The Failure of Artifact

Intro

Valve Corporation (referred to as “Valve”), a multibillion dollar gaming and publishing company (Chiang, 2011), released a digital collectible card game (CCG) called Artifact in November 2018. Over 60,000 players started playing the game on the first day. By March 2019 — only 4 months after launch — less than a thousand players remained, and Valve released a statement that they would cease updating the game to focus on re-examining Artifact’s core issues (Valve Corporation, 2019).

I am fascinated by Artifact’s failure for many reasons. First, Valve has, in the past, released million dollar game franchises like Half-Life and Dota 2, the latter being the intellectual property (IP) that Artifact is based on (Faylor, 2008; Gough, 2019). Second, Valve hired Richard Garfield, the creator of the iconic trading card game Magic: The Gathering, as lead game designer. And finally, Valve announced a $1 million USD tournament scheduled for the first quarter of 2019. Neither money, talent, nor the promise of a rewarding esports scene helped Valve capture market share in the CCG industry. Many blamed the game economy (Fogel, 2018); you had to pay for the game, pay for the cards, and pay for tickets to play some of the game modes. However, these facts were known before the release of the game. From the time of its release, Artifact was not appealing to casual players: people who had bought the game, but had no intention of competing or creating content for the game. Casual gamers are important for games to have because they are generally the largest portion of the player base, and the most profitable. For example, the largest digital CCG, Hearthstone, announced they had over 100 million players last year (Fogel, 2018). Of those players, only 48 players could compete globally in Hearthstone’s $500,000 USD Grandmaster tournaments (Blizzard, 2019). In terms of profitability, Hearthstone’s revenues for 2017 from in-game purchases through the PC platform were $217 million USD despite being free-to-play on all platforms (Goldman Sachs, 2018). Valve likely anticipated a large player base in order to pledge $1 million to their first tournament, yet did not test the game with potential casual players. Looking at the plunge in player count over four months, casual players were not intrinsically motivated to play Artifact over a sustained period of time; it wasn’t “fun” (reynad, 2018).

I will examine how that lack of motivation contributed to Artifact’s failure, and mention how hubris caused Valve to inaccurately predict the needs of their casual player base. Although the future of Artifact is uncertain, if Valve is able to attract potential players and sustain their enjoyment, it could have a chance at a second life.

Figure 1. A website that exists solely to track current players of Artifact. Adapted from Is Artifact a dead game? Retrieved November 13, 2019, from https://isartifactdead.com/

Context

Artifact was first announced in August 2017 at Dota 2’s International tournament. The CEO of Valve, Gabe Newell, described how his team wanted to “build the best card game” (PCGamesN, 2018). It had an alpha phase with high-profile players from other CCGs like Hearthstone and Magic: The Gathering Arena, which helped generate excitement for the game early on with their fans and their respective game communities (Akshon Esports, 2019). Nine days before launch, a closed beta version of the game was released to beta key holders — attendees of Dota 2’s The International 2018 and PAX West 2018 (Horti, 2018). Artifact was launched on November 28, 2018. It cost $20 USD to play the game, which came with 10 card packs. Players could not collect more cards unless they bought them from other players on Steam Marketplace (the full collection cost approximately $330 USD at the time of launch (Madigan, 2019)), or bought tickets to play the competitive draft mode (a mode that allows you to create a deck based on a random pool of cards) that would reward you with card packs if you won at least 3 out of 5 games.

In general concepts, a match in Artifact is played across 3 boards with a tower on each board, with players trying to defend their towers from being destroyed. The player wins the match when they destroy 2 of their opponents’ towers from different boards, or 1 tower on the same board twice. Players create a deck from cards they own or have drafted, which consists of 3 “hero” cards and at least 37 supporting cards. Players use their cards on any of the 3 boards to attack or defend against the opponent’s cards, or they can use special cards that change the cards already on the board. There is also a mini-shop during the match where items can be purchased and put into play (Bedford, 2018).

Figure 2. A view of one of three boards in a match of Artifact. Adapted from Imgur. Retrieved November 13, 2019 from https://i.imgur.com/f54F4Mm.png

In addition to the monetary aspect of the game, there were many elements of the game that were criticized by the player community (reynad, 2018; Margaris, 2019):

1. High variability (in gaming jargon, this is referred to as “RNG” for random number generator like dice to add unique permutations so that matches can play out differently) both in the game mechanics and certain card effects. For example, the card Cheating Death initially read: “If there is an allied green hero in this lane, allies have a 50% chance of surviving with 1 Health when they would die.” The end result was that half the time your allies would all survive for another turn, and the other half the time no effect would be applied.
2. Lack of clarity in the decision-making aspect of the game. Imagine an archer trying to hit a target. They need to decide where to aim the arrow and how much distance the arrow needs to fly. If the arrow falls short of the target, the archer understands intuitively that they need to draw their bow back more to make the arrow go farther to hit it. In Artifact, that kind of tacit knowledge of understanding which decisions would affect the outcome of the match was significantly harder to grasp, even among veteran CCG players.
3. No skill-based matchmaking in draft mode. In typical competitive games, each player is assigned a matchmaking rating (MMR) by the game, a number that correlates to their skill level. Every win or loss affects a player’s MMR, with good players having high MMRs and bad or new players having low MMRs. With no purposeful matchmaking using MMRs, it was possible for professional players to be matched against new players.
4. No global leaderboard. Implementing a MMR system also allows for an additional feature — the ability to show a global leaderboard of every player from best to worst. Leaderboards are particularly important to the players competing for the top spots, but even casual players use them for setting goals. Artifact was later updated to include a global leaderboard, but one that would reset every 2 weeks.
5. Difficult to understand as a viewer. Although the game consisted of 3 boards, usually only 1 board was shown on-screen. Viewers of the game were forced into memorizing 2 other boards for the majority of the match as they were beholden to the players navigating the boards at their own pace. This is akin to having viewers memorize 2 chess boards while observing a third in progress.

Figure 3. Peak player count for Artifact in its opening months. Retrieved from Steam Charts, 2019, https://steamcharts.com/app/583950#1y

Valve released two patches (a set of changes to the game) within a month after the initial release that implemented a leaderboard and changed some of the problematic cards (Valve Corporation, 2018). In January, despite Valve’s promise of a tournament, there were no announcements for qualifiers or tournament dates. Then in early March, it was confirmed that Valve had terminated Richard Garfield’s contract (rokman, 2019). The player count continued to drop significantly, resulting in a 99% loss of player count in 4 months (Steam Charts, 2019). Valve’s final communication to the community regarding Artifact was on March 28, 2019, and there have been no updates since then about Artifact’s revival.

Human Factors

There are two psychological human factors that contributed to Artifact’s failure. One examines the player experience as the average player was not intrinsically motivated to play the game despite investing money into it, and the second, hubris, examines Valve’s attitude during the game’s development process. They only tested their game with professional CCG players and a subset of Valve and Dota2 fans. Valve believed that the ideal CCG their in-house team would want to play would also be successful in the marketplace (Eurogamer, 2018).

The first human factor is the players’ intrinsic motivation to play the game. Under Self-Determination Theory, intrinsic motivation is described as people’s “inherent growth tendencies and innate psychological needs” (Ryan & Deci, 2000, p 227) being met. These needs are autonomy, competence, and relatedness. Generally the needs for autonomy and competence must be met for people to find intrinsic motivation to accomplish a goal, while the need for relatedness can be applied to tasks that contain a social aspect (eg: hiking versus soccer).

The need for autonomy is met when people feel they have a sense of volition and choice over their actions. While choosing to play a game is an autonomous action, within a game itself, people must feel like they have freedom to affect the game’s outcome.

The need for competence is met when people either see opportunities for improvement, feel their own skill level correlates to their success, or receive positive feedback for their perceived skill (Deci & Ryan, 2000). Another important aspect to the need for competence is having optimally challenging tasks. “Too much challenge relative to a person’s skills leads to anxiety and disengagement, whereas too little leads to boredom and alienation” (Deci & Ryan, 2000, p260).

The need for relatedness is met when people feel a sense of connectedness to others through the activity. In this particular instance of gaming, relatedness can be interpreted as players understanding how they perform compared to others and the formation of communities around the game (Ryan, Rigby, Przybylski, 2006).

The second human factor is Valve’s hubris that led to their misunderstanding of the needs of their primary player base, namely, casual CCG players. Hubris can be described as “arrogance and overconfidence that prevent someone from understanding or accepting human limits” (Picone, Dagnino, Mina, 2014, p 449). Specifically, hubris can manifest in three ways (Moor & Healy, 2008). Firstly, the person overestimates their own capabilities, outcomes, and likelihood of success. Secondly, the person perceives that they are better than others in comparable tasks and activities. Lastly, the person is overconfident in the accuracy of their own beliefs.

Between these two factors, Artifact had serious flaws and misconceptions that Valve failed to address. While the hubris aspect is important, in this paper I will delve deeper into the first factor as I believe if Valve finds a way to address it in the future, their game could return with a better chance of success.

Analysis

The game elements previously described influenced player enjoyment of Artifact, which in turn affected their motivation to play (reynad, 2018). All five elements eroded the players’ inherent psychological needs for autonomy, competence, and relatedness.

Variability in many games create interesting and unique match permutations that players then have to “solve” in order to win; most standard card games, for example, create variability by shuffling the deck before dealing cards. However, the amount and type of variability affect how players feel because every instance of variability has the potential to diminish the player’s ability to make choices and decisions that matter. In Artifact, there were so many elements of variability that the player’s autonomy was diminished. In the card Cheating Death, players felt frustrated when they played the card and nothing happened — yet players also felt frustrated when their opponent played the card and the effect took place, simply because the results were out of their control. Having high variability that could potentially decide wins and losses meant that players had to look elsewhere for key decisions that would increase their chances of winning.

This leads into the second problem of the game: lack of clarity in the decision-making aspect. During a match, a player would be able to choose which hero they wanted to play on which board, and which items from the mini-shop the hero could consume, but the game randomly decided where the cards would be situated on the board and in which direction they attacked (left, right, or centre). If we compared Artifact to war waged on a battlefield, it would be similar to the player, as commander of an army, choosing which units to participate in the battle (a high-level, strategic decision), and what kind of armour they would wear (a low-level decision), but not their positions on the battlefield nor which opposing units to attack (mid-level decisions). With arbitrary parts of the game decided upon randomly, the players had a difficult time figuring out which of their decisions would have changed the outcome of the match. This difficulty in analysis affected their ability to improve their competency at the game.

By having random matchmaking in the game, players could be matched with anyone. Relating to players’ need for competence, this system of matchmaking led to unoptimized challenges. Veteran or good players would have found new or bad opponents too easy to beat, leading to boredom. Conversely, new or bad players would have found veteran or good players too difficult to beat, leading to disengagement. In either scenario, players were discouraged psychologically from playing.

Without a global leaderboard, players had a difficult time comparing themselves to others or how well they did in the past. The game initially offered no statistics to show players their successes or their progression in any metrics (common metrics in other games are match history, deck win rate, competitive rank, etc.). Even with a global leaderboard that reset every two weeks, players still found it difficult to relate to other players simply because every two weeks, all of their progress would be erased and the climb would begin anew. There was no way for players to see their improvement over a long period of time. This impacted their need for competence, and their need for relatedness to other players.

Finally, Artifact was difficult to understand as a viewer. With 3 different boards to follow, and match length anywhere between 15 to 30 minutes (TomRMarks, 2018), casual viewers needed to pay close attention to the match being played over a sustained period of time, which proved too burdensome for most viewers (average viewership was just over 24 000 viewers on Twitch within a day of release, with average viewership dropping below 10 000 after two weeks of release (Twitch Tracker, 2019)). Watching games online allows for a fairly high level of interactivity on a platform like Twitch where viewers can chat directly with streamers and other viewers(Takahashi, 2019). This interactivity provides viewers a measure of relatedness to other members of the community, a need that Artifact did not satisfy because matches were too difficult to watch.

With all of these issues outlined, Artifact failed to meet the intrinsic needs of their casual player base and did not foresee it due to their hubris. Fulfilling intrinsic motivations of people has been not just a challenge for Valve, but also for schools and work organizations (Deci & Ryan, 2000). To address these challenges, more and more industries are looking at gamification as a solution. Gamification as an idea is “to use elements of game design in non-game contexts, products, and services to motivate desired behaviors”(Deterding, 2012, p.14). I think it is pertinent to look closely at a game that failed despite all indications for success in order to understand which game aspects discouraged players from engaging with it. By looking at what failed, we can create a better understanding of what might succeed in future games and in non-traditional applications.

Conclusion

From the initial announcement right up until release, Artifact seemed to have so much potential: it was visually appealing, the rules were simple but strategies to win were complex, it attracted a large pool of competitive players from other CCGs, and it was built on an incredibly successful IP. However, it was missing the “fun” factor (reynad, 2018). Valve failed to meet the intrinsic needs of their casual players. By only allowing competitive gamers to provide feedback to the game, believing they were the only important audience to capture, Valve could not predict the game’s failure because it had not tested the game with the people who would make the game profitable. In the future, if Valve can fix the core issues of the game — high variability, unclear decision-making, no MMR-based matchmaking, ineffective leaderboard, and unappealing viewing experience — the game might increase its chance of success. Otherwise, Artifact might just end up being an artifact of the collectible card game industry.

Bibliography

Admiral Garp’s Grandson (2018, November 30). The most powerful card is credit card. [Steam review]. Retrieved from https://store.steampowered.com/app/583950/Artifact/#app_reviews_hash

Akshon Esports. (2019, February 26). Artifact: A Launch — The Story Of Valve’s Digital Card Game Disaster [Video file]. Retrieved from https://www.youtube.com/watch?v=rjjudgRAlC0

Bedford, J. (2018, November 20). Artifact: Guide. Metabomb. Retrieved from https://www.metabomb.net/artifact/gameplay-guides/artifact-the-dota-card-game-guide-8

Blizzard. (2019). Hearthstone Grandmasters. Retrieved November 13, 2019 from https://playhearthstone.com/en-us/esports/programs/grandmasters

Chiang, O. (2011, February 15). Valve And Steam Worth Billions. Forbes. Retrieved from https://www.forbes.com/sites/oliverchiang/2011/02/15/valve-and-steam-worth-billions/#454e5e6633f4

Deci, E. L., & Ryan, R. M. (2000). The” what” and” why” of goal pursuits: Human needs and the self-determination of behavior. Psychological inquiry, 11(4), 227–268.

Deterding, S. (2012). Gamification: designing for motivation. interactions, 19(4), 14–17.

Eurogamer. (2018, May 2). We interview Artifact’s lead designer, Richard Garfield (AKA the creator of Magic: The Gathering) [Video file]. Retrieved from https://www.youtube.com/watch?v=_If41SYSg3c

Faylor, C. (2008, December 3). Valve Reveals Lifetime Retail Sales of Half-Life, Counter-Strike Series. Shack News. Retrieved from https://www.shacknews.com/article/56193/valve-reveals-lifetime-retail-sales

Fogel, S. (2018, November 5). ‘Hearthstone’ Now Has Over 100 Million Players. Variety. Retrieved from https://variety.com/2018/gaming/news/hearthstone-has-over-100-million-players-1203019919/

Fogel, S. (2018, November 19). Valve Fans Not Happy About ‘Artifact’ Monetization System. Variety. Retrieved from https://variety.com/2018/gaming/news/valve-artifact-monetization-backlash-1203032287/

Goldman Sachs. (2018, June 26). eSports: From Wild West to Mainstream. Retrieved from Strive Sponsorship website: http://strivesponsorship.com/wp-content/uploads/2018/06/Goldman-Sachs-Esports-Report.pdf

Gough, C. (2019). Revenue of DOTA 2 worldwide 2015–2017. Retrieved from Statista: https://www.statista.com/statistics/807617/dota-2-revenue/

Horti, S. (2018, October 20). Artifact’s beta appears to have been delayed until November 19 [Web log post]. Retrieved from https://www.pcgamer.com/artifacts-open-beta-appears-to-have-been-delayed-until-november-19-just-nine-days-before-full-release/

Madigan, S. (2019). How much does Artifact cost? Retrieved from https://www.howmuchdoesartifactcost.com/

Margaris, J. (2019, August 12). Why Artifact Failed [Web log post]. Retrieved from https://www.gamasutra.com/blogs/JamesMargaris/20190812/343376/Why_Artifact_Failed.php

Moore, D. A., & Healy, P. J. (2008). The trouble with overconfidence. Psychological review, 115(2), 502.

PCGamesN. (2018, March 9). Gabe Newell discusses Artifact at Valve HQ [Video file]. Retrieved from https://www.youtube.com/watch?v=mERhtoD21rU

Picone, P. M., Dagnino, G. B., & Minà, A. (2014). The origin of failure: A multidisciplinary appraisal of the hubris hypothesis and proposed research agenda. Academy of Management Perspectives, 28(4), 447–468.

reynad. (2018, November 28). Reynad’s Thoughts On Artifact | Game Review [Video file]. Retrieved from https://www.youtube.com/watch?v=DV-YlwC0sPw

rokman. (2019, March 8). Garfield is no longer at Valve [Web log post]. Retrieved from https://www.artibuff.com/blog/2019-03-08-garfield-is-no-longer-at-valve

Ryan, R. M., & Deci, E.L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American psychologist, 55(1), 68.

Ryan, R. M., Rigby, C. S., & Przybylski, A. (2006). The motivational pull of video games: A self-determination theory approach. Motivation and emotion, 30(4), 344–360.

Software Testing Help. (2019, November 10). Alpha Testing And Beta Testing (A Complete Guide). Retrieved November 13, 2019 from https://www.softwaretestinghelp.com/what-is-alpha-testing-beta-testing/

Steam Charts. (2019). Artifact. Retrieved from https://steamcharts.com/app/583950#1y

Takahashi, D. (2019, April 13). Our deep dive into how esports broadcasting differs from traditional sports. Venture Beat. Retrieved from https://venturebeat.com/2019/04/13/our-deep-dive-into-how-esports-broadcasting-differs-from-traditional-sports/