In the very first issue of the online Game Studies journal back in 2001, Markku Eskelinen noted that [here’s a cached version if game studies is still down],

In narratives and many other kinds of fiction it is acceptable and sometimes even preferable that users are misled by being given wrong instructions. But in games the deliberate frustration of action seems clearly to be an intolerable option. One might think of unreliable maps giving false and incorrect information about the location of the player or of the objects he’s seeking – that’s something almost every writer would like to do, and almost every player and game designer to avoid…

That may have been the situation in 2001 however in 2010 it could be viewed as a sign of maturity that there are now many games that have attempted, some even quite successfully, to trick the player by employing unreliability, and specifically an unreliable narrator.

The first example that comes to mind is Bioshock (2007). David Carlton notes in his piece on the game from December “[when] you meet Tenenbaum for the first time; she makes a case that “rescuing” the little sisters is good for them, but does so in a context that paints her as an unreliable narrator.” A commenter by the name of Julian also notes on Mitch Krpata’s Games of the Decade post on Bioshock that “Atlas is practically the definition of an unreliable narrator.”

Ken Levine admitted in an interview with ShackNews that he considered and deliberately employed this technique in Bioshock. Of course, Bioshock wasn’t the first game in the ‘shock series to use an unreliable narrator, System Shock 2 (1999) had a notoriously unreliable narrator in the form of Shodan pretended to be a completely different person for large sections of the game. Kieron Gillen’s ‘The Girl Who Wanted To Be God’ is an excellent treatment and discussion of the routinely untrustworthy Shodan. You may have noticed the fact that System Shock 2’s release predates Eskilenen’s quote from Game Studies. This can be forgiven, however, for the dual reasons of the lengthy process of peer reviewed academic publishing and the fact that SS2 was, according to Giant Bomb, more of a ‘cult hit’ than commercial success.

No such excuse, however, exists for the failure to consider Final Fantasy VII (1997) and its remarkably unreliable and often disjointed narrator Cloud who is also the main player controlled character. Kurt Kalata references FFVII & Cloud’s mental delusion/confusion in the middle of a mammoth 20 page history of the Japanese role-playing game (JRPG). Kalta claims it was as one of the first implementations of an unreliable narrator in a JRPG.

Speaking of the Final Fantasy series, a reading of Final Fantasy VIII (1999) that interprets the vast majority of the game as a delusional dream occurring as the protagonist Squall dies could also be considered an implementation of an unreliable narrator.

Gregory Weir has noted that “Many games have unreliable narrators, especially in survival horror games where the player character is being affected by mental influences.” To preserve the brevity of this essay I’m going to list some of the other recent games and link to where it has been suggested they employ ambiguous or unreliable narrators; Braid (2008) in more than just the one place; The Half-Life 2 (2004) modification ‘Dear Esther‘ (2009); Emily Short posits that it is often used in Interactive Fiction works; Trent Polack has discussed unreliable narrators for Gamasutra and a commenter mentioned Portal (2007) as being another; engaging with almost any ARG ever made also involves some level of player questioning of facts presented to them (See Chris Dahlen’s “They’re here. They’re fake. Get used to it” for a great discussion of ARG’s) and if that seems like a bit of a stretch: when an ARG becomes a videogame blog and people as cluey as Michael Abbott and Simon Carless are fooled I’m willing to say that we’re dealing with some kind of an unreliable narrator.

It would not be entirely wrong to criticise my line of reasoning so far by suggesting that it has not quite addressed the specific issue Eskelinen raises. Judging from the rest of his article, Eskelinen is generally referring to games qua rules & systems and not their interpreted or perceived narrative meaning. With his mentioning of “unreliable maps giving false and incorrect information about the location of the player or of the objects he’s seeking” it’s clear that Eskelinen is talking about the seeming impossibility (or at least that it’s something “almost every player and game designer [would like] to avoid”) that mechanics, even the very geography of the game, could ever be ‘unreliable’.

STALKER: Call of Pipyat (2010) is less the unreliable narrator than it is unreliable geographer. Two key features of the game make it so – the first is the mutable nature of the location of dangerous anomalies in the zone. Whenever an emission happens the dangerous anomalies (predominantly invisible except to a specialised scanner) rearrange themselves in new patterns, theoretically ruining any safe routes previously learned. In practice, however, the anomalies are still confined to specific areas of anomalous activity limiting their potential for geographic ‘unreliability’ and one hardly ever ventures into such an area without a scanner anyway.

The second feature of the game is one particular section of underground geography that literally defies the laws of physics. Entering through a concrete door in the side of a hill, the player enters a bunker connected to a series of traversable ventilation shafts. Successfully navigating these poorly illuminated shafts (often populated with small, fast moving enemies) leads to an eerily deserted room with an open ceiling grate overhead, above which fly the games ever present crows. After exploring the empty room the player can continue on to the next, only to be faced with an identical and similarly empty room. The player continues, finding empty room after empty room, all identical, one after the other. The player consults her map, notices that she hasn’t gone very far from the entry in the side of the hill in spite of having passed through several large empty rooms…

And then the penny drops. The player turns back rather than pressing on and instead of passing through several identical rooms finds herself immediately back at the ventilation shafts. Suspicions confirmed, she tests moving forwards again this time watching the mini-map. It is the only giveaway that the player ever makes a jump through space – back to the entry point of the room she has just left. Otherwise it is visually identical and completely connected but in the games (virtual) reality she is translocated every time she attempts to pass a certain spot.

I can only speak for myself and my own experience of this area but the revelation was certainly not one that I would want “to avoid” but instead wanted to cherish and share with friends. It was an exciting personal discovery and made perfect sense in accordance with the strange paranormal logics of the STALKER universe.

To conclude, there are clearly a myriad of games from before and since Eskilenen’s article that address or deploy unreliable narrators and considering it’s desirability in writing it’s no small wonder that there are so many. However it would seem that it is much rarer to find one that is geographically unreliable, or to find one willing to explore space in such a deliberately and strategically disorienting manner. It’s seems indicative of a highly advanced gaming audience that is able to make and a similarly accomplished developer to create this kind of unreliability, even if it is in such a limited quantity. Regardless of other factors, the fact that such a game even exists is a positive sign.