Early in the process of playing LA Noire, right around when I finished the Traffic Desk, I took some notes on how I was feeling about the game. It’s interesting to look back on them and compare my early impressions to how I feel about the game now, having finished it.
I initially felt that the story and characters were interesting and well-drawn. Now I’m not so sure. I feel like the game was strongest when it was telling one-off stories. The serialism of nearly all the cases in the second half of the game hurts. The villains are too obviously evil or deranged. Phelps himself is never compelling, it’s always the world around him that provokes interest.
I initially felt like the facial animation was incredible, and it still is. I even found myself paying more attention to faces in the real world after playing. Altering the way I perceive the world is a great accomplishment for a game, and LA Noire should be lauded for that.
That said, gameplay that the facial animation generates – the interrogation system – is a failure. It neither produces realistic-feeling interrogations, nor enjoyable gameplay. It devolves into a game of trying to read the developers’ minds as expressed in the facial animation system.
Given that cases still progress regardless of the player’s success or failure in the interrogation sequences, I was puzzled that the game felt compelled to give question-by-question feedback on how the player was doing in the interrogations. It drives a wedge between the player and Phelps, since the player knows there was more information to be learned from the interviewee, but Phelps generally doesn’t. It also tends to trigger a compulsive desire to reload and retry the interviews, which can’t be the designed result. It would be much better if the game left the player to judge himself on how he did in the interview, and then provide more concrete feedback after the case.
I thought the driving and action gameplay was inoffensive, and while it began to wear on me later on in the game, it was still fine. As others have pointed out, it speaks volumes that the game lets the player skip these sequences. How much worse would the game have come across if the action sequences couldn’t be skipped, and if there was now fast-travel system.
I initially felt like it was an amazing feat that the cases were such that they couldn’t be failed. But having completed the game, and now being able to see behind the curtain, I’m much more reticent to praise this feature. Too many cases fall back on having the ultimate suspect be located through some means largely independent of the investigation, such as via the coroner phoning in about some conclusive piece of evidence, nullifying whatever investigation was conducted through the first three-quarters of the case. It leaves an unsatisfying feeling when becomes clear that most of what I’ve been doing has no influence on the ultimate outcome.
Lastly, it seems deeply weird to me how rarely Cole actually gets his man. After the Traffic Desk, I’m hard-pressed to come up with cases where Phelps successfully arrests the correct suspect. In every Homicide case but the last one, he arrests the wrong guy, and then in the last one he shoots the guy. I don’t remember Vice very well, but my feeling is that, while he generally fingers the right guy, he ends up shooting him in most cases. And the two Arson cases that Phelps is on, he once arrests the wrong guy, and once shoots the wrong guy. It seems bizarre to me that the game chose to play out that way.
So, in conclusion, it’s a deeply flawed game. It’s at its best in one-off cases, where the investigation is much more open-ended about where it could go, and where the villain is much less obvious (another failing of the Vice and Arson cases – while I didn’t know who did the dirty work, it was obvious from very early on who was giving the orders). I think the turning point of the game can be traced to exactly the point where I realized I had the wrong guy in the early homicide cases. The single biggest hit had to be when it was revealed during the last homicide case that there was, in an earlier case, a major piece of evidence linking the murder to the Dahlia killer (writing on the body, under her clothes) that the coroner had inexplicably withheld. Either this is because the case wasn’t originally designed to be a Dahlia murder and was changed later on in the design, or it’s because the designers were worried it would be too obvious that all the homicide cases were connected. Either way, it was a cheap trick. Actually, “trick” seems like an apt way to describe the game. Under examination it just doesn’t hold up as a satisfying experience.