Category Archives: Games

Exceptional Letterpress Strategy

Posted on by

There are numerous strategies (or “tactics” or “rules,” whatever) that can help you win Letterpress games. I wrote an article, previously posted here, detailing several such guidelines.

Once you begin to master these guidelines, the next step is to realize that there are situations where the winning approach is to ignore the conventional wisdom and make an exception to the rule.

In this article, I explore three such exceptions.

Get off the “see-saw”

In games between closely-matched opponents, the mid-game often boils down to a swinging back and forth, where the overall position remains largely unchanged. Typically, there is a “neutral zone” line of squares dividing two areas of the board, each area largely controlled by one of the opponents.

Both players naturally go after the neutral zone squares. However, as these letters are adjacent to the opposite player’s protected squares[*] and/or to unclaimed squares, there is no way either player can protect them. This means each player will, if possible, retake the lost squares on his next turn. That’s what makes it a neutral zone. Nothing much changes over a series of turns. Metaphorically, it’s like being on a see-saw.

Game 1

A good example of a see-saw situation can be seen in the game depicted in Figure 1. To replay the entire game, click here.

Figure 1: By playing JUMPSUITS, I got off the see-saw.

After coming up with SURVIVALISTS for my third move, I thought I had the game just about won. However, one obstacle remained. The JPF squares in the lower-right corner remained a “third-rail.” Neither of us could safely touch them. This was because: (1) there was no single word that contained all three letters and (2) playing a word that contained just one or two of the letters would allow the other player to take the remaining unclaimed square(s) and almost certainly win the game.

So, instead, we went into see-saw mode. We eventually focused on a neutral zone consisting of a vertical line of squares containing RSPTU, plus the I and the S in the last two columns. You can clearly see the see-saw effect in the sequence from UPSTIRS to PLURALISTS shown in the first three panels of Figure 1.

There are two ways to successfully get off these see-saws.

The first and more common way is to outlast your opponent. This means that you are able to come up with one more “neutral zone” word than your opponent. Once your opponent fails to match you (and thus leaves one or more neutral zone squares untouched), you take advantage by expanding your territory in the newly accessible area. Eventually, you hope to gain enough territory that you can comfortably take “third rail” letters and close out the game with an easy win. Even one misstep may be sufficient to lead to a loss; two missteps are almost certainly fatal.

This is the “conventional wisdom” strategy because it is a “safe” one in the short term: you maintain the status quo while waiting for a sure advantage. However, it contains a long-term risk: you may be the one who runs out of words first, ceding the advantage to your opponent. Since you usually can’t tell in advance which way things will go in the end, it is functionally akin to a toin-coss. Heads you win, tails your opponent wins.

I prefer better odds. That’s why, in these situations, I am always scanning for the second way to successfully get off the see-saw: Rather than wait to see who runs out of words first, be pre-emptive. Play a word that contains one or more of the “third-rail” letters and yet guarantees a quick win. The key here is knowing when this is likely to succeed.

In this game, my pre-emptive word was JUMPSUITS. This word claimed the “third-rail” J, as seen in the final panel of Figure 1. After a conventional response from my opponent, I could play a word such as SPIRITFUL on my next turn and win the game. My only concern was whether or not I would get a “next turn.” The risk here is that my opponent might come up with a word that included P and F, plus sufficient other squares, to give him an instant win. Either way, the game was going to end on the next round of play. The see-saw ride was over.

For those of you familiar with the game of Hearts, I consider this pre-emptive strategy to be somewhat similar to “shooting the moon.” If it works, you win spectacularly. If not, you lose spectacularly. That’s why you want to be confident the move will succeed before embarking down this path. However, your confidence need not reach 100% certainty to justify taking the risk — which means you may lose on occasion.

In Letterpress, before I make a pre-emptive move, I test out all the possibly dangerous return words I can think of that my opponent might play…and count out the resulting score. If any word gets him to 13-12 or better win, I abandon the move. Such an analysis can take a good deal of time, and there’s no guarantee that you will find every relevant word. You just do the best you can.

In this game, after my opponent played PLURALISTS, I looked at playing JUMPSUITS and searched for any words that my opponent could reply and get a win. I could not find any. So I went for it. It paid off. My opponent played SPIRITFUL. Close, but no cigar. He ended the game — but I still won with a score of 13-12!

Game 2

Another good example of this technique can be seen in the game depicted in Figure 2. To replay the entire game, click here.

Figure 2: By playing GRUMPILY, I got off the see-saw.

After playing MATERIALIZING on my second turn, the game had already advanced to the point that the only remaining unclaimed letters were the middle PZW squares. However, these were now “third-rail” squares. For the next several turns, neither of us would risk claiming any of them. Instead, we focused on a neutral zone that consisted of the ring of squares that surrounded these three letters. [As an aside, I did not mean to play RHYTHMIZES when I did, but that’s a story for another day.]

After I played THERMALIZING, my opponent responded with FRAGMENTARILY. Despite being a 13-letter word, it actually hurt his position overall — by leaving the H as blue. The safe move here would have been for me to respond with a word like LADYFINGERS, reclaiming all the neutral zone squares (except for the M in this case) plus extending my position by turning the now exposed G in the top row to blue. However, this would not tilt the game decisively in my favor. A prolonged battle remained likely.

So, after careful consideration, I instead chose to “shoot the moon.” I played GRUMPILY, claiming the third-rail P (as seen in the third panel of Figure 2). This also turned both the G and the U in the upper right corner to blue, as well as newly protecting the R and the H squares, leaving me with nine (9) protected squares.

I knew I could win on the next turn with any of several words that contained Z and W. Would I get that chance? Not exactly. But again it didn’t matter. Similar to Game 1, my opponent, realizing his hopeless position, fell on his sword by playing TWIZZLING. This ended the game but still left me with a 13-12 win.

Game 3

As I said, even after a thorough analysis of the position, you can still lose by going for the pre-emptive move. This happened to me in the game depicted in Figure 3. To replay the entire game, click here.

Figure 3: By playing DISHEVELED, I got off the see-saw — and was sorry I did.

After my opponent played BLEACHINGS, I believed I would have won the game by playing BEKNIGHTED. However, I didn’t see this word until after I played BLIGHTED. This led to a see-sawing over the next several turns, involving the vertical column of GDWIS, with some looseness for control of the D and W.

When my opponent played HEADWINDS, it was the first time in this sequence that he allowed the G at the top of the second column to remain blue after his turn. This gave me what looked like a potential opening for a quick pre-emptive win.

I did consider the “conventional” safe moves of either SANDWICHED or BANDWIDTHS. As things turned out, I wish I had played one of those words. Even though this would have maintained the see-saw, I stood a good chance to win in the end.

But I did not know how things would turn out. So I instead played DISHEVELED, finally claiming the V, one of the 4 “third-rail” letters. I had spent a good deal of time checking out the consequences of words my opponent might play. None of them led to a victory for him. However, my opponent surprised me with BACKFLIPPED, a word I had not considered. It gave him a 13-12 win. It might be the only word in the dictionary that allowed him to win. But that’s the way it goes. I still think it was worth the risk.

Despite the occasional loss, I consider the pre-emptive strategy to be one of the most important weapons in my arsenal. You won’t always have an opportunity to use it. You certainly don’t want to take such moves recklessly. But I’ve found that good opportunities come up quite often in these see-saw situations. And when you do take the risk, and it works out like you hoped, it’s a great way to win. It will certainly surprise your opponents.

Win with one corner

Conventional wisdom, as covered in my previous article, says that protecting a corner is a good thing. I obviously agree. If protecting one corner is good, it should follow that gaining possession of more than one corner is even better. Generally, this too is true.

However, the quality of a protected corner position also matters. A strong corner position is one in which you have several, say as many as 7 or 8, connected squares extending out from the corner, at least three of which are protected. A weak corner position is where you only have 1 or 2 protected squares, usually one being the corner square itself, with little or no extension.

This leads to the exception to the “rule” about corners: Possessing one strong corner can often be superior to having control of two or even three weak corners. On the surface, this might seem fairly obvious. However, you may be surprised by the extent to which this logic can apply. Sometimes, even a single fairly weak corner can be sufficient to win a game.

As an example, look again at the game depicted in Figure 3. For most of the game, I had possession of two corners, a strong one in the lower right and a weak one in the upper left. The corner in the upper left may have been weak but it was critical. If I let it go, my opponent would likely quickly gain control of the entire left side of the board and march to an easy victory. In contrast, my opponent possessed only one corner, the lower left.

By playing DISHEVELED, I gained a foothold in the upper right, giving me a degree of control in a third corner. Meanwhile, my opponent seemed to be struggling to maintain his lone corner. And yet…with his next move, my opponent went on to win the game!

Pass up a corner

Your opponent goes first and protects a corner. Now it’s your move. As I covered in my prior article, ideally you want to both weaken your opponent’s corner and establish at least one corner of your own. However, this is not always possible. Sometimes, you have to choose between these two goals. When this happens, the question becomes: which goal to pursue first?

The conventional wisdom is to prefer to set up your own corner first. I agree. The rationale is that, if all you do is weaken your opponent’s position, he can often undo your damage as well as expand his advantage on his next turn. You wind up worse off than before your move. At least, if you protect a corner, it cannot be entirely undone by your opponent in one turn.

Once again, however, there are exceptions. If it looks like, left unchecked, your opponent can quickly gain an insurmountable advantage, you need to go after his corner. Occasionally, the rationale may be more subtle. For example, take a look at the game depicted in Figure 4. To replay the entire game, click here.

Figure 4: For my first move, I decided to attack my opponent’s position rather than establish a corner of my own.

After my opponent opened with JAUNTY, I could have responded by taking the lower left corner with WANTONLY. However, this fails to touch his upper right corner. Alternatively, I could play JUNKET and dismantle his corner, but fail to establish a corner of my own. I couldn’t find a word that accomplished both goals. Which to do? In this case, I went against the conventional wisdom and opted for JUNKET.

Why? Because I knew that by playing a word with a J, my opponent would similarly have to follow with a J-word — if he wanted to continue to protect his corner. As far as I could tell, there weren’t any J-words that would allow him to improve his position significantly. So I felt safe in not establishing my own corner immediately. After he played GLUTTONY, abandoning taking a J-word, I finally took the corner with WANTONLY.

Bottom line

The “conventional wisdom” move is the one you should usually take — almost by definition. However, the common thread in all of the above games is that the best moves were ones that did not follow the conventional wisdom. The specifics of a given board position will determine when the time is ripe to make an exception. Your goal is to always be on the look-out for these “exceptional” positions. If you can reliably spot them, and take the appropriate action, you are certain to increase your winning percentage.

—————

* The term “protecting” means surrounding a square with other squares of the same color so that the square turns the darker shade. This means its blue/red color cannot change on the next turn. “Protecting a corner” or “establishing a corner” means protecting a corner square. A typical strategy is to expand this corner base, on subsequent turns, by protecting adjacent squares.

[Update: Several minor changes, for clarity, have been made since the article was first posted earlier today.]

Win at Letterpress: Start Second, Finish First

Posted on by

I like Letterpress. A lot. In fact, it is my favorite new game since Angry Birds. For me, it is a nearly perfect merger of my dual interests in word puzzles and strategy board games. My hat is off to developer Loren Brichter for creating this delightful app.

I’m not going to review the basics of the game here. I figure that, if you don’t already know how to play, you’re not going to be reading this article anyway. If you do prefer a review of the rules and essential strategy, I highly recommend Josh Centers’ Letterdepressed in Marco Arment’s The Magazine.

My focus here is on how to play when you go second after your opponent has made a great opening move. In this regard, Josh Centers writes:

The first move in Letterpress confers a huge advantage. A well-played opening can devastate your opponent. If you’re opening the game, always defend a corner letter and make the longest word you can.

If you play following an opponent’s really great opening, you’re at a disadvantage, but not an insurmountable one. Like Microsoft in the ’90s, you want to “embrace and extend.” “Embrace” the opponent’s letters by using as many as possible, and “extend” by using unclaimed letters, preferably taking another corner as you do so.

However, the current first-mover advantage might be short lived. Developer Loren Brichter told me that he’s considering adding a “pie rule,” which would allow the second player to veto the opening move.

While Josh acknowledges that a great opening is not an “insurmountable” advantage, it sure comes close to sounding like one. If not, Loren wouldn’t be considering a “pie rule” (which I hope he doesn’t do). While it’s certainly an advantage to go first, I wouldn’t be too concerned — no matter how big the advantage seems.

I have currently won my last 60+ games. In almost two-thirds of those games, I played second — usually after my opponent got off to a solid start. Yet, in every case, I won.

There’s no big secret to how to do this. Essentially, I followed the advice outlined in the quote above. However, going from abstract advice to practical implementation may prove a bit tricky. That’s where a game replay can help.

What follows is a move-by-move analysis of a tightly fought game, explaining the thinking and strategy that went behind each move. I also include briefer analyses of two other games. My hope is that these annotated replays can help develop your own skills.

Game 1

As the figures below do not include every move of the game, you should ideally follow along with the full replay of the game.

Move 1. My opponent opens with the word DRAPERY, leaving the strong position shown in the figure below.

Game 1-1

She has solid control of the northeast corner, with two protected (dark red) squares.  It’s an especially strong start because the corner contains the letters d, r and e. These are desirable letters to re-use for long words — as they form the suffixes -er and -ed. For example, you could turn a word like SLAM into (with the second M) the much better SLAMMED or SLAMMER.

After a start like this, I assume the corner (perhaps 6-8 squares) will still be in my opponent’s color at the end of the game. I may be able to do better, but I don’t count on it. The good news is that, even with the corner lost, I can still theoretically win by at least 17-8. Accomplishing this, however, will require playing catch-up for most of the game and being very careful not to make mistakes. One or two, even minor, errors can quickly turn a disadvantage into a sure loss.

Move 2. I give a great deal of thought to my first move. Unless an obviously great word presents itself, I typically spend more time on my first move than any two or three moves for the rest of the game.

When playing second, I do my best to have my move accomplish two goals simultaneously: (1) Force my opponent to defend their advantage and, if possible, (2) establish a corner of my own. Ideally, this requires my opponent to fight back on two fronts. By continuing this dual pressure over several turns, I hope to eventually build an advantage while limiting the ability of my opponent to expand their lead.

After I play SLANDERS

I attempted to accomplish this goal in this game with the word SLANDERS. It gave me the desired foothold in the northwest corner plus turned every light red square in the northeast corner to blue. Perfect!

Moves 3-5. My opponent came back with BREADY. This looked pretty good; it re-established her control of the NE corner, even extending it a bit. However, it failed to undo my NW corner foothold. This was a significant oversight in my view— and I quickly took advantage of it.

My opponent would have likely done better if she had played BLANDER. This would have had the added bonus of turning the N and L in my corner to red, forcing me to work much harder for a good reply move.

I played REBRANDS. This turned almost the entire top two rows to blue, including extending my dark blue squares from one to three. It also attacks the NE corner again. Overall, a very good move.

I was almost sure my opponent would come back with BRANDERS, using the exact same letters in a different word. However, since the SNR squares were now dark blue, playing them would not help her. As such, I thought I would still be ahead after the exchange. Not a great exchange for me, but the best I could see at the moment.

As it turned out, it didn’t matter, as my opponent played BLARED.

Moves 6-10. With BRAMBLED, MARBLED and MANGLED, we are pretty much treading water. I gain some traction with my move and my opponent reverses the gain with her move. When she played SPAMMED, I began to feel some additional heat (see figure). By moving into the southwest corner, especially by protecting the M, she was threatening to obtain control of the entire corner region. If she succeeded, she would almost certainly win the game. I was in trouble.

After my opponent plays SPAMMED

After much experimentation, I came up with SWORDPLAYERS. It accomplished my ongoing key goal of simultaneously attacking and defending. In particular, it protected the O (now dark blue) in the SW corner, thereby removing the danger at least for the moment.

Moves 11-15. My opponent came back strong with WORDPLAYS, using most of the letters I had just played. I returned the “favor” by playing SWORDPLAYS.

At this point, all the squares in the first two columns are blue except for the two M squares. If I could retain the eight blue squares plus add the two M squares, the entire first column would be protected (turn dark blue). More often than not, this translates into an unstoppable win. Would my opponent play a word that allowed me to do this?

She played BADGERS, an excellent comeback. It stopped me in my tracks. She attacked the G and S in the first column plus protected the M. Not at all what I was hoping to see.

I believed my opponent now had the lead. In fact, if it were possible to switch sides here and my opponent asked me to do so, I’d probably say yes.

The best I could think of for my next move was BADGES, duplicating the letters of her just-played word except for the R. For obvious reasons, I never like playing a word inferior to what my opponent just played. However, in this case, as all the R squares were either blue or dark red (protected), playing one of them would have made no difference in the position.

My opponent came back with BLADES. This was a huge error, in my view, because it left the first two columns exposed.

I believe my opponent would have done better had she replied with DEBAGS (using the same letters as BADGES) or even BARGED. Either of these would have kept an M protected.

As it stood, I at last had my chance to protect the entire first column — if I could come up with a word that used B, M, M and S.

Moves 16-20. At move 16, I played BAMMERS. I wasn’t even sure it was a word when I submitted it. But it is, because Letterpress accepted it. Bingo! For the first time, I had confidence that I would wind up winning the game. I would now be able to go on the attack more, with my opponent being on the defensive.

With BOMBARDERS, WARMONGERS, and SOMBER, we spent more time treading water. I tried to solidify and expand my western wall. My opponent tried to stop me.

At move 20, I had a major decision to make. I could have probably quickly finished and won the game, by playing a word like DAZZLERS. This would have turned both Z squares to blue and given me possession of the first three(!) columns. In retrospect, I believe I should have done exactly that.

After I play ROBBERY

However, I was greedy. I was now thinking not only about winning, but about winning with a crushing margin of victory (not very friendly, I know). So I instead went with ROBBERY. This turned every light red square on the board to blue, leaving my opponent with just the three (previously all dark) red squares in the NE corner. This seemed a potential crusher, but I was wrong.

By the way, I wasn’t worried about my opponent filling in the four unclaimed squares with a word that would end the game and give her a victory. I am almost certain that there isn’t any word that contains Z, Z, X, and K — certainly not one long enough to give my opponent a win. Having the unclaimed letters be uncommon ones was working to my advantage here.

Moves 21-25. With PASSERBY, my opponent gave me unanticipated trouble, causing me to regret my previous play. By turning the Y to red, my previously protected O in the SW corner was in jeopardy. My next word would need to include a Y, in order to re-protect the O. I had not expected this. So I played PANDERLY. With PRAYED, and BEARDY, we see-sawed again.

My opponent then gave up going after the Y square and played WORMED. I’m not certain whether this was a mistake or not. But it gave me the opportunity to go on the attack again.

Moves 26-30. I went with SWAMPED, bringing me back to about the same place I was after playing ROBBERY. My opponent played PREBOARDS, again leaving the Y untouched. I was now ready to pounce. I played ZAPPED, at last gaining possession of the three left columns.

After I play ZAPPED

If my opponent had any chance of winning, she lost it with SPARROW. With this word, she gave up control of the P in the NE corner, the location she had guarded since her very first move. Although there weren’t any great choices for her, this was perhaps the worst one. I can only assume the move was a mistake; she failed to see the consequences until after she had played. It happens.

I countered with WRAPPERS, leaving a score of 21-1.

Moves 31-end.  The next several moves are relatively uninteresting — with the two of us exchanging similar words such as SWAMPY and SWAMPS. Essentially, I am jockeying for a maximum win position, while my opponent is trying to hang on to as many squares as she can. With BOXERS, I claimed the X. With MAWKS, I claimed the K. The game was about to end.

After I play BLAZERS, victory!

My opponent responded with WREAK. I was surprised. I thought she would play a word with Z, finishing the game even though she would lose. Instead, she left me to finish the game with BLAZERS — handing me a 22-3 victory.

Game 2

This second game demonstrates the same principles. Here, my opponent starts off with the NW corner and retains it till the end. I initially fight back by gaining control of the SW corner (see figure below). As the game develops, the outcome hinges on who will eventually possess the eastern end of the board. That turns out to be me, and I win 17-8.

After my opponent plays FREIGHTERS

Game 3

The third game is the shortest of the trio, lasting only thirteen (13) moves. Here, my opponent starts out with STRONGLY, grabbing the NE corner. She will never lose it. I thought I had a near-devastating first move reply with SYMPTOMOLOGY. However, she completely turned the game around with OBNOXIOUSLY (see figure below). Suddenly, I had the sinking feeling that the game might already be lost. Still, I fought on and came back with some good replies of my own. Despite some strong play from my opponent, I was able to secure a 16-9 win.

After my opponent plays OBNOXIOUSLY

Bottom Line

Playing the longest word you can is typically a fine thing to do. Choosing a word that turns the most amount of your opponent’s squares to your color is often a better, quite excellent, thing to do. But neither of these things, by themselves, are sufficient to win consistently. I have seen many boards where one player owns almost all the currently claimed squares, yet winds up in a hopelessly lost position within the next move or two.

The key to winning is to figure out the strategically best squares to claim and figure out the word that best acquires them. In that regard, I often start a turn by selecting six or so squares that I would most like to acquire. I then see what words I can construct that include those letters. Take your time here. Don’t rush to make a move you will regret.

How do you know which are the best squares to claim? This requires an ability to look ahead and see the consequences of your move and the possible retaliatory consequences of your opponent’s next move. Hopefully, this article provides some insight on how to do this. I plan to write additional articles that explore this further, going back to some of the basics. Beyond that, the best way to learn is to play the game.