The Stack

OK, I finally convinced myself that the last Battleships puzzle really is impossible, and gave up on it. This decision hinged on the realization that the puzzle could be wrong without any errors in the row and column numbers. See, most Battleships puzzles start off with at least one tile already filled in, either with a ship piece or empty water, to eliminate ambiguity and guarantee unique solutions. (Annoyingly, the game lets you delete the pre-set tiles just like any tiles you’ve placed yourself, so you have to remember where they are so you don’t do that. The print version doesn’t have this problem.) This puzzle had one pre-set tile, a right-side endcap for a ship. When the game revealed the correct solution, sure enough: that endcap had been placed one space away from where it should have been. In retrospect, it’s a miracle that there weren’t more mistakes like this throughout the logic puzzles, given how many there are elsewhere.

The remainder of the game went quickly. I finished the rest of the Trivia section, discovering as I did that one entire set of trivia questions had been accidentally put in twice under different names, then blasted through the Visual section in a sitting. Mainly it’s more of the visual puzzle types from the first game, this time with color added. In one case, the added color actually interferes with the puzzle: you’re asked to identify the one clown in a clown crowd that doesn’t match any of the others, and the clowns were colored without regard to whether they were supposed to match or not. But then, the scan was bad enough that it was more or less ruined anyway.

There’s a whole lot of puzzles where you have to type in phrases clued by cartoon images, and I’ll note that the typing interface here doesn’t have the problems it did in the Trivia section. There’s three of the sort where you have a line drawing overlaid on a grid, and have to find squares in the picture that match a set of clue squares. This was never all that hard a puzzle type, but the addition of color makes it downright trivial. Also, in the first game, I noticed that the clue squares, which were supposed to be in random orientations, were all in the same random orientation. It seems like they didn’t fix that. Probably they didn’t even notice it.

There’s one Cross Comics puzzle, where you have to arrange a bunch of comic panels into crossing but independent strips, like in Scott McCloud’s Choose Your Own Carl. The first game had one of these as well, although I didn’t describe it before. The UI for it there was terrible, and here it’s slightly less terrible — it’s a little easier to select panels — but still terrible in fundamentally the same ways. The basic problem is that the panels are displayed at a size that’s illegible at the game’s resolution. When you select a panel, a blown-up version is displayed on the right so you can actually read it. However, once you’ve selected a panel, you have to place it in the grid. You can’t even click on another unplaced panel to change your selection until you’ve placed the one you’ve selected. So you can’t read all the panels in advance, and it’s hard to check what you’ve already placed without accidentally rearranging it. Also, there’s one panel already placed, and you can’t select that one, which means you can’t read it. I can think of a simple change that would solve all of this: displaying the blow-up on rollover instead of click.

That done, all that remained was the final puzzle. I may well be the only person in the world to see it, other than the developers.

The main menu displays your options as green circles, and has one spot that looks like an empty depression where a green circle belongs. I had been assuming that this was where the button to access the final puzzle would appear when it was unlocked, but no. Instead, the final puzzle comes up automatically when you unlock it, and also every time you try to play any of the puzzles afterward. That’s right, unlocking the final puzzle locks you out of everything else. I can see a little sense to this: it’s a metapuzzle that’s intended to test your memory of the other puzzles. But you still can’t access the other puzzles after you’ve solved it, except by making a new profile.

The way it works is: At the top of the screen, you get a bunch of pictures from the other puzzles in the game. The first letters of the names of those puzzles form the key to a cryptogram. Again this shows the designers’ ignorance of what they were making. To get to this point, the player has been through an entire section of cryptograms far harder than this one. The key is unnecessary.

When you solve the cryptogram, and enter the correct answer, the game will tell you that it’s wrong. This is because their solution has the word “then” where it should have “than”. “Then” doesn’t even fit their own key. This is pretty much the perfect ending to the game, if you ask me. Just one last really obvious boneheaded mistake before saying goodbye.

It still hasn’t fully sunk in that it’s all finally over. This game is going to haunt me for a while. I knew before starting my posts that the the two games were buggy, but I didn’t realize just how bad it was going to be. But it feels good to have a record of it all. Perhaps it can serve as a warning.

There are really two opposite ways to be stuck on a logic puzzle. The more usual way is to be unable to eliminate possibilities, so that the number of solutions under consideration is too large to simply iterate through. The other way is to eliminate all the possibilities. If you can do that, and the puzzle actually does have a solution, it means there’s an error in your reasoning somewhere, a variation that you’re forgetting about. But knowing that there’s an error doesn’t help much in finding it.

That’s the position I’m in with the last Battleships puzzle. There’s one four-long ship in every puzzle, and this puzzle has only one row or column that can hold it. Since the grid is 10×10, that means there are only seven positions this ship can go in. And I can prove every single one impossible.

There must be a mistake. That much is clear. But is the mistake mine or theirs? I actually think it’s probably mine in this instance. Battleships has a certain amount of error-detection built in. There are exactly 20 tiles worth of ships in every grid, so you can tell if there’s a mistake in one of the row or column numbers by adding them all up and seeing if you get 20. This would fail if there are two or more errors that cancel each other out perfectly, but how likely is that? So I think I’ll have to keep on looking for a solution here.

I can’t say the same for the last remaining Cross Math puzzle. Here, a similar summing technique proved that the puzzle is wrong. In these puzzles, you have to fill the numbers 1 through 9 in such a way to satisfy all six horizontal and vertical equations. Now, observe that the horizontal equations in this one have no multiplication or division. If they had nothing but addition, then the sum of all three rows would have to add up to the sum of all the numbers from 1 through 9, which is to say, 45. But some of the numbers are subtracted. The effect of subtracting a number instead of adding it is to subtract that number from the total sum twice: once for removing it as an addend, and once for the actual subtraction. So no matter what numbers have their signs flipped, the result will differ from 45 by an even number. The sum of all three rows therefore has to be odd. But this puzzle gives us rows that sum to 18.

Having thus convinced myself that at least one of the sums here was wrong, I told the game that I was done with the puzzle, and was pleased to be vindicated: in the game’s solution, the top row makes 8, not the 7 claimed. Proving a puzzle incorrect in this way is a sort of metapuzzle, and at least as satisfying as solving the puzzle itself.

Yeesh, did I really just make a “definition of ‘game'” post? I must really be running out of steam. Doubtless this has to do with my continued inability to solve the very last Battleships puzzle. (And I do mean the very last — I’ve solved the rest of the set.)

To counteract this, I’ve moved on to the Trivia section for a bit. It turns out to be annoying — enough so that I didn’t finish it in one sitting like I had planned, even though it’s half the size of the other sections. Partly this is due to the design. You may have noticed from my screenshots that every puzzle type in Games Interactive 2 has its own logo, with a picture labeled in a zany font. Usually this is something you can ignore, but for Trivia, they animated it and made it part of the game. The picture is a caricature of a game show host, which makes sense, I suppose, but honestly isn’t the feel I’d have gone for. He wiggles his head back and forth while waiting. His facial expression changes with your guesses. When you get something wrong, a buzzer sounds and the light on his podium turns red; when you get something right, a brief latin-rhythmed victory ditty plays. If you have sound turned on, you hear this enough for it to become irritating. If you don’t, you just get a pointless pause for a few seconds before it goes on to the next question.

But the worst part of the Trivia section is the questions where you have to type in an answer. This isn’t always the case — most of the questions are multiple choice. But some are type-ins, and a lot of them aren’t really designed for this game. For example, at one point it asks you for the two ways, other than Lisa’s saxophone tune, that the opening credits sequence from The Simpsons varies from episode to episode. Nobody is going to get that one right, because nobody, even if they know the answers, is going to phrase it in exactly the ways that the computer is checking for. But let that slide for the moment; a lot of the questions are reasonably answerable. The more consistently annoying problem is that the game isn’t very responsive to typing. You pretty much have to hold keys down for half a second or so to make sure they register. This is basically intolerable, and I don’t remember the first Games Interactive having this problem.

It is with some relief that I report that I have finished the Paint by Numbers puzzles, and made some progress on the Battleships. I’ll be closing out the Logic section entirely before long, and that will leave just the three most lightweight categories: Trivia, Visual, and one that’s new to Games Interactive 2: Word Search. I haven’t mentioned the word searches yet, because they’ve never been the focus of my play sessions. And yet I’ve almost finished them.

Let me get this out of the way: Word searches are a joke. They’re the puzzle type that fans of all other types of puzzles look down on. Logic puzzles require real thought, crosswords test your knowledge and vocabulary, and cryptograms combine all of the above, but all a word search requires is the ability to not be bored by word searches. The reason I’ve gotten so many of them done is that I’ve been doing them one at a time between other puzzles in order to avoid having to do them all at once. It’s one thing for a variety puzzle game to have one or two word searches for variety’s sake — one per issue of the magazine was pretty much the right amount, if you ask me — but this game has twenty. That’s more than any other single puzzle type except crosswords.

The word searches from Games Magazine have one real point of interest: The unused letters spell out sentences, usually some quotation linked to the puzzle’s theme. As far as I know, this was never stated outright in the magazine. It was just a little bonus for those who noticed it. The closest they came to spilling the beans was in a March issue with an Irish-themed word search for St. Patrick’s Day: the editor’s message for the issue mentioned that the editor had suggested slipping some Irish names into the unused letters, only to be told “Wait til you see what we’ve already got going on there”. So at that point, not even the editor of the magazine was in on the secret. But that anecdote was enough for me to catch on, and now, the information proves useful: looking for the hidden sentences provides a welcome shortcut to finding all the words, in addition to making the whole process less tiresome.

As you’re probably expecting if you’ve been following these posts, there are problems. There are occasional typos in the word lists: SUBSITUTE for SUBSTITUTE, BARBEQUE where the grid contains BARBECUE. Such words cannot be marked in the grid; the game only accepts what’s in the word list. Occasionally there’s a word that’s left out of the word list entirely. You can tell that it’s supposed to be part of the solution because it interferes with the secret unused-letters text, but it’s just not there. One particular puzzle is missing something like half of its word list. I only have a vague sense of what the unused letters were supposed to spell out in that one. In one of the puzzles, part of the grid was actually misplaced, showing letters one space to the right of where they should have been. It took me a while to figure out what was going on there. Once again, I find myself thinking that some of these bugs could make good puzzles in their own right if deployed deliberately.

The UI is mainly reasonable, but has one biggish problem: words are marked with ovals around them, and these ovals are sometimes wide enough to intersect with adjacent letters, especially when the ovals are diagonal. Now, the puzzles come in different sizes, and the larger grids are fit on the screen by using a smaller font. Throw a couple of lines across the smaller font and it becomes illegible. It’s not hard to think of solutions to this. Like, instead of circling the words, put a line through the marked letters! Ideally, make the line a light color and display the letters on top of it so you can still read them. But I suppose you can’t do that in the print version, so it violates the game’s prime directive.

As before, Paint by Numbers is weekend work. The game has six sets. I’ve done two so far, and expect to manage a third today, leaving the other half for next week. The palette this time is missing the dithered grey that I previously used to mark uncertain squares, so I’ve been using yellow for that purpose instead. At least the two-semantically-distinct-whites problem is gone.

The chief characteristic of Paint by Numbers is that it’s methodical. I described the basic solving process in a previous post: you look for a thing that just barely fits in the space available, which gives you information about its position, which imposes constraints on other things, and so on. It takes time, but you can keep on extending the chain of logic for as long as you keep finding leads. Every once in a while, though, the chain seems to end. Maybe you reach a point where the necessary reasoning takes more steps before yielding concrete results. Maybe there really is a simple next step, but it’s hard to spot. Either way, I wind up marking, in a different color, the things that I suspect rather than just the things that I know.

I don’t like taking this step, because it seems so much less certain than the plodding and methodical reasoning. But is it really? In theory, logic should never steer you wrong, but it’s being applied by a fallible human mind. I frequently do make mistakes. Occasionally, I even make mistakes so severe that I have to correct them by wiping the entire grid and starting over. Intuition may well be a better guide, especially as I train it up by solving these things. It would certainly be faster. I remember struggling with anagrams in crosswords as a child, seeing no good way to solve them other than by writing down every possible arrangement of letters to the point where they stopped making sense. Nowadays, all I usually have to do is look at the letters and, if I have enough constraint from the cross-letters and clue, the solution just pops out at me unbidden. Could a human mind get to the point of solving nonograms the same way? I suspect so, given the prodigious mental feats that “Human Calculators” have trained themselves to do. Heck, after all this intensive practice, I may well be closer to that point than I’ve been allowing myself to acknowledge.

I decided to ease myself back into the Logic section by tackling the one-off puzzles. This time around, it turns out none of them are logic puzzles in the traditional sense. There’s an easy tile-based dissection puzzle, a memory challenge, a sort of visual trivia thing where it gives you the top and bottom thirds of a bunch of circular logos and other symbols and you have to match them up. This last one has a non-intuitive UI: all the pieces are displayed on the screen at once, but whenever you want to match a pair, you have to click “Done”. I don’t think I’d have guessed that if I hadn’t seen it do something similar in the Cryptograms.

There’s one puzzle in the assortment that I could see calling a logic puzzle, although it’s really more of a math puzzle. It’s a more complicated variant on the old chestnut where you’re given a head count and a leg count and have to derive how many cows and how many chickens there are. The premise here is that there’s a planet occupied by six species of aliens with lame jokes for names. The number of eyes, noses, mouths, eyebrows, ears, and heads varies by species. Given a total count of each body part, can you figure out how many of each alien there are? Well, maybe you could in the magazine. Here in Games Interactive 2 it’s impossible because they left out the body part count. That’s like asking us to solve a crossword without the clues. We’ve seen other puzzles where there was information missing, and I’ve always done the best I could. But this time, the best I can do is just hit the “Done” button and let it give me a rating of -100% for not having any of the answers right.

Well, I’ve completed the Special Crosswords. The remaining Orneries fell in a single session. It’s a lot quicker to get these things done when you’re fully alert, and can get your mind into a thinking-of-words zone. But then, it’s also quicker to get them done when you already did half the puzzle in a previous session, and that’s probably the larger factor here.

That left just the sole remaining Clueless, which was one of the difficult ones I mentioned before, where all the words come in just two lengths and there isn’t much obvious constraint about what goes where. Such a puzzle offers no clean way to get started. The only workable approach I’ve found is trial-and-error, picking a space with lots of crossings and trying out each of the words from the word list in that spot and exploring the consequences until you hit an impossibility. If you’ve picked a good spot — which I didn’t on my first couple of attempts — then you eventually find a word that forces a lot of other words, and the pattern grows until not being right would be a weird coincidence. If you haven’t, you eventually find a few different words that don’t force anything much.

This last and hardest Clueless turned out to be the only one in Games Interactive 2 with a typo in the word list: the word “power” was listed twice and “loser” omitted, as if the extra power had forced the loser out. Fortunately, due to changes in the UI, I was able to get a perfect score anyway. In the first Games Interactive, if you tried to put a word in a place where it conflicted with existing letters, the game just didn’t let you. In Games Interactive 2, it lets you and it just overwrites any conflicting letters. And the way the word “loser” was situated, it had words crossing it at the L and S. So I could put “power” into its spot, then re-enter the crossing words to overwrite the P and W. I can’t help but think that there’s a metaphor in that, but I don’t know what.

Next up: the Logic section. As before, we’ve got Battleships, Paint By Numbers, Cross Math (not to be confused with Cross Numbers), and a bunch of one-offs that may or may not really be logic puzzles. I’m pleased to see that, although the Battleships and Paint By Numbers puzzles are still grouped into unwieldy sets, the navigation between puzzles within a set is more sensible now. As you may remember, the behavior in the first Games Interactive was: Pressing “Done” scores the current puzzle and goes to the next one, pressing “Next” goes to the next puzzle without increasing your score, and either way, there’s no going back. In Games Interactive 2, it works the way I originally expected: You can navigate freely through the entire set with “Next” and “Back”, and pressing the “Done” button indicates that you’re done with the entire set, at which point it scores all the puzzles at once. But the irony of suddenly switching to what I originally expected is that I didn’t expect it any more. After the first game, and the cryptogram section in this game, the obvious thing to do on completing the first puzzle in the first Battleships set was to press “Done”.

I do appreciate the freedom to switch around between the puzzles in a set. I’ve pooh-poohed the decision to keep the sets together as units instead of dividing them up into individual puzzles, but as long as you have the entire set available simultaneously rather than sequentially, it’s a better imitation of how you’d solve them in the magazine, skipping around whenever you run out of leads. Unfortunately, it seems to be unsafe to do! Every once in a while, on paging backward in Battleships, I’m horrified to find that the entire set of puzzles, including ones I had completely solved, has gone completely blank. Understand that most Battleships puzzles don’t even start out blank; they usually have a tile or two already revealed and playing a crucial role in figuring out the rest. I don’t know what causes this, but since it seems to only happen when I page backward, I guess I’m going to stop doing that. So long, lovely new feature that I wanted to use.

Previously, I lamented the lack of cryptograms in Games Interactive. Well, in the sequel, my wish is granted! Cryptograms are an entire category — although, weirdly, the name of the category is “Crypto Funnies”. I guess this has something to do with the fact that it’s the name of the first five puzzles in the list. As with the logic puzzles, the category contains four distinct sub-types: Crypto Funnies (four-panel cartoons with ciphered word balloons, which gives you enough context for the deciphering to be really easy), Cryptolists (ciphered lists of things that fit some theme, without the cues you’d get from full sentences), Variety Cryptograms (collections of ciphered texts fitting some theme), and “Dszquphsbnt!”.

This last one is the name of Games Magazine’s regular cryptogram section; it’s the word “Cryptograms” shifted forward one place in the alphabet. The individual cryptograms within a Dszquphsbnt! are unrelated, and, as in Battleships and Paint by Numbers, are only grouped together here because they were originally published that way. Dszquphsbnt! is where the really tough cryptograms are — the ones where they make sentences without articles or other short words and with weirdly skewed letter frequencies, where your only way to get started is by noticing a long word (or, worse, combination of words) with an unusual pattern that identifies it. They don’t start out that way, though. Each Dszquphsbnt! set starts out easy and works its way up. In fact, the first cryptogram in each Dszquphsbnt! is a “Cryptoon”, which is basically the same idea as Crypto Funnies but with one panel and a caption instead of four panels and balloons. Unfortunately, in Dszquphsbnt!, this game leaves out the pictures. The Cryptoons are quite solvable without them, but what you end up with is a punch line without its context, and sometimes it’s a really inscrutable punch line, like “That looks like it says, ‘Machine wash warm, tumble dry medium, made in France'”. I think some of the Variety Cryptograms may have been originally published with pictures too, but that’s just a guess.

The basic cryptogram UI here, shared by all the puzzles in the category, isn’t the best I’ve ever seen, but it’s okay. It lets you select letters with either keyboard or mouse, and in the case of mouse, it lets you click on either a displayed alphabet or directly on the cryptogram. I mostly wound using it in a sort of hybrid style, clicking on the cryptogram to select a ciphertext letter and then typing the plaintext version via keyboard. Selecting any letter highlights all instances of it in the cryptogram, which is handy for eyeballing letter frequencies.

On the downside, it occasionally fails to respond to the keyboard, making me press a key multiple times to get it to register. Also, it lets you bind multiple ciphertext letters to the same plaintext, so sometimes I accidentally wound up with multiple distinct kinds of T on the screen. This is exacerbated by the way it removes the ciphertext letters from view when you bind them to plain text, so you have nothing visible to tell apart identically-bound letters. That is, it doesn’t take them away completely — there’s still a very faint ghost of the letter there, like they tried to gray it out but went too far. At least you can still click on it to highlight it, but what it really makes me want is a way to remove a letter’s binding, and the game doesn’t give us that. The ability to mess up the display without being able to unmess it interferes with the way I want to use the interface: not just as a way of entering answers that I’m sure of, but as a medium for exploring possibilities.

I said the cryptogram UI isn’t the best I’ve ever seen. You know what is? It’s the one where you have just one version of the text displayed, and selecting two letters swaps them in it — that is, selecting A and J, for example, replaces every A with a J and every J with an A. It’s simple, and it just naturally avoids the problems here. And if I’m not mistaken, this was the interface used by Cliff Johnson in games such as The Fool’s Errand and At the Carnival back in the 80s, so it’s not like it was unknown.

Outside and around the cryptogram UI, there’s the UI for navigating through the puzzles within a collection, and that’s where we run into real trouble. Dszquphsbnt! and Variety Cryptograms commit the same sin I previously observed in the Battleships and Paint by Numbers in the previous game: they expect you to solve each puzzle in the group in sequence, and press the “Done” button after each one to score it, but the “Next” button, which advances to the next puzzle without scoring the current one, is still available, even though there’s no way to go back once you’ve pressed it. Cryptolists spreads a single puzzle over multiple pages, one list element per page, and thus allows you to page back and forth freely with the “Next” and “Back” buttons, but it still expects you to press “Done” on every single page to get credit for it. Hitting “Done” on every page isn’t enough to finish the puzzle, though. You signal that you’re finished with a Cryptolist by pressing “Next” on the very last page, which is way too easy to do accidentally, because you’re pressing that button a lot just to see the entire puzzle.

The very worst thing, though, is the Crypto Funnies. Like Cryptolists, Crypto Funnies spreads a single puzzle over multiple pages, one page for each panel of the comic. And like Cryptolists, it lets you page around with “Next” and “Back”, and expects you to hit “Done” on each page. But this time, there doesn’t seem to be any way to signal completion. Pressing “Next” on the last page just keeps you at the last page. The only way to get out is to just quit the puzzle, which leaves no record that you ever attempted it.

Remember that there’s a final puzzle only available to people who have played all the puzzles. It looks like this is impossible to reach without cheating. So I cheated. I think it’s permissible in this case: I’m not lying to the game about my accomplishments, I’m just making it acknowledge the truth by the only means available. Luckily, the game’s record of player progress turns out to be stored as an easily-editable text file. The only complication was that I failed to realize at first that the first thing in that file is a count of the records it contains; if I didn’t increase that, anything I appended to the end would be ignored.

[Update] It turns out there is a non-cheating way to get credit for attempting the Crypto Funnies: pressing Next on the final page works if and only if you have not made any attempt at solving the puzzle. If any letters are bound, it fails. I guess this explains how the unlocking of the final puzzle passed the developer’s tests, if they performed any.

Like the first game, Games Interactive 2 divides its puzzles into categories. But this time it gives the categories in alphabetical order, so Crosswords come first.

The UI has changed in several ways. The clues are in a smaller font, so it can fit more of them on the screen at once, but it’s wasting even more screen space on headers and frames, so it’s basically a wash. The backspace key, which used to delete the previous letter in the word unless the cursor is on the last letter and the last letter is filled in, now behaves consistently: no matter where you are in the word, it deletes the letter under the cursor and then backs up if possible. This takes a while to get used to, because it’s not how backspace behaves in any other context I can think of. Clicking the grid to navigate, which always selected Across in preference to Down before, now alternates — if you had an Across clue selected, it selects a Down, and vice versa. (Clicking a square on your currently-selected word to switch to the word that crosses it is thus now just an application of the general rule, rather than a special-case behavior.) I thought this was weird at first, but I’ve come to appreciate how it fits a certain solving pattern, where you fill in a word and it gives you enough information to fill in a word that crosses it. But it doesn’t exactly fit the just as frequent pattern where you want to fill in several words that cross it. I still can’t help but feel that there’s a better way to handle this.

Anyway, it’s not all improvements. For one thing, it’s slower. I remember when I tried this game for the first time being disappointed that they hadn’t fixed the speed problems of the crosswords in the first Games Interactive, but now that I’m playing on a faster machine, I see they actually made it worse. When you select a word, it highlights it in the grid, one square at a time, a flashy little transition effect. For the longer words, the resulting delay, during which the UI is unresponsive, can be quite irksome. Another thing: Whenever you select a word, it moves the cursor to the beginning of that word, even if you clicked on the middle. This is something that the first game got right and GI2 gets wrong. And where the first game had some problems with navigating the grid with the arrow keys, this game solves them by scrapping that functionality entirely. You can only use the arrow keys to move the cursor within the current word.

I haven’t seen a whole lot of bugs within the grid data. There were a couple of cases where its notion of where a word ended was a few letters short of where it should have been, but that just meant I had to fill in the rest of the letters via cross clues. Instead, the chief problem this time is typos in the clues. Most commonly, there are a bunch of clues missing their first letters. This is fairly benign; when you see a clue like “ouis Quinze, e. g.”, you can tell what it’s supposed to be. But there are other places where typos just obscure the meaning. For example, “Cop cabana site” was a plausible enough construction that it wasn’t until I worked backward from the cross-letters that I realized it was missing an A.