« 2012 SH novelettes and short stories | Main | Forward motion »

Some Letterpress strategy notes


Letterpress is an iOS word game. I played it a fair bit in spare moments for a couple months, mostly against randomly chosen players. I had set it aside for a while, but a recent discussion on Facebook reminded me that I'd never posted my strategy notes. So here they are. (Wrote half of this in early November of 2012, then set it aside and didn't finish it 'til now.)

(Added later: Nothing here was intended as a comment on any game currently in progress.)

(If anyone wants to play me, my username on Game Center is my middle name; if you don't know my middle name but you know me and want to play, drop me a note with your username.)

  • My general strategy is to start with at least one defended tile in a corner or on an edge, and then on each turn try to restore defenses to all of the previous turn's defended tiles and to add at least one new defended tile. In practice, this often takes the form of growing a pyramid of defended tiles out of one corner.
  • Another aspect of my general strategy: I don't focus on the other person's defended tiles. If I try to attack them, I'll never manage to grow my own far enough to win.
  • Paying the $2 to unlock the view-played-words feature was totally worth it; not being able to view what words have been played makes the game significantly harder, unless you have a better memory than mine.
  • Opening: I look at the four corners of the board. (Unless the other person has gone first and has established at least one defended corner; in that case, I ignore that corner and look only at the other three.) I pick a corner where the four tiles that form a square in that corner can easily be used in a word; ideally in multiple words. If there isn't such a corner, then I may settle for using the corner tile and the two adjacent edge tiles (which still makes the corner tile defended), or for four tiles in a T shape against an edge (making one edge tile defended). I also factor in the other nearby tiles, because I'll be trying to take those on subsequent moves; if a corner has three nice easy-to-use tiles but also a Q and two Ks, I generally choose a different corner.
  • After I start in on a corner or edge, chances are good that the other player will try to take back the tiles I just took. If they do, then on my next turn, I try to retake those same tiles plus a couple more that will let me build my protected region. If I protect at least one additional tile every turn, then my opponent can't retake everything I've taken.
  • If the other player ignores me and goes off to build their own protected region, great! I take the opportunity to expand my region.
  • On any given move, if I can expand my own space and take a tile or two of theirs, great. But if I can't, then I focus on my own.
  • I try to remind myself not to be greedy. In most games of Letterpress that I've played, slow and steady wins.
  • Regarding the actual finding of words: one thing that sometimes helps me is to add extra letters beyond the tiles I need; sometimes that lets me see words I didn't notice before.
  • I try to remember to always look for suffixes before playing a word: -S, -ING, -ED, etc. Sometimes I play goofy plurals that I don't consider real words, just to prevent the other person from playing them. (Made-up example: If I want to play SNEAKING, I might try SNEAKINGS even though I don't think that's a real word, just because I would be really annoyed if the other player played it and took back all those tiles.)
  • Other than that, I don't usually try to play words that I'm not absolutely sure are words.
  • I also try to play for Maximum Fun Quotient in some other ways. For example, I dislike going back and forth on the same word with different suffixes or prefixes; if I play SNEAK and they play SNEAKS and I play SNEAKER and they play SNEAKERS and so on, that just seems boring to me. So I try to never play essentially the same word they played, unless they just did that to me and I don't see any other good options.
  • Sometimes I just can't expand, or even hold, my territory. In that situation, I sometimes just cede a tile or two in hopes of getting it back next turn; but sometimes it works best to instead abandon my original pyramidal expansion and shift to a rectangle-shaped zone, ideally two full rows or columns. Sometimes I even end up shifting to a different corner than the one I started in. This kind of shift doesn't always work, but it sometimes does, and it often seems to surprise opponents, though of course I have no way of knowing whether they're really surprised or I'm just reading that into their actions.
  • If I have a choice between two tiles with the same letter, and one belongs to the other player and one is untaken, and all else is equal (like if neither one expands my region), then I usually take the one from the other player, because that shifts the difference between our scores by two points instead of just one.
  • If I can get to thirteen protected tiles, then I can take risks, because the other player can't win on the next turn no matter what. It's not worth focusing on this as a longterm goal; I don't usually end up with that many protected tiles. But sometimes it can be useful to try for that if I already have, say, nine or ten.
  • Endgame: When the number of untaken tiles drops below about eight, I try to start keeping an eye on them. (Which is tough if I've gotten too focused on building my protected region.) If I can spell a word that uses all of them, I'll likely win. But if I can spell a word that uses only most of them, that's usually not a good idea. If I leave my opponent with only a few open tiles, chances are good that they'll take them all (plus a couple of my unprotected ones) and win. If I leave them an unlikely letter combination, then that may prevent them from winning in this way—but then again, they may turn out to be better at making words with obscure letter combinations than I am.

Do I always win with this strategy? No. I probably win about 19 out of 20 games against random opponents, usually pretty easily. But certainly not always.

How do they win? Well, there are at least four general strategies that I've seen that can beat mine:

  • The Blitz. On every move, play a word that's (say) ten or more letters long. You'll win quickly while I'm building my slow empire. I will be frustrated but admiring.
  • Trench Warfare. On every move, take back all the letters that I took back, plus a few more next to them. This turns into a tedious game of seeing who can come up with one more anagram for those letters than the other player. Eventually, one of us will run out of anagrams, and the other one will win. I generally find these games boring and irritating, but sometimes going back and forth over a border is the only place left to play.
  • Beat Me at My Own Game. If you take another corner, and you build faster/better than I do, you'll probably win.
  • The Surprise Ending. If you notice before I do that the last n untaken tiles can be played in a single word that gets you to 13 or more tiles, then you win.

So there you have it: now you know how I play, and some ways to beat me. I may have to stop playing.


Regarding MFQ, I would also mention that modifying your opponents word takes away some of the Fun Factor - make up your own words for crying out loud! Also, it gives you tunnel vision - trying to create permutations of a word and missing other opportunities.

Also, you mention getting 13 protected tiles and then you can't possibly lose on the next turn. But, that means you have all of the tiles save 7 which is pretty much the entire board. How can you possibly lose in that scenario?

Lastly, all things being equal, when should you take an unclaimed tile or your opponents tile? You say that you should take your opponents tile since that gives you a point advantage. However, I would posit that you should always be looking at positional advantage. Which tile gives you a strategic advantage. Thinking points in an early game is unprofitable.

In general, I agree with your assessments. You would no doubt be a worthy opponent! ;)

I just found this comment from April that fell into the moderation queue. Sorry about that; posted now.

I agree that, for me, modifying your opponent's word takes away some of the fun. Sadly, that doesn't seem to be true for some players.

Regarding number of tiles: It's a 5x5 board, so there are 25 tiles, so if you get 13 protected tiles then there are 12 remaining.

Re unclaimed vs opponent tiles: Yeah, that's what I meant by "and all else is equal (like if neither one expands my region)." Though, sure, there can also be positional advantages that don't expand my region per se. But much of the time, I don't see any useful distinction between two tiles with the same letter, except that one is owned by my opponent and one is unclaimed; in that case, it seems clear that I should go with my opponent's tile.

Post a comment