 16 rounds won during regulation. Should a tie of 15-15 occur, an overtime would begin with shortened, 3-round halves. If a team wins 4 out of these 6 rounds they will win the map, if the overtime ends in 3-3 split another overtime will commence. This continues until a team has won as there are no draws in CSGO. A typical playoff match in a CSGO tournament is a Best-of-3. This is akin to Tennis in some regards. For a team to win a match they must win enough rounds to win the map, and win 2 out of 3 maps to win the match. For professional play there are 7 possible maps that can be played in a series. The teams conduct a pick-ban selection process before the match to decide the maps that the match will unfold on.

Tournament Theory

To start any conversation on tournament theory it is important to understand the Principle-Agent Dilemma and who the principals and agents are in this competition. The Principle-Agent problem occurs because a principal cannot effectively monitor an agent at all times to ensure 100% effort and quality by the agent. Therefore, the agent has the opportunity to make mistakes or act against the principal's best interests and in their own self-interest instead. A way to attempt to solve this dilemma is for the principal and agent to enter into a contract which aligns the agent’s incentives with the interests of the principal. This gives direct motivation for the agent to perform to the best of their abilities and enact the principal’s vision. In professional CSGO the principal is the ownership group of the team and the agents are the 5 players competing in a team. When looking at this specific match there are 10 agents, the 5 players on Liquid (Elige, Nitro, Stewie, Twistz, Naf) and the 5 players on Astralis (Device, Dupreeh, Glaive, Magisk, Xyp9x). For Liquid the principals are the co-CEO’s of the team, Steve Arhancet and Victor Goossens. For Astralis the principals are their co-CEO’s, Nikolaj Nyholm and Anders Hørsholt. However, it’s important to keep in mind that for both teams the CEO’s are also agents with the shareholders being their principals who put their trust in the CEO’s to enact their vision for the company. Team Liquid is a private company with many investors including NBA legend Magic Johnson, Steve Case, and others. Astralis on the other hand is a public company traded on Nasdaq’s First Growth Market Denmark.

To properly incite effort the environment needs to allow for effort to directly impact productivity. In CSGO, this is the case. The more hours spent aim training or studying opponents past matches, the greater the performance in game tends to be. Another big factor in inducing effort is the difference between winning and losing a competition. In this match the losing side would leave with only $35,000 while the winners would secure at least $70,000 and play to win a total of $500,000 should they come in first place. This is a minimum 100% increase in prize money. In a 2 party competition with higher payouts for the winner, the incentive to win is higher which will lead to a higher investment of effort by the parties to maximize their probability of winning the competition (Lazaer & Rozen, 1981). In addition to this, it has been shown that the bigger the spread between winning and losing, the larger the investment by the players. This effect is amplified even higher as a competition moves towards the final few matches or days (Ehrenberg & Bognanno, 1990). The matchup in question was an elimination match in the quarterfinals of one of the largest monetary events on the professional CSGO calendar, meaning that not only was there a large monetary difference for winning or losing, but that the contest was in its later stages with even more incentive to exude more effort. Something unique to esports and CSGO in particular is how player’s contracts are incentivized. Not only do top level players have salaries between $20,000-$50,000 a month, but it’s commonplace for the players to get 100% of all prize money won, split evenly between all players (possibly a smaller share going to the team’s coach if they have one). Therefore in this situation the players are even more incentivized to win more prize money as they will keep all of it without any split to their organization.

In addition to the $1,000,000 prize pool the 8 teams that make the playoffs in CSGO Majors are automatically invited to the next Major as long as the retain 3/5 of their current roster. This invite is akin to the PGA Tour “exempt list” on a much smaller scale. With the PGA, players who have won any major tournament within the last 8-10 years were eligible to enter any tournament they pleased (Ehrenberg & Bognanno, 1990). Since CSGO Majors are the pinnacle achievement in the scene, instant invitation is worth its weight in gold and therefore further incentivizes teams to reach the playoffs. Furthermore, by securing an invite to the next Major, each player on the roster has greatly increased their own bargaining power. Since teams must retain 3 out 5 current players, each player on the team has become exponentially more valuable and the risk of being cut and becoming a free agent drastically decreases with the securement of an invitation to compete at the next Major.

Another element of tournament theory is the need to sort through all possible players to have only those with the best skills to reach the top teams. In CSGO there are no minor leagues or farm systems that allow players to truly work their way up and ensure they are talented enough to compete against the best teams. Instead, the rise of young talent is done in one of 2 main ways. The first and more common is where players rise through the ranks of pick up game ladders and attract the attention of top teams. These games are not professional play but more akin to playing 5-on-5 pickup basketball and live streaming it to grow one’s brand. The more marketable and well known these players can become, the higher chance they have to create connections with current top level players, who have large influence over their team’s roster decisions. The second way to get to the top is to win your way there. Since CSGO is completely open-circuit, it is possible for any group of 5 players to band together and win any tournament in the circuit. This is less common as the best talent from up and coming teams are usually poached by larger organizations before they can reach this point. However, this open circuit allows for self-selection of the best talent as the only way to reach these high level events is to field top teams. Since there are no franchised spots, a team that decides to spend less on talent and field a lesser team will no longer qualify for events and be resigned to play smaller events against worse competition, out of the spotlight that top events have. Similarly, because there are next to no trade deadlines or contract restrictions, player’s are incentivized to put in their maximum effort throughout the year as they can quickly be replaced from an ever-growing talent pool. This job insecurity coupled with the monetary motivation of winning guarantees that top level professional players that want to keep their spot on a team will consistently be putting in the required hours to train and invest in their skills to constantly perform at the best of their abilities.

The organizer of this contest, Starladder, wants the best matchups between the best teams to occur later in the tournament in order to maximize viewership. To increase the chance of having the best teams competing in the playoffs Starladder modified the Swiss System group stage from chess and added in an elo rating system created by Arpad Elo, after which the playoffs would be played in single-elimination Best-of-3 matches. In this format teams are initially seeded by their average world ranking over the past 5 months according to the CSGO news website HLTV.org. The first round matchups between the 16 teams are based on these seedings, so #1 team played the #16 team, the #2 vs #15 and so on. Using an elo rating system based on the strength of the opponent all the teams have their ranking slightly adjusted due to their victory or defeat. Subsequent rounds following the initial rounds are “power-matched” where teams can only face each other if they have the same record (Hua, 2017). So after the first round in this contest 8 teams are 1-0 and the other 8 are 0-1. Now the highest ranked 1-0 team will play the lowest ranked 1-0 team, the highest ranked 0-1 team will play the lowest ranked 0-1 team. In addition, a team cannot play the same team more than once in the swiss system (Ólafsson 1990). This format continues for every team until they have either won 3 matches and advance to the playoffs or lose 3 matches and are eliminated from the tournament. Finally, to ensure the best teams would advance and the worst would be eliminated, all elimination and promotion matches were played in Best-of-3 format, which is typically reserved for the playoffs. This was so that if a team got a bad draw early on and went down 0-2, they could show they were superior in the Best-of-3’s and have a chance to win 3 straight matches and qualify for playoffs. Likewise, this made it so no team could make it to the playoffs without winning at least 1 Best-of-3 series. These series induce different efforts from the players at different points. It was found that in tennis matches the winner of the first set exerts more effort than their opponent, on average, during the second set and if the match goes to a third set, both competitors exert equal effort (Malueg & Yates, 2010). In this group stage, 8/13 Best-of-3’s ended 2-0, meaning the winner of the first map carried the momentum into the second for a sweep, following Malueg & Yates’ findings.

For Astralis and Liquid, who came into the tournament seeded #1 and #2 respectively (although Astralis had been slumping while Liquid was the clear favorite, Astralis had dominated the rankings in the first half of the year due to previous results and as such did not leave the #1 spot on the world rankings until May when Liquid dethroned them and began their own win streak), neither escaped the group stage with a perfect 3-0 record. Liquid was defeated by North-American rival NRG in the second round 1-0 matchups, falling to 1-1. There, they were shocked in overtime by huge underdogs AVANGAR from Kazakhstan, who came into the tournament ranked 14th out of the 16 teams competing. Now down to 1-2, Liquid were fighting for their tournament life and were able to claw out of the hole following back to back 2-0 victories against Danish side North and international squad Mousesports. Importantly, due to these losses and the structure of re-seeding teams after each round, Liquid entered the playoffs ranked #6 out of the 8 remaining teams, which meant they would face the #3 seed in the quarterfinals.

Astralis was able to avoid the early struggles that plagued their rivals in Liquid and reached the promotion 2-0 match without much issue. They did take a loss in this round against NRG but bounced back with a 2-1 win vs the Serbian team CR4ZY to advance to the playoffs with a 3-1 record. Due to this 1 loss Astralis were based on the elo seedings and were slated into the #3 spot, setting up a date with Liquid in the quarterfinals. It is vital to understand this information as the unique format of this tournament is what allowed the #1 and #2 ranked teams to face each other early on in the playoffs, a matchup that most spectators would have hoped to be the grand final.

League Governance

Traditional Sports

In traditional sports the league organizer aims to have the best product to sell to consumers and therefore are incentivized to have exciting matchups and more parity than dominance in a league. The greater the competition, the greater the profits are for the league and the teams in it in the long-run. This goes against a typical firm’s goals to achieve a monopoly to maximize profits due to the Louis-Schmelling Paradox, whereby it is in the best interests of the teams to create a cartel as sheer dominance by one team will lead to burnout of their audience and decreasing interest in each game (Neale, 1964). Teams can become dominant and become exciting but if their reign lasts too long without contest, fans will lose interest in their games and the league. Sports leagues take strides to achieve this parity thought competitive balance acts. These include salary caps, drafts, revenue sharing, and more. The end goal for all parties is to have their league be entertaining enough to fill stadiums throughout the year and garner large television audiences for every broadcast. Teams benefit from this via gate receipts from their home and away games, while the league itself massively benefits from large network television contracts to the rights of the league.

To achieve this goal, the competitive balancing measures must allow the league to be uncertain. In order to reach higher levels of uncertainty the league must have an equal distribution of top talent throughout the league (Rottenburg, 1956). As teams move towards being homogeneously talented, the uncertainty of each matchup increases up until it is a 50/50 tossup of who will win any single game. This is more hypothetical as Rottenburg mentions how there will always be teams with larger home markets or revenue and they can afford to pay top talent more than small market teams, leaving true uncertainty out of reach. In professional sports leagues across North America and Europe, competitive balance is enforced by an overarching governing body. For European Football this body is UEFA, which in turn is just one part of FIFA which governs all of international soccer. For closed franchise leagues such as the MLB, the MLB itself is the governing body which enters into a collective bargaining agreement with the MLB Players Association in order to effectively control the league. In these agreements both players and teams must agree to guidelines on topics such as revenue splits, free agency rules, contract minimums, salary cap, and more.

In North American sports the market for players has many constraints. For a team to acquire another player they either draft them, win the bidding war during free agency, or make a trade to acquire the play from another team. In soccer there is also the ability to purchase a player directly from another team for a set buyout price determined in their contract. Player’s rights have generally increased in sports throughout the late 20th century into today and have come a long way from the reserve clause of the early MLB where they were fully at the whim of their owners.

The market for CSGO players is based on that of soccer with some major differences. Just like soccer, CSGO has a large market for buyouts where a larger team will purchase the rights to a talented player to enhance their team. However, in CSGO, there is no union that has reached a CBA with the leagues running the events. This is due to many factors including top organizations acting as a cartel, the large pool of players willing to take cheaper salary for the opportunity to play, and the divided nature of esports with teams based across North and South America, Europe, Asia, and Australia. Experts believe that as time goes on and esports continues to grow a player’s union will rise and collective bargaining agreements will be formed and esports will start to mature (Ridenhour, 2020). But for now, CSGO players are unprotected and their leagues remain unbalanced. And without this CBA there exists no competitive balancing mechanisms such as a salary cap, draft, or any real governing force to balance the teams to create parity. Teams with enough capital can purchase the top players from all other teams if they wanted to and this has happened in the past. This had led to the HHI for CSGO Majors to be above 1500 as will be shown later.

While European Soccer is more of an open circuit than the franchised leagues of North America, CSGO embodies a much more pure iteration of the open circuit ideals. The developer and publisher of CSGO is Valve, who own the game. Unlike any other professional sports league, this company owns the IP of the entire sport. This means that if they wished to have unilateral control across the globe with no restrictions, they could do so. However, Valve is very famously laissez-faire about esports and controlling the circuit for their own games. Valve has stated that no company needs a license to host and operate a CSGO tournament with any teams in the world. This allows for a free market to operate for tournament organizers who want to enter the scene. The only event that Valve directly controls are the Majors. The Major was introduced by Valve in 2013 in order to reinvigorate attention around CSGO. Valve themselves funded the prize pool for the first Major, at a then impressive $250,000, and have done so for each of the subsequent 14 Majors that have occurred. In addition to giving out this prize money Valve have also created in-game cosmetics for all the teams and players competing in each Major. These include virtual stickers of team logos and player signatures that can be used in game, of which a percentage of each sale is given directly to the player or team that the item belongs to. For the Starladder Major Berlin 2019 Valve paid out over $11 million to players and teams from these cosmetics alone. Valve themselves do not organize the event though, instead they award complete control to independent organizers from a number of pitches. Similar to how FIFA awards the World Cup, these pitches are judged by country location, prior tournament organization experience, and scale. The company that wins this pitch is given near full control of the Major. They are able to sell the broadcast rights, receive 100% of ticket sales, as well as sell sponsorships for segments of the event and even a title/presenting sponsor. The Fall Major 2019 was awarded to Starladder, a Ukrainian company founded in 2001 which is predominately known for its events in the CIS (Commomwealth of Independent States) area. Starladder had already hosted over 20 top tier CSGO tournaments with great success. The distribution of which company receives the chance to host and run a Major is another issue to competitive balance. The organizer who runs the Major is given control over the most important segment of the CSGO competitive calendar and as such could purposely schedule qualifiers for the Major over their competitors event dates to increase their own market share.

When looking at how balanced CSGO is there is an important distinction that needs to be made from traditional sports, the idea of a team. In traditional sports the organization is the ownership and brand that pays the players, but in CSGO it is more common to associate the “core” of a team’s players as the team. For instance, in 2016 an upstart group of 5 Brazilian players won both Majors in a show of dominance. However, after their first victory they gained massive leverage to look for better contracts and the entire team of 5 was purchased by a different company which they competed under. The players were the same, the fans followed the players and not the organization. So when looking at competitive results it makes sense to consider 2 groups the same “team” if they share 3 out of 5 players in common. Now, dissecting competitive balance is slightly harder in CSGO as we cannot really look at win% as due to the open nature of the scene, a team can have a really large win percentage but play only a few top teams. Instead we will look at HHI of the Major championships to understand the parity in competitive CSGO. Table 2.0 below showcases both the HHI for the winners of each Major and also the tournament organizer who was chosen to host. Teams that played under different organizations but we consider the same “team” are numbered by their Core.

From this table we can see that the market for Major winners is slightly unbalanced but not towards a complete monopoly with an HHI of 1639. This shows that within the free market of CSGO, the Majors (which are the equivalent to a sports world championship), are not dominated by a few players and have more parity than some major sports. The more startling number is the HHI of organizers who have been chosen to host the Majors at 3954 which would be considered a highly concentrated market. The biggest player in this space is MTG (Modern Times Group) which owns 2 different tournament organizers, Dreamhack and ESL. This is concerning as MTG has hosted 60% of all Majors ever which means they have a strong market dominance over the CSGO competitive calendar.

Other organizer’s have tried to chip away at MTG’s hold on CSGO to little success. In 2020 3 different organizers have started to create revenue shares with the team’s competing in their leagues to incentive certain organizations to prioritize their events over their competitors. In February 2020 MTG met with the top teams across the globe in Paris in order to come to an agreement on how they can organize their event structure to the teams and players needs and requirements. Titled the “Louvre Agreement”, the meeting outlined that top teams have the ability to become partnered with MTG. Being an MTG partner entails automatic invites to every MTG event on the calendar with the requirement that they attend a certain amount each year to ensure each event has a high level of competition. These partner teams also gain power over the leagues that MTG operates with any new partner teams requiring 2/3rds approval from existing partners before being admitted.

With this agreement, partner teams will split 25% of operating revenue from MTG leagues before expenses, representing the first time that competitive CSGO has moved towards competitive balancing measures and structure of more traditional sports leagues. In addition these revenue shares are not equal and instead each partner’s individual revenue share is based on the number of events attended, competitive results and viewership. This provides further incentive for organizations to field highly competitive and marketable players to maximize their potential profits from MTG leagues.

Finally, the players for these top teams have come together to form a proto-players union called the Counter-Strike Professional Players' Association (CSPPA). The CSPPA was present during the Louvre Agreement, giving the player’s a seat at the table for the first time in CSGO history. Through this new agency players are now entitled to receive 15% of the partner’s revenue share for MTG events. Now CSGO players are not only increasing their payouts by winning prize money but also from the event they are attending reaching higher viewership and driving more revenue. This in turn motivates the teams and players to exert more effort during media days and other obligations while attending the event.

The Louvre Agreement represents the dawn of a new era in competitive CSGO. As esports continues to grow every year, CSGO needs to adapt to the new attention that has been shined into their world. With this agreement teams, players, and organizers have created mutually beneficial relationships that motivate all parties to cooperate in order to maximize their own profits (Louvre Agreement. 2020).

Other Esports

It is important to note that other esports leagues have been ahead of the curve on CSGO in terms of moving towards traditional sports models. Other large publishers such as Activision-Blizzard or Riot Games take a more hands on approach and have franchised many of their own games with a slot costing upwards of $40 million for one team. As esports has grown globally these publishers and other tournament organizers are looking to transition to a more traditional competitive model based on that of the North American franchised leagues. This allows the publisher to increase revenue by charging franchise fees in the millions to enter their closed league as opposed to having a free market that operated before. This also appeals to the larger esports organizations who can afford to purchase these franchised spots as they no longer have to worry about qualifying to events or needing a top ranked team to maintain relevance. Since their business model is based on sponsorships, having their team getting guaranteed air time and viewership is preferred to needing a winning team to get air time.

These leagues also include collective bargaining agreements which in theory should increase the leverage and power for players in the league. However, the player’s union that represents all professional players in each of 3 current franchised esports leagues (Overwatch, Call of Duty, and League of Legends) is also owned and operated by the publisher who is running the league itself (Melrose, 2018). Esports is ripe with conflicts of interest and clearly the players in these leagues have not yet created a coordinated effort to make their own union and instead are now at the complete control of the publisher.

Valve have made it clear they have no interest in going down this route and want CSGO to maintain an open circuit with any organizer able to host a tournament should they have the capabilities. They have fought against circuit exclusivity by issuing statements like the following as a warning to those who wish to restrict a team's ability to play in their competitors events, “At this time we are not interested in providing licenses for events that restrict participating teams from attending other events” (Valve, 2019). With this in mind the future of the competitive Counter-Strike events will be a battle to take control of the tournament market with MTG currently controlling over 60% of the calendar and no other organizer being able to compete on the same scale and volume.

Strategy

There are many different ways to analyze the strategy of a CSGO match. From Counter-terrorist side setups, to full site executes, fake executes, clutches, flashes, and more. But for this particular match arguably the most important strategy took place before the players even entered the server with the map selection process.

In CSGO there are 7 competitive maps in the pool at any one time. This means that in any match you could potentially play any of the 7 maps. Each team has strengths and weaknesses on every map and can have extremely varying win rates for each map. During the Starladder Major Berlin 2019 the 7 maps in the playing pool were: Dust2, Inferno, Mirage, Nuke, Overpass, Train, and Vertigo. Liquid and Astralis’s respective win rates and stats for each map factored into the map veto preparation.

The map selection process for a competitive CSGO Best-of-3 is a 6 step pick/ban process whereby 3 maps are chosen to be played in the series. Before the selection, one team is assigned to be “Team A” and the other “Team B” this can be decided by coin flip or possibly by whoever is the higher seed in the tournament. This typically doesn’t have a huge impact but does change which team goes first in the selection process. For the Majors the veto process goes in the following order.

1. Team A bans a map
2. Team B bans a map
3. Team A picks a map, Team B picks which side to start on
4. Team B picks a map, Team A picks which side to start on
5. Team B bans a map
6. Team A bans a map and the leftover map is made the decider if necessary

Liquid’s map pool statistics up to their Starladder Major quarterfinal faceoff with Astralis are shown in Table 3.0 below:

These stats reflect Liquid’s incredible dominance over the Spring and Summer of 2019 in CSGO. Their worst map, Nuke, had a win rate of 64.3% and they had won their last 5 matches on Nuke coming into the Major. The only map they don’t play is Train and due to the format of the map selection, they can have Train as the “permaban” to ensure it never comes into play. This left 6 maps at an elite level of play and had won their last 6 tournaments, which led to them having a 24 Best-of-3/5 winning streak, one of the longest in CSGO history. Their other active winstrikes included 13 in a row on Mirage, 9 in a row on Inferno, and the 100% win rate on the 5 matches they’ve played on Vertigo. In the last 55 maps Liquid had played up to this match, they had only lost 9 maps.

For Astralis, who were not as active in the Spring and Summer months of 2019, their map pool statistics for the 4 months leading up to their matchup with Liquid are represented in Table 3.1

From these statistics it is clear why Liquid was favored going into their quarterfinal matchup. Similar to Liquid’s permaban of Train, Astralis had a permaban of Mirage which they would ban in their first decision during the veto process. Astralis’s few strong maps were also strong maps for Liquid. This meant that even if they got a map veto to end up in their 3 highest win percentage maps being played, Liquid would still be favored to win. However, these statistics are against any team, to get a much better picture it’s important to look at the last time Astralis and Liquid played each other and how the map selection process unfolded.

On June 19th 2019, in the quarterfinals of the ESL Pro League Season 9 Finals, Astralis and Liquid faced off in a Best-of-3 series where Liquid won in a closely contested 2–1 victory. The map veto for this series went as follows:

1. Liquid ban Train
2. Astralis ban Mirage
3. Liquid pick Overpass
4. Astralis pick Inferno
5. Astralis ban Dust2
6. Liquid ban Nuke, Vertigo is the decider

The results for this series were an Astralis victory on Overpass (16–12), a Liquid victory on Inferno (16–6) and a Liquid victory on Vertigo (16–13) to take the series. Based on this previous map veto the idea of dominated strategies starts to take shape for both teams. Astralis and Liquid each have a map that they believe they are so weak on that if they play it, they will lose. As such these maps represent a teams dominated strategy and we can therefore simplify the game tree by using the Iterated Elimination of Strictly Dominated Strategies. A strategy is considered strictly dominated if there exists another strategy whereby the payoffs are strictly better than the previous strategy for all outcomes (Dutta 1999). For these teams the dominated strategy would be picking their permanently banned maps, or leaving them for other teams to pick against them. Liquid’s dominated strategy is not banning Train and it is dominated by removing it during their first ban phase. Similarly, Astralis’s dominated strategy is not banning Mirage and it is dominated by removing it during their first ban phase. Now the game is simplified by removing the first 2 bans as they both are the removal of the dominated maps for each team. The new map veto process after elimination consists of 5 maps (Dust2, Inferno, Nuke, Overpass, Vertigo) and 4 steps:

1. Team A picks a map
2. Team B picks a map
3. Team B bans a map
4. Team A bans a map and the last map is made the decider

On September 6th, 2019, when Astralis and Liquid faced off in Berlin in front of a sold out Mercedes-Benz Arena, the greatest shock came during the map selection process. Astralis was the higher seed so they were Team A while Liquid became Team B. The full veto process went as follows:

1. Astralis ban Mirage
2. Liquid ban Train
3. Astralis pick Vertigo
4. Liquid pick Overpass
5. Liquid ban Nuke
6. Astralis ban Dust2, Inferno is the decider map

Looking at this game in extensive-form after removing the dominated strategies is visualized in Graph 3.2:

The green lines represent a team’s pick options while the red lines represent their ban options

This veto results in the same 3 maps being played as the last matchup between the 2 teams with Overpass, Inferno, and Vertigo, however in this game they appear in a different order and this drastically impacted the outcome of the game. The most interesting decision made in this selection process was the surprise pick of Vertigo by Astralis. A dissection of this choice is illustrated in Graph 3.3

Based on these win percentages conventional knowledge and experts predictions believed that Astralis would pick into Inferno as a safety pick as it was currently their most successful map even though Liquid beat them on it the last time they played. Or they would pick Nuke as a show of confidence, a map where they had previously won 31 straight and for a long time was considered the “home map” of Astralis. However, Astralis did neither and instead picked Vertigo where they had won 2 out of 3 with the sole loss being a 16–13 defeat to Liquid, who were an undefeated 5–0 on the map since its release. The live reaction to this pick was deafening, the crowd roared, the analysts claimed that this picked showed that Astralis thought they were an underdog and needed a curveball in order to have a chance to win, and Liquid themselves had a look of bewilderment at the confusing pick in front of them.

Looking further into the potential payouts for each of the 5 options Astralis faced it becomes clear why Vertigo was the optimal choice for them. The best choice for Astralis is the map where they have the best chance to win vs Liquid. On Dust2, Inferno, and Nuke, their current form was poor or worse than that of Liquid. Also, Astralis was expected to pick one of either Inferno or Nuke, meaning that Liquid would be able to prepare against this pick in preparation, which would lower Astralis’s chance to win even further. Down to Vertigo and Overpass as the remaining 2 options it would seem that Astralis would want to play Overpass as it is the only map where their past win percentage is higher than that of Liquid. However, the format of this game is considered a sequential move game with imperfect information. In this disparity of information lies the deciding factor that led Astralis to pick Vertigo.

The true reason behind the confusion was the information difference between the 2 teams. Liquid believed they knew Astralis’s strengths, weaknesses, and current form but it had been over 3 months since the 2 sides last faced off. Astralis had taken this time to secretly prepare a completely new game plan on Vertigo in anticipation of this very moment in order to pull out the victory. Part of this secret strategy was maintaining it’s secrecy and purposely avoiding playing Vertigo much in the months leading up to the Major and even when playing it, never picking it but instead having it be the decider. Astralis’ coach, zonic, had this to say when asked about the surprise Vertigo pick after the match, “When I was looking at Liquid’s win streak on different maps, it was kind of like ‘what to pick?…We kind of wanted to catch them off-guard. We’ve been really good in practice on Vertigo, extremely good playing against all different teams, so it was also a case of not falling into old habits…It was about going a different route this time and then winning the mind game” (HLTV.org, 2019).

Liquid on the other hand, were completely caught off guard and had not prepared to play Vertigo against Astralis in this match. Had Liquid known Astralis’s prowess on Vertigo they would have instead chosen to ban it instead of Train. However, since teams are naturally risk averse, the risk of potentially playing their worst map in Train outweighed any other possibility Liquid thought they could face. They played it safe and Astralis took advantage of this unequal information between the teams. Due to the large disparity in preparation spent for Vertigo, the chance of Astralis winning this match heavily swung towards the Danish side. And since Astralis had banked on this pick being a surprise, they knew they were going to play a wild card and hoped to take the momentum from this pick and steamroll into a full match victory on the second map of Overpass. Following along the ideas mentioned in Malueg & Yates this is exactly what unfolded. After shocking the world with their Vertigo pick, Astralis was able to win with a convincing scoreline of 16–8. Moving towards Overpass, Astralis had all the momentum having just broken Liquid’s undefeated streak on Vertigo. With this added momentum their effort exerted was most likely higher than that of the now slightly demoralized Liquid and this led Astralis to taking the map victory, 16–13. With this Astralis won the quarterfinal 2–0, with the entire victory stemming from their decision to pick Vertigo during the map veto.

Conclusion

Counter-Strike: Global Offensive’s Starladder Major Berlin 2019 quarterfinals between Astralis and Liquid represents the essence of competitive Counter-Strike. Astralis’s surprise victory propelled them to their 3rd straight Major victory and their 4th overall while Liquid’s failure to perform on the world’s biggest stage remains a talking point with this match as the prime example.

This match elicited large interest from fans across the world, both in person at the sold out 17,000 person arena in Berlin and the hundreds of thousands watching live on the internet. CSGO still has a long way to go before it can be considered in the same sentence as traditional sports leagues such as the NFL, MLB, UEFA, and others, but all the pieces necessary to build a league that scales over time are there.

Valve Software has created a tournament structure where effort is correlated to success and more importantly, effort is rewarded with increased payoffs for winning. Although no large-scale competitive balancing has taken place, the Louvre Agreement signed by top teams and MTG signifies the start of stricter league governance and more balancing to create a more competitive scene that is not dominated by organizations with the most money to spend.

Finally, it is clear that decisions matter in Counter-Strike. In this quarterfinal the most crucial choice was made before the 10 players even entered the server. Astralis’s ability to understand their opponent and base their decisions on this information is what led to their eventual victory and Major championship run.

References

Dutta, P. K. (1999). Strategies and Games: Theory and Practice. Retrieved November, 2020, from https://drive.google.com/file/d/0BxeSXHD37Bt5TlI5VHpyNEFPa28/view

Ehrenberg, R. G. & Bognanno, M. L. (1990). Do tournaments have incentive effects? Journal of Political Economy 98(6), 1307–1324.

HLTV.org. (2020). Retrieved November, 2020, from https://www.hltv.org/

Hua, C. (2017). The Swiss Tournament Model. University of Pennsylvania Scholars Commons. Retrieved November, 2020, from https://repository.upenn.edu/wharton_research_scholars/149/.

Lazear, E. P., & Rosen, S. (1981). Rank-Order Tournaments as Optimum Labor Contracts. Journal of Political Economy, 89(5), 841–864. doi:10.1086/261010

Louvre Agreement. (2020, February 18). Retrieved November, 2020, from https://pro.eslgaming.com/csgo/proleague/wp-content/uploads/2020/02/Louvre-Agreement.pdf

Malueg, D., & Yates, A. (2010). TESTING CONTEST THEORY: EVIDENCE FROM BEST-OF-THREE TENNIS MATCHES. The Review of Economics and Statistics, 92(3), 689–692.

Melrose, A. (2018, April 20). Legal Analysis of the Overwatch League Structure and the Code of Condu. Retrieved November, 2020, from https://www.lawsofesports.com/single-post/2018/04/07/Legal-Analysis-of-the-Overwatch-League-Structure-and-the-Code-of-Conduct-A-Comparison-Study

Neale, W. C. (1964). The Peculiar Economics of Professional Sports: A Contribution to the Theory of the Firm in Sporting Competition and in Market Competition. The Quarterly

Journal of Economics, 78(1), 1. doi:10.2307/1880543

Olafsson, S. (1990). Weighted Matching in Chess Tournaments. Journal of the Operational Research Society, 41(1), 17–24. doi:10.1057/jors.1990.3

Ridenhour, K. F. (2020). Traditional Sports and Esports: The Path to Collective Bargaining. Iowa Law Review, 105(4), 1857–1897.

Rottenberg, S. (1956). The Baseball Player’s; Labor Market. Journal of Political Economy, 64(3), 242–258. doi:10.1086/257790

Valve. (2019, September 18). Keeping Things Competitive. Retrieved November, 2020, from https://blog.counter-strike.net/index.php/2019/09/25557/