Enter the title of any mainstream video game followed by the word “gameplay” into a YouTube search, and you will soon be presented with a wealth of user-captured evidence of what video games look like. Some of this comes from directly captured audio-visual signals and appears essentially seamless, while some is more roughly captured from television screens with handheld digital cameras. Some includes audio commentary and shows, blow by blow, how to complete a given game, some displays users’ impressive achievements in gameplay, and some provides a record of games which may no longer be easily accessible to users. This, of course, is only an incomplete picture of the wide variety of game-related content that can be found on YouTube.
To begin this essay I would like to point to a particular thread within all of this video content: gameplay samples uploaded by users complaining about the most recent iteration in the Madden NFL series of football games, Madden NFL 12. While this might seem like a minor topic, the number of examples to be found is quite staggering. Grumbling through their frustration (and often their impressive profanity) these users describe artificially prescient computer-controlled opponents, poorly animated movement, limbs and torsos passing through each other, flying footballs inexplicably changing course in mid air, superhuman movement, and scores of other problems.
Though Madden is one of the most successful video game franchises of all time and Madden NFL 12 sold over a million copies in the week after it was released, these users have no trouble finding flaws in the game, or supporting their claims with video evidence. Furthermore, Madden NFL 12 is not a particularly poor representation of the series; these sorts of complaints are made on a yearly basis.
While perusing these grievances, I came upon a video titled “Madden 12’s magical 1 handed interception” that led me to consider some unique issues in users’ dissatisfaction. The video’s creator, identified as “bballashotkalla,” protests that “these players can jump up twenty feet in the fucking air, these linebackers, and intercept these passes like they’re fucking Jesus Christ or something.” The video indeed shows Washington Redskins linebacker London Fletcher leaping to make a one-handed interception against the Kansas City Chiefs. The play seems improbable, either in “real” life or in a Madden game, especially because Fletcher is a hard-tackling linebacker and is not expected to be so dexterous. Logos on the virtual field reveal that the recorded game was a virtual Super Bowl, meaning that it is likely to have had particular significance in the National Football League season narrative supported by Madden NFL 12. This helps explain bballashotkalla’s extreme frustration with the game: he feels that he has been cheated at an important moment.
What makes this particular video especially telling, however, is its resemblance to an athletic act which took place in a real-world football game. In January of this year, J.J. Watt of the Houston Texans leaped to make a one-handed interception against the Cincinnati Bengals, and returned the ball for a touchdown. In a critical single-elimination playoff game, this gave the Texans a lead that they would not relinquish. The real play by Watt is visually similar to and as impressive as the play by the virtual Fletcher, and the two plays came in games of very similar import. Watt’s play may actually be more surprising for, as a defensive end, he is expected to be even less apt to catch a football than the linebacker Fletcher.
An important difference between the two situations is that, while the virtual Fletcher is seen as having done something impossible or even wrong, Watt’s play is seen as impressive. NBC commentator Mike Mayock indeed suggested “I think [Bengals quarterback Andy] Dalton knows that there’s not much you can do about that. That’s just a great individual effort by an opposing player.” It seems that in both cases the offensive player (Dalton or bballashotkalla) felt helpless in the face of what their opposition (Watt or the virtual Fletcher) did, but the real play can be accepted while the virtual play appears unacceptable and unfair. This is in spite of the fact that the two plays are about as near to identical as two football plays can be. Why is it that one event is merely unusual, while the other is deplorable? The root cause seems to be something that generally goes unmentioned in discussions surrounding sports games, perhaps because it seems too obvious to mention out loud: what happens in sports games and what happens in real sports are not the same.
Difficulties of Visual Resemblance and The Uncanny Valley
The statement that sports games and real sports aren’t the same is quite broad and encompasses many differences between the two. These differences seem to be problematic because sports games aspire to be perfect simulations of real sports. In other words, they aim to be entirely the same as real sports, or at least as close to that as possible. Of course, as so many YouTube users have pointed out, a game like Madden NFL 12 hardly accomplishes this goal.
However, even these dissatisfied users would have to admit that the game does at least make impressive strides towards visually resembling the real National Football League, down to the uniforms, tattoos, hairstyles, jumping, catching, and tackling. Nevertheless, perhaps the most easily apparent difference between sports games and real sports is the fact that, at least with current programming, graphics, and computing technology, the visual resemblance between sports games and real sports is incomplete. Sports games certainly don’t look exactly like real sports seen in person, nor do they look exactly like televised real sports, despite the fact that televised real sports, like video games, are viewed on television screens. Technical limitations cause Madden to struggle in rendering everything from virtual athletes’ facial expressions and movement, to more general things like lighting and the textures of different materials. While a televised football game is a very believable simulacrum of the real game it depicts, these flaws prevent Madden from being a perfect simulation.
The concept most frequently deployed in examining these visual imperfections is the idea of the “uncanny valley,” first described by Japanese robotics engineer Masahiro Mori in 1970. In considering a graph of our comfort with robots versus their visual likeness to humans, Mori said he “noticed that, as robots appear more humanlike, our sense of their familiarity increases until we come to a valley. I call this relation the ‘uncanny valley,’” referring to the dip seen in the graph.
Research into the causes of the uncanny valley remains ongoing, but it does seem that there is an unpleasantness about something which appears very human, but not quite human. This is why Mori ultimately recommends that engineers building robots should take advantage of the first peak in the graph, rather than strive for the second, for “although the second peak is higher, there is a far greater risk of falling into the uncanny valley.” Applying these insights on robots to virtual athletes in sports games, it seems that this may also be why the visual imperfections in Madden NFL 12 are more forcefully apparent than those in, say, an older Madden game like Madden NFL 95. The later game looks a great deal more like real-world football, but as the Madden series’ developers have produced more lifelike graphics, they have actually exposed their games’ visual appearance to harsher scrutiny.
Another study in robotics by Jennifer Goetz, Sara Kiesler, and Aaron Powers adds an additional element to the uphill battle for visual verisimilitude. These authors note that “humanoid robots convey animistic and anthropomorphic cues that evoke automatic perceptions of lifelikeness in the robot. These perceptions will lead to people making attributions of ability and personality to the robot. In turn, their social responses and expectations will be shaped by these initial attributions.” In other words, robots’ appearances are pleasing when they match their function and abilities. A robot should look very humanlike only if it can behave in a humanlike manner, while a microwave is better off looking like a box. Applying these insights to sports games suggests that the more lifelike sports games appear, the more will be expected of their ability to simulate real-world situations. This seems like a fairly common-sense point about the Madden series: the more lifelike the games look, the more unacceptable it will be when one virtual player’s hand passes through another’s arm, or a football changes course in mid air.
The aims of the Madden series and other mainstream sports games seem to fly in the face of these observations. Attempting to make a game with a lifelike appearance is very demanding, and it seems that it can only be truly a successful endeavor if it is absolutely successful. Nevertheless, game developers have continued to pursue visual verisimilitude, and perhaps they will succeed with enough technological advancement, in both the techniques for making games and in the computing power to run them. Though this doesn’t seem achievable in the near future, it’s a fine hypothetical solution for any complaints about momentary amputation.
But, looking back to the London Fletcher example, do the visual problems of the uncanny valley and the need to match humanlike appearance with humanlike behavior entirely explain the frustration displayed in bballashotkalla’s video? Would making the virtual Fletcher completely humanlike in appearance and behavior, i.e. making him look like J.J. Watt, completely resolve this problem? My suspicion is that it would not; after all, the two already look quite similar, that was the reason to compare them in the first place. Greater visual verisimilitude would address some issues to be sure; giving Madden a more lifelike appearance would make it difficult for users to suggest that a linebacker jumped unreasonably high, or that he made an unreasonably dexterous catch. Nevertheless, I believe that a user like bballashotkalla will still feel somewhat cheated by an unexpected interception, even if that interception appeared entirely lifelike.
Difficulties of Credibility, Simulation, and Realism
To construct a more complete explanation for this sense of frustration, it is necessary to locate another way in which sports games and real sports are not the same; something beyond visual appearance. It seems helpful, then, to revisit how Watt’s interception was acceptable, as evidenced by Mike Mayock’s commentary, while Fletcher’s was not, as evidenced by bballashotkalla’s video. This was despite the fact that the two plays were, from a football standpoint, essentially the same. Since that is the case, the difference must stem from the context of each play, rather than any feature of the plays themselves. It could be said that both plays were “incredible,” but in different senses of the word. Watt’s catch was “incredible,” meaning “so extraordinary as to seem impossible,” while Fletcher’s was “incredible,” meaning “not credible; hard to believe.” The question, then, is what about the context of Fletcher’s interception sapped it of its credibility?
The answer can be found easily enough, for once again one only has to state the obvious: outcomes in real sports really happen, and so they are entirely credible, while outcomes in sports games are simulated, and so they are not entirely credible. If something incredible happens in real life, there is no choice but to accept it. No matter how strong an understanding of the world one may have held before a truly incredible event, that understanding must be reconsidered in the face of the inconceivable becoming conceivable. If you believed that a defensive end could not make a leaping, one-handed interception, J.J. Watt has proven that your belief was ill founded, and that it must be reconsidered. Alternatively, if something incredible happens in a video game, it’s very easy to question it. An incredible play by the virtual London Fletcher can hardly convince us that the real London Fletcher is capable of doing likewise. The only belief which must be reconsidered is the belief that the game in question is a well-designed simulation. This, in essence, is what all of the users who uploaded videos pointing out Madden NFL 12’s failings were doing: suggesting that the game is not a well-designed simulation.
In the case of Madden NFL 12, with so much visual evidence available on YouTube, this claim actually seems to be a very strong one; the game doesn’t seem to be well-designed. However, it seems that even a well-designed simulation would struggle with a lack of credibility in the face of incredible events. MLB 11 The Show might be considered a better-designed simulation, but when users’ perfectly located pitches are hit for a home runs, the game will likely struggle with its credibility in the eyes of its users, even though good pitches are frequently hit for home runs in real Major League Baseball games. The same goes for goalies making incredible saves against users’ well-placed shots in FIFA 12 or NHL 12, players missing wide-open in-rhythm jump shots in NBA 2K12, or a ball taking an unfortunate bounce in Tiger Woods PGA Tour 13. If these things aren’t enough to strain players’ belief in sports games, what happens when a goalie makes incredible saves throughout an entire game, or a basketball player misses ten consecutive jump shots? These things certainly do happen in real sports, but in sports games they will eventually be hard to believe. And, while users may react more strongly to their virtual misfortunes, one’s good fortune in games may also seem all too incredible when, say, a player makes ten consecutive jump shots or a golf ball always bounces the right way. Again, these things certainly do happen in real sports, but may be hard to take in sports games.
Though incredible outcomes may cause games to lose their credibility as simulations, sports games also cannot gain credibility by backing away from incredible outcomes. Real sports are expected to be unpredictable and dramatic, rather than entirely consistent. A sports game without incredible outcomes would thus not only be boring, it would also feel unnaturally homogenous. As a result, players need to go on sudden hot and cold streaks, balls need to take unpredictable bounces, and mediocre players need to make unexpected, incredible plays at least some of the time, but not too often. No player should always make a certain sort of catch; sometimes he should drop the ball. These aspects of sports games, then, must be determined by probabilistic abstractions: the player will make the catch 79 out of every 100 times. And, because sports are unpredictable, these probabilities cannot be absolute: the player will make the catch 79 out of 100 times, 87 out of 100 times. The developers who design sports games have to set up these probabilistic models and try to match them to real-world statistics.
For users, however, experiencing the outcomes produced by a very good probabilistic model is still different from experiencing the outcomes which occur in the real world; the outcomes of such a model are never entirely credible in the way that real events are entirely credible. As a result, games that seek to simulate sports can never achieve the “reality” of real sports. Instead, they can only achieve “realism,” an effect produced in the minds of users; the sense that sports games are very similar to reality. Realism, it seems, can never be absolute, as incredible and unexpected outcomes must always intervene, but preclude any sort of stability. A game that allows London Fletcher to make a one-handed catch will always lose credibility, but so will a game that never allows Fletcher to make a one-handed catch. A perfect balance between the two, a balance that would yield absolute realism, does not appear to be achievable.
Difficulties of Control and Difficulty
So far I have considered sports games as largely separate from the input provided by the users who interact with them, as I have focused on the frustration and disbelief involved in moments that are well beyond users’ control. Interactivity, however, introduces additional problems for sports games. The first set of these stems once again from a way in which what happens in sports games is not the same as what happens in real sports; namely, from the fact that real sports are played with one’s own body, while sports games are controlled through much more indirect mechanisms.
Consider attempting to block a shot in a real game of basketball. A player blocks a shot by using their body; they jump as high as their legs cause them to, their arms move as fast as they cause them to, and they react to the offensive player as quickly as possible. If they deflect the ball their body has done something successful, if the ball continues unimpeded their body has done something not entirely successful. The entire truth of the matter is readily apparent. Alternatively, consider blocking a shot in a basketball sports game. Since the advent of basketball video games (or at least since the mid ‘90s) success or failure in virtual shot blocking has usually depended on a user positioning the virtual player under their control with a directional pad or control stick, and then pressing a button assigned to blocking at the right moment. While this certainly requires quick reflexes, good timing, and other skills, attempting to block a shot in a real game of basketball is much more complicated than pressing a button. As a result, the game must invent the movement of the player under the user’s control. In more recent games, players programmed to be good shot blockers may move quickly and decisively, while players programmed to be less effective shot blockers may move more slowly, even though the user presses the same button to cause each to attempt to block a shot. Even if probability is not introduced into the situation and a user will always succeed in blocking a shot with the proper positioning and timing, it’s easy for a user to second-guess a sports game when it judges their timing to have been off the mark. This is because the truth of the situation is not readily apparent, as it is in real basketball. In real sports, outcomes are directly and physically determined, while in sports games outcomes are mediated by the user’s means of controlling aspects of the game and by the game’s programmed simulation of the real world.
The advent of so-called “motion controls” in the form of the Nintendo Wii, Kinect for Xbox 360, and PlayStation Move might appear to offer solutions to this problem through their ability to respond to users’ nuanced physical movement. However, these devices cannot truly combine video game space and real space. A Wii controller and a virtual baseball are not the same as a real bat and ball because they cannot touch each other; a real physical relationship is not present. Furthermore, each of these devices only works within a small area in front of users’ televisions, and there is a fundamental disconnect between simulating, say, an entire football field within the space of a living room. Additionally, users still only have images from their televisions to react to, rather than full three-dimensional, panoramic vision which a user can reorient by turning their head. As such, the information they can receive is limited, and so they cannot be expected to react as quickly as an athlete in a real sport. In short, motion controls cannot produce fully embodied experiences, and at each of the points of disconnect discussed here they must involve invention or abstraction beyond what users can input. As a result, games using motion controls can still be second-guessed in the same way that button-based games can be. Perhaps the sort of seamless virtual reality envisioned by science fiction like The Matrix will solve these problems, but until then our conception of sports games will retain these limitations.
Because there is such a distance between users’ input and the outcomes involved in sports games, the success or failure of users is arbitrary; it is simple to program a game that users will either always win or always lose. However, as has been discussed already, sports games strive to reproduce the appearances and outcomes of real sports. Thus, if professional sporting leagues generally produce close contests, so should sports games modeled on those leagues. If they fail to do this, sports games lose their credibility and their essential relationship with reality. Interactivity causes another problem here, however, because some users are more skilled at playing sports games than others. The New York Yankees should beat the Kansas City Royals maybe two out of three times if a sports game is to conform to reality, and they should have this success rate whether they are controlled by a relatively skilled user or by a relatively unskilled user, because the game seeks to be credible for all users.
Sports games achieve this by introducing variable levels of difficulty, ranging from “Easy” to “Hard.” As difficulty increases, users will face stiffer opposition and will find it more challenging to achieve positive outcomes. For instance, users playing a football game set with increasing levels of difficulty will find that computer-controlled defensive players react more quickly to their offensive plays. Along similar lines, blocking the same sort of shot in a basketball game set with increasing levels of difficulty will require increasingly well timed user input. The same goes for the timing involved in hitting pitches in baseball games; once users have mastered one level of difficulty and find themselves achieving positive outcomes with incredible consistency, they are encouraged to increase the game’s difficulty in order to restore the game’s balance, credibility, and relationship with reality.
This is an unpleasant situation, however, because it ultimately forces users to decide how often they would like to win in sports games. If a user is able to take a squad of unheralded players on an incredible journey to a championship, was it because they played well or because they didn’t have the game’s difficulty set on an appropriate level? Once this difficulty of difficulty, as it were, is apprehended, the arbitrariness of users’ successes and failures comes to the fore once again, and sports games lose their essential relationship with reality.
One possible response to this situation is to take the setting of difficulty levels out of the hands of users and instead program difficulty to be set dynamically in response to the outcomes users achieve. This is what Madden NFL 09 attempted with its “Madden IQ” difficulty setting, which would increase difficulty as users won games and decrease it as they lost. Hiding dynamically determined difficulty behind the scenes might make for an even stronger effect, concealing the arbitrariness of users’ successes and failures even further. However, dynamic difficulty necessarily enforces a “normal” level of outcomes for users, which could cause sports games to resists the incredible to an undue degree and become homogenous. It also seems that dynamic difficulty could make users’ efforts to achieve better outcomes in a sports game meaningless, as doing so would only cause the game to pose greater challenges. In the end, it seems that the arbitrariness of users’ successes and failures will come through at some point.
Games Creating their Own Reality
Given these difficulties with visual verisimilitude, credibility and realism, and interactivity and difficulty, it seems that sports games will continue to be deeply flawed for the foreseeable future. Perhaps the development of a form of seamless virtual reality or the introduction of other new technology could resolve some of these issues by allowing sports games to attain more of the reality or directness of real sports. Judgment on such developments will have to be withheld until the fundamental relationship between sports games, their users, and reality is indeed changed. Until then, however, the flaws of sports games can be used to shine light on the strengths of other varieties of video games.
Think of some of the most successful arcade games; Space Invaders and Defender can hardly be lauded as brilliant simulations of space combat, Joust won’t garner praise as a realistic representation of ostrich-mounted jousting, and PacMan can hardly be said to relate to the real world at all. The same can be said for recent games like Super Stardust HD or Shatter which owe a great deal to classic arcade concepts, or games like The Last Guy which only relate to the real world very abstractly. These games don’t tie themselves to the reality of the real world through an effect of realism; instead, they create their own reality.
Looking through the flaws found in sports games, it’s clear that they aren’t present in these more direct video games. Visual verisimilitude and realism are not at issue in a direct game like Tetris because Tetris makes no attempt to look like the real world. There are no problems with credibility in a direct game like Asteroids because Asteroids does not attempt to simulate anything; instead it presents only itself. If something incredible happens in a game of Asteroids it is not incredible because it seems unlike what is expected to happen in the real world, it is incredible because it is unlike what is expected to happen in Asteroids. As such, “incredible” shifts back to its kinder sense, and incredible outcomes in direct games can be believed without question. Interactivity is not a cause for concern in a direct game like Super Mario Bros. because the actions Mario is expected to perform match users’ ability to provide input. Mario is expected to run from side to side at two different speeds, duck, go down pipes, jump, and use a special abilities like throwing fireballs, and users are given a four-way directional pad and two buttons, which is enough to control these actions entirely. If Mario jumps too high, it is not because game invented the height of his jump or determined it probabilistically, it is because the user held the jump button for too long. Thus, the efficacy of Mario’s movement completely matches the efficacy of users’ input. Finally, difficulty is not difficult in a direct game like Centipede because the arbitrariness of the game’s difficulty is manifest. It is clear that direct games are entirely invented by their programmers, and so how difficult it is for users to achieve successful outcomes is not beholden to the real world. The difficulty of direct games is thus free to become a matter of users’ taste; some users will prefer to succeed with ease, while others will prefer a greater challenge. Centipede actually anticipates users’ different tastes, or users’ different levels of ability, by providing three difficulty settings. Difficulty settings in sports games are introduced to strengthen sports games’ connection to the real world by counteracting users’ different levels of ability to keep outcomes in line with real-world expectations. Thereby, however, sports games make users’ achievement of particular outcomes arbitrary. Alternatively, difficulty settings in direct games allow games like Centipede to simply offer users different challenges.
The categories of sports games and direct games are, of course, not absolute. For example, Madden NFL 95 was mentioned earlier for its more abstract graphics, and these indeed allow the game to move away from the real world and towards creating its own reality. This is the case to an even greater degree with simpler sports games like Baseball Simulator 1.000 and Tecmo Bowl, which make no attempts to simulate the full complexities of athletic acts or the full appearance of the real world, leaving less for users to question. Nevertheless, a fielding error or a poorly-thrown ball in these games can still be blamed on their use of probabilistic determinations, and thus can be second-guessed. Others sports games are able to match users’ ability to provide input with what would be necessary in the real world; racing games like Gran Turismo 5 or flying games like Pilot Wings 64 have had little trouble equating mechanical game controllers with the mechanical controls of cars and aircraft. Driving a car and playing Gran Turismo 5 will always be quite different, while playing an entirely direct game like Super Mario Bros. cannot be unlike itself, but it can also be said that if a user’s car goes off of the road in Gran Turismo, it’s almost entirely their fault. In these cases, then, directness in input also helps sports games’ credibility.
Many of the flaws that I have identified in sports games are also present in other games that relate to the real world. That includes everything from the Uncharted series back through Grand Theft Auto 3, Goldeneye 007, and beyond. All games of this sort will have at least some problems with lifelike appearance, credibility and realism, and interactivity and difficulty. Ultimately, it seems that these failings originate from the truism that what happens in video games and what happens in the real world are not the same. Along these lines, it could be said that what happens in these “realistic” games is not real at all, while what happens in direct games is real in and of itself.
This is not to say, however, that games striving for realism should be abandoned. First and foremost, games which resemble the real world allow us to simulate experiences, like playing in a top tier professional sports league or fighting in World War II, which most of us will not have a chance to undergo firsthand. Clearly enough, they also relate to the real world in ways that direct games do not, and thus can tell stories, change users’ understanding of the world, make political statements, and serve other purposes in ways that direct games cannot. Even if experiences with realistic games are indeed fraught, their defects are not damning. I, for one, keep coming back to and enjoying realistic games anyway, so they can’t be all bad. Still, I do think that, for realistic games to move forward, both developers and users need to be aware of these problems and to work towards new ways of addressing them.
 Here I will refer to people playing video games as “users,” rather than as “players,” to avoid them being confused with the virtual “players” involved in sports games. Please note, however, that “user” doesn’t really seem to be an adequate word, since someone “using” Microsoft Office and someone “using” Super Mario Bros. seem to be doing pretty different things.
 This story available at http://www.1up.com/news/madden-12-sells-14-million-first-week. It’s the same every year, really.
 I’m going to keep using the word “real” throughout this essay to refer to the physical or tangible world outside of video games. The caveat, of course, is that games are also “real” in a sense. They also happen, and indeed happen in the “real” world; that’s where one plays them or sees them. But the three-dimensional space of a game like Madden isn’t “real,” it’s illusionistic, representational, or “virtual,” while “real” football is played in “real” space. So that I can drop the scare quotes, consider this caveat about the word “real” to be in effect for the rest of this essay.
 That is, insofar as real and virtual sport can be compared; such a comparison has to be made with virtual sport’s goal of simulation in mind. Here, Fletcher’s play seems impressive because the terms of Madden, as a simulation, make it seem impressive. Under other terms, Fletcher actually jumping twenty feet into the air might not be surprising at all. Similar comparisons between real and virtual sport will be made here, so consider this point about simulation to be another proviso that will be in effect for the rest of this essay. All of these problems with language may suggest that video games require a new vocabulary, perhaps with some newly invented words.
 Masahiro Mori, “The Uncanny Valley” Energy, vol. 7 no. 4. (1970), 33-35. Translated by Karl F. MacDorman and Takashi Minato, 2005, available at: http://www.androidscience.com/theuncannyvalley/proceedings2005/uncannyvalley.html. The above version of Mori’s graph was also produced by MacDorman and Minato.
 Mori, “The Uncanny Valley.”
 Jennifer Goetz, Sara Kiesler, and Aaron Powers, “Matching Robot Appearance and Behavior to Tasks to Improve Human-Robot Cooperation,” Proceedings of the 2003 IEEE International Workshop on Robot and Human Communication (2003), 55-60.
 http://dictionary.reference.com/browse/incredible?s=t. Used in a sentence: Isn’t it incredible that I have no qualms about citing dictionary.com in my term paper? Thanks, internet!
 Or “Very Hard.” Or “Grueling.” Or “All-Madden.” The names, like the levels of difficulty, are arbitrary.
 I’ll go with the name “direct games” to refer to games that have this quality. “Arcade games” often refers to the same thing because most arcade games are indeed direct games, but “arcade games” points to a particular part of the history of video games. “Direct games,” alternatively, can refer to games created at any time and for any platform.
 Users may occasionally suspect that a “glitch” or shoddy programming has altered an outcome in a direct game. It seems, then, that an important concern for the directness of direct games is ensuring that outcomes are consistent. In Asteroids, for example, this means ensuring things like users’ projectiles always causing asteroids to explode when they are properly aimed.