The dilemma of open development and polish

We are big advocates of "open development": the philosophy of sharing all our information about Wolfire and Overgrowth with the public as often as we can. Each week we release whatever we happen to have in the SVN repository to people who have preordered the game. This has treated us well for our early development process, and I'm confident that this approach is responsible for the majority of our preorders and general support. However, in the later stages of development, is it always good to be open?

A movie analogy

It's not uncommon for film studios to release early information about movies they're shooting. They leak information like cast lists and pieces of the plot, and give out posters and other promotional materials. Sometimes they release early footage of scenery, costumes, and perhaps cast interviews. However, have you ever seen a studio release a 400 minute, uncut block of video with no special effects or editing to the general public? Would you actually want to follow the director through his countless edits and painstaking efforts to bring his film to final production quality? Such an inside look at the process would be similar to late-stage open development -- something we will have to deal with soon.

The uncanny valley

The uncanny valley theory for example is an extension of this idea. It states that people tend to feel at ease with intentionally unrealistic human forms like cartoons. They also feel at ease viewing videos and photographs of actual people. However, somehow in the middle as human forms transition from the abstract to the realistic they appear to drop into a valley of viewer discomfort and rejection.

'

How does this apply to game development?

Based on anecdotal evidence we've observed on GameTrailers and YouTube, we sometimes wonder if game development might also follow a similar curve. When a game is early in development, it is obviously just a prototype, and people cut it some slack. However, there is a period in which it looks and feels like a finished game... just a bad one with very little polish. If this is true, then while it certainly makes sense to show as many people as possible the initial positive ascent along the development path, it might also be optimal to go into "stealth mode" for the duration of the "valley phase".

By some measures, the death of Rabbot could be interpreted as our descent into the valley. While Rabbot was obviously not a ninja rabbit, Robbert is. While Rabbot didn't have any combat animations, Robbert does.

The essence of the dilemma is this: What's more important; spreading the word or setting a stellar first impression? Is it even a bad thing if someone curses us out on GameTrailers or YouTube? Would that person be more likely to purchase Overgrowth, had they not been exposed to a development video and does that outweigh the immediate support we get from more open-minded viewers?

Perhaps it's cool that they are even hearing about us at all, and we will be able to win their support with the next video. What do you guys think? Is all exposure good exposure or do we need to start being careful of what we publicly show off on the blog, GameTrailers, and the rest?