I am (co)designing a pickup-and-deliver strategy game.
Extensive playtesting has revealed what I believe to be a slightly broken aspect of the endgame (basically, there comes a point when everyone can visualize the rest of the game, and there's nothing to be done to change the outcome/winner). The first 45 minutes of the game play smoothly and are quite enjoyable; it's only the last couple minutes that get stale when the end is in sight.
Knowing that "KC" will likely be submitted to an established publisher - and expecting their developers to take over - we have only taken the theme, artwork, and rulebook so far.
Will a publisher accept games that have slightly broken aspects to them, or is the designer responsible for having a mechanically polished game at the start of the submission process?
Consider Settlers, but without development cards (and also without longest road, as those points are temporary). This hypothetical game would be "broken" in the same way that mine is - not to mention missing out on what is a great aspect of the game.
There would be no hidden information. A king would rise, and at some point everyone would be able to see the same writing on the wall: if something extremely unlucky doesn't happen, the final victory position of each player will remain unchanged.
So in terms of this hypothetical Setllers game, my original question might be reworded:
Have any games been submitted to publishers with noticeable problems that were subsequently fixed by the developer's addition of a new mechanic? (If so, what are they, and who are the publishers?!)
Edit 2: I believe the solution to lie in an alternative scoring mechanism that introduces hidden information and (more) variance/randomness. Rather than advancing a certain number on the VP track for completing certain tasks, basically players will take a tile corresponding to the task they completed. The tiles will have various VP's on them with a certain distribution (think Jaipur bonus tiles).
A simple example: Ticket to Ride. In TtR, you immediately score 15/10/7 etc. for claiming routes of length 6/5/4. If, instead of scoring this immediately, players took tiles from the 6/5/4 "stack" upon completion of corresponding routes, this would be like my game. The tiles for a "6" route would probably range from 14-17 VP each, "5"s would range from 9-11, etc.
The larger issue remains (for future games of mine and of others): how much should the designer do before submitting to a publisher?