I've thought about a little piece to sum up what can be learned from Deadly Premonition. So far I've arrived at the following conclusions:
"Tank controls" are outdated. You don't want to use them for controlling a person. You actually don't want to use them at all. I also assure you that you don't want to use them if there's a possible change in the camera angle, like the camera suddenly changing to a position over your head. This change will make the player lose focus for a moment and the momentum is broken since he has to rethink what exactly the game character is facing right now since the point of view is no longer an over-the-shoulder one. If - for a reason I might not be able to imagine and really don't want to guess - you use tank controls make sure the rest of the game fits this scheme or is that good that it doesn't matter. Believe me though when I say it's easier to avoid them altogether. Tank controls are especially bad from a bird's-eye view.
The loading screens in Deadly Premonition are generally okay, except for the fact that they're infuriatingly intrusive at times. It is most noticeable the moment when you get from cutscene to boss fight to cutscene to exploration (might not occur in that order, see end of game). The moment you're eagerly beating a tough opponent into submission in order to progress with the story and have a loading screen block your path. Queue some of those in a short period and a certain level of frustration is almost guaranteed.
I really try to see the other side of the medal here. Sometimes the engine may already present before you start developing your game. Sometimes the engine may be built and somehow abused into doing something it wasn't developed for. Still, the immersion breaking that the loading screens are doing are probably not for the best.
Telling Your Story
There's a significant difference between foreshadowing - which can improve a player's emotional investment in a given story - and spoiling your own game - which can be particularly harmful to a game as storydriven as Deadly Premonition. I'll try to give a good and a bad example directly taken from the game.
Bad: Some trading cards have content and description that might spoil your joy because of the revelations they present at a point of the game where you will almost certainly not know the information presented. You gain insight in personal details of chacters which you might not yet know or gained the trust of.
Good: There are some parts that are inaccessible for story reasons, but the game paints you an entirely different image why things are not not 1.
The gamer playing the Let's Play already pointed out a fatal flaw of the ingame UI - the map markers. I am quite aware that permanently marking some optional important locations apart from the main goal was probably an idea born from good intentions. The current implementation - using differently colored map markers for various targets - is lacking in thoughtful design. That is most evident when having bought all of the Spirit Maps and having cleared all of the related combat areas does not clear the markers from your map.
A solution to the problem might have been a map system which let's the player choose what things he wants to display (or filter out). A fine example can be found in the Assassin's Creed series2. There the player can choose to filter by categories or by map marker type. - Or simply the removal of the markers upon completion of the related objectives.
Consistency is very important3. I'm quite sure that the mismatch of colors between the quest text and the item description in a certain scene is merely a development oversight. I'm however not so sure how that managed to get through quality control because fixing this is probably only as little effort as exchanging one string in the code (or localization files).
Many games which feature mini games deem it to be a great idea to shove them into the user's face at least once per playthrough. While most of the time I'd argue that is not the best practice to force a user to do something that is not considered part of the core gameplay there is a high chance that users might miss such a optional element entirely.
Digital Actor Concept
The concept of using "digital actors" seems quite interesting. On one hand one might argue that it's essentially a recycling of previous material, but on the other hand there are real life actors who have played many, many different roles in their careers quite successfully.
I am not so sure how players of both or all games in which a characters stars feel about this reuse though. It might be interesting to check for emotional investment upon recognition of an existing character with a certain level of depth to his back story and traits who previously starred in other games4.
Purely story driven games without any competitive aspects don't need leaderboards. Maybe you're having fun implementing them and tracking certain things like playtime and the number of times the player shot himself in the foot, but a statistic is enough to show you care and not enough to make people wonder why you'd build a leaderboard around those activities.
I know this is quite vague but I don't want to give everything away during the final analysis. If you want to know, you should either play the game or check out an LP. I'm talking about certain houses or rooms. ↩
Sadly, I'm not quite sure whether all games of the series feature this option. I'm certain that Assassin's Creed: Brotherhood and Assassin's Creed: Revelations do. ↩
Every time I say that it makes me chuckle because this is what caused me to first switch from Windows to GNOME (Ubuntu) and afterwards to OS X. ↩