Stealth: Confession and Complaint

Recently I mentioned I was playing Alpha Protocol and I quite liking it.  While I haven't changed my mind about that (though I've still not finished it), and I'd still recommend it to people. I'll confess that part of the reason I like it is that it allows me to use stealth.  Unfortunately, they have some issues with it.

One mission in a US facility saw me receive a post-mission debriefing that I'd killed American soldiers.  Yet I'd made a specific effort NOT to do so, incapacitating all of them through non-lethal means rather than getting involved in a gunfight.  Another mission in a train station had a particularly difficult section where there were five guards. As I took out each one stealthily, another guard would magically appear to take his place. After I killed or incapacitated 3 guards, the guard that then appeared was automatically hostile (even though he couldn't see me) and alert his four friends to my presence.  In yet another mission, my handler warned me of a room filled with guards saying that they might have heard the gunfire... yet not a single shot had been fired. That mission also saw me leave the area by tossing an explosive to dispatch an enemy combatant - which was odd given I'd made my way stealthily through the entire level.

Thank goodness people are too stupid to look up.

The problem is that stealth players are perfectionist fanatics. If, as a designer, you offer stealth as an option for gameplay, then you really need to set aside a lot of time for testing. Why? Stealth players are going to find a way to "break" your level design in ways you haven't considered. Stealth players will wait for minutes for the timing of patrolling guards to match up so that they can get from one hiding place to another without being seen.  When you add in mechanics that allow the player to "go invisible" (which Alpha Protocol does), they will find a way to get around without raising an alarm. What's more, if they do get spotted, they're probably going to reload and try again. And over, and over, and over again, until they get it "right".

This kind of perfectionist behaviour lead to the user-created "ghost" rules for the Thief series of games. There are a few variations, but this is one:
1) No guard must see you, or even be alerted to your presence. If they so much as say "Somethin' there?", that's it, you've failed.
2) You must not deal or take damage of any form to anyone or anything. (Whether animate or inanimate)
3) When you leave the level, everything must be as you found it. Opened doors must be shut, doused torches must be relit, etc. (If there was a mechanic for relocking lockpicked doors, that would be in there too, but there's not, so it isn't)
Needless to say, trying to finish a level in a game (even one designed specifically for stealth like Thief) in this manner is a significant challenge, and highly time consuming.

Playtime's over. Time for bed.

So now I must confess that I'm one of the fanatics who went through Thief games playing as a ghost, which I why I deplore the use of some of the cheap tactics employed in Alpha Protocol.  No, I don't expect to be able to "ghost" in Alpha Protocol, though I'll confess I'd love if it was possible. That said, I have no problem being forced out into a firefight in the open. That's fine, because there are some circumstances in which that has to occur. But magically respawning enemies? Unacceptable. The incapacitation issue seemed like a bug rather than design issue, but the assumption that I was going to be fighting my way through one particular level was again unacceptable.

However, the highlight (or lowlight?) was the first "boss" I faced. With an objective to reach him without being spotted, I immediately decided I was up for the challenge. Unfortunately, the game wasn't. It took me numerous attempts to figure out a means to eliminate his guards and get behind him without alerting him, but I persevered despite some ridiculous AI quirks.  When I finally succeeded, I approached him from behind and pressed the "A" button on my controller to take him out. Only it wouldn't work. This "boss" was immune to an instant kill sneak attack... something that had worked on every single person so far... why was this man immune to death by knife from behind? What's more, is that it didn't seem to be possible to dispatch him in a single burst of silenced gunfire either. Maybe I was missing something, but I could not find any means to dispatch him without him at least being able to turn and return fire for an instant.

So much for the subtle approach...

The reason I considered this such a significant issue is because I was presented with a seemingly unachievable goal. The game explicitly offered stealth as an option, and in this case it was even stated that was the preferred approach. Yet the implementation means that it is not possible to complete using the mechanics that the player is already familiar with and has been using successfully throughout the game up until this point. As a designer, you cannot nullify a tactic presented to the player without a clear indication and reason for doing so. Is the fact that the enemy is a "boss" sufficient to justify their inability to use a stealth takedown? Especially when you're specifically told that you should take him down without being seen?  I considered not.

Maybe I'm overreacting. Maybe this is just sour grapes because I couldn't stealth my way through a difficult encounter. Or maybe I'm just annoyed because a developer offered stealth without considering the playstyle of a stealth fanatic like me. I've said before that you can't cater for the few for the sake of the many, and I stand by that. My issue is that Obsidian took deliberate measures to prevent stealth from being used, actively preventing a tactic rather than failing to cater for it.