The other problem was that the positions of tiles meant nothing. You could use any tile from anywhere as part of your word and it would turn your color, so the game was very one-dimensional. I figured I needed to give the positions of tiles meaning, so I came up with a new scoring system: rather than one point per tile that was your color, an individual tile would have a score based on the number of neighbors of the same color. So a single blue tile would be worth 1, but a blue tile with a blue neighbor would be worth 2, up to having 4 blue neighbors being worth 16. With the new scoring system there was incentive to clump your tiles together, and to break up your opponent's tiles, so where you played mattered just as much as what you played.
This system had a fatal flaw however. Every turn would result in the score swinging wildly back and forth. So you'd get into situations where there would only be a handful of tiles left, and using them would leave you vulnerable elsewhere, and you couldn't build up enough of a lead to be able to stretch out and close the game.
So I went back to the drawing board and killed that scoring system. Instead, I gave tile "positions" a different reason for being: neighbors could be used as defense. With a defense mechanic it would be possible to stake out an area of the board and not have to worry as much about points swinging so wildly, making it easier to build up a big enough lead so it would be possible to stretch out and close the game.
Did you lean on certain conventions of existing word games for your mechanics e.g., letter distribution? Where did you break with the past?
I ran some analysis on my dictionary to come up with the letter distribution. I'll admit I didn't look to other word-search games for ideas, I'm sure the distribution is pretty different than Boggle. I keep the consonant generator separate from the vowel generator, and always guarantee at least a certain number of vowels. I don't want too many or the game goes too quickly. If there's a "Q I" guarantee an I, as even a "U" wouldn't guarantee the "Q" would be playable.
Letterpress uses Game Center differently than most iOS games I've seen. Tell us a bit about the integration there.
Letterpress is one of the few apps to use Game Center for turn-based gaming. Apple made that API available starting with iOS 5. It handles pretty much everything, matchmaking, game state hosting, notifications. It's nice that I don't have to run my own servers. I noticed things start to buckle around 24 hours after launch, right when things were ramping up like crazy. It's been a few weeks and things have gotten a ton better.
What are some favorite traditional board games? Anything you've started playing recently?
It's hard to beat chess. Nothing recently, though I have been waiting for a chance to try Cards Against Humanity.
Could you see tablet and phone apps gradually replacing traditional table games? Would you ever want to design a physical game?
Letterpress might be fun as a traditional physical game. Though I'm not sure how tedious it'd be to keep track of protected states and swapping colors and all that. I don't know if the physicality of the board and the pieces is as important as actually interacting with people around you through the game — something like a coffee-table-sized 50-inch iPad would be more up my alley. You could build some neat games (and apps) with that.