Advice on learning to construct crosswords

[BarbaraK]

So I'm kind of sort of maybe thinking about trying to learn to make a crossword. I'm not highly committed to this; it's entirely possible I'll play with it for a bit and then lose interest. So I don't want to invest a lot in it at this time. But I'd appreciate any advice from our muggle constructors on free or cheap resources and suggestions about how to get started. (Or maybe advise me that it's not worth it if I'm not ready to go all in.)

I found this New York Times series which gives the steps and some general info.

And remembered Brian writing about his journey and found this post

I banged by head against a wall on and off for nearly 10 years trying to construct a crossword puzzle. Every few months I would give it a try over the course of a weekend and then quit in frustration, vowing to never try again. Then someone tipped me off that all but the most hard-core pros use commercially available software. There are two main programs -- the most popular are Crossword Compiler (for PC) and CrossFire (for Mac). Both are about $50.

Once I decided to get serious, I tried both and found CrossFire much more user-friendly. I also joined Cruciverb ($40/year), which has an excellent and easily searchable database for looking up past puzzles and theme entries. It also has a helpful page on constructing tips (you do not need to be member to access the tips). Finally, I ponied up about $200 for Jeff Chen's word list. The Patrick Berry Book is also invaluable.

All of this was a moderate investment, and it took a full year of effort and 15 rejections, but the end result was finally worth it! It was a great experience start to finish and I encourage anyone who is interested to go for it! I am far from an expert, but I would be happy to help with any tips or I look at draft puzzles or theme ideas.

But I believe https://crosshare.org/ is new since then, and I'm wondering, is that a reasonable place for a beginner to start? Or do you need to already know what you're doing there. I see that CrossFire has a free demo version - you can play with it but not save anything - would that be better? Or do I really need to buy the program?

Is it easier to start with smaller grids and work up to the bigger ones? Or is 15x15 the optimum size?

Any general advice? What do you wish you'd known before you started constructing? Or what was the best advice someone gave you?

[MikeyG]

I am the biggest neophyte ever for this, but I will state that I love Crossword Compiler. It does require a bit of an investment, and for some reason, Crossfire and I never really got along - I found it more confusing to handle. But, honestly, that could have been luck: I used it first and then used Compiler second, and so maybe I was at a better place then.

This is critical; this is an ever-growing wordlist that has really helped me. I actually love, I must admit, crosswordese: I don't mind seeing ANOA and ETUI and STOA in my puzzles at all, but I know that the fill should be more sterling than that if we're attempting to market them elsewhere. With metas, which I love, you can play around a bit more with the fill since if there's a few entries that are "meh," that doesn't matter as much.

I probably should also be reaching out more to polished constructors, but I'm a huge introvert and also feel like I'd be annoying them, haha - but I'm sure there are a lot out there who would be very happy to give you pointers or take you under your wing!

One thing that Will of Pandora's Blocks puts in his e-mails is: "Nobody should ever do anything crossword related if it’s not fun for them." That's probably my biggest piece of advice. I started making more metas to the point that it's really all I make now because, well, I have the most fun making them! Apparently, themeless crosswords are a huge thing, but they're not really something that interests me too much in constructing (solving is fine), so I'll shy away from making those.

Always be on the lookout for ideas as well. Sometimes, I'll think of a phrase or something and I'll be like, "Oooh! That could be a theme revealer!" Legend has it Merv Griffin did that a lot for Wheel of Fortune. On the Griffin angle, the Jeopardy! writers are genius, and I bet sometimes watching that could be a springboard to ideas and such.

Other polished people can give more targeted advice, but those are some early rambling thoughts. Welcome to the wonderful world of constructing!!

Mikey G

[vandono]

I typed a painfully long comment but fortunately for you I somehow refreshed the screen and lost it. So this should be much more brief.

I spent decades being very fascinated with how puzzle constructors weave words together so tightly, but despite occasional efforts I've never had success at it.

However, this forum helped me finally do this. While working on someone's weekly meta on Crosshare, I saw that you can create your own puzzles. This is a free way that is a great place to start if you are hesitant to buy one of the more feature-rich packages for now.

You just select the size of your grid and get cracking. As you build it, it will suggest words that fit the row and column for the selected cell. (These will be listed to the right of the grid.) When you mark a cell as blank, by default it will automatically blank the corresponding cell on the other side to keep it symmetrical. (It doesn't warn you so if you had already put a letter there it will silently blank it out.)

As you get far enough along, it may automatically try to complete your grid. (It does this right off the bat on a 5x5 grid.) This caught me by surprise as I was nearing completion of my first 15x15. It will use words from its built-in list and some of those may be too obscure for your taste; some you may have trouble even finding a definition for. This can be disabled.

When you're ready to create the clues, you can click a button for a suggestion. Even when I have something in mind, I may scroll through these to see if I like any better than mine.

When you are done, you can publish it. This can be kept private from the Crosshare community but shared via a link, or you can share it with the community. It can also be printed from the app and downloaded as a .puz file.

Another very useful tool is OneLook Dictionary Search. This is a tool for searching multiple dictionaries (of varying quality) and other references. The beauty of it for a puzzle constructor is that you can do a search such as 'c??ss???d' to find words matching a pattern you're having trouble completing in your grid. With the ':' operator, you can also search for words related to a given word (like, say, ':puzzle'). This searches a Thesaurus and can be a way to come up with words for a clue to your word or to come up with words for a theme. The home page describes the search options.

Hmm. this still ended up rather long. Sorry.

Whichever tools you use, I hope you have fun!

[Cap'n Rick]

I'll echo what @vandono said above re: Crosshare. Several years ago I used a product called Crossword Weaver (I think it's still available). Last year when I asked @AnswerPfinder for his recommendation, he suggested Crosshare. I gave it a look, and I've been using it ever since. It's ad-free and free to use (there is a link for donations), and I find it to be very well-designed and user-friendly. In addition to the built-in wordlist, I like its instant-feedback capability for meta submissions. There is a surprisingly large community of Crosshare constructors of many different puzzle types who are very supportive of each other. I'd give it a try, @BarbaraK.

[BrianMac]

Agree on crosshare. It's a new entrant since my original post above, and for the money, there's really no reason not to try your hand at it using that before investing in any software.

[BarbaraK]

Thank you for all the replies - here and the PMs! So far, no one has said, "don't do it!" Not really a surprise with this supportive crowd.

I've played around a bit on crosshare, and I did manage to make a 5x5 just for my husband. But it's amazing how, once you put a few things in, the choices for the rest rapidly dwindle. I guess getting a feel for what will be workable and when you're heading for a dead end will come with practice.

I'm also trying to figure out, as I try the larger grids, how do I decide where the black squares should go.

[TMart]

I've constructed a few puzzles over the past couple of years (with a few in the hopper - stay tuned!), and I've found that Crossword Compiler is pretty easy to use. I haven't tried Crosshare, but it seems to have become pretty popular with the MMM crowd. I should probably try it.

You should DEFINITELY do it - solving a regular crossword is fun, solving a meta is more fun, but construction is a whole new level of fun as far as puzzle solving, because there is no actual answer until the grid is finally filled, the clues are written, and the title and prompt are carefully worded. Sometimes, you just can't get it to work, but when it does, it's even better than solving a week 5, especially as the solves start rolling in.

The first puzzle is the meta idea. The second puzzle is the grid size and layout - I like to do 15x15s, but sometime that changes on the fly if the meta dictates it. The third puzzle is the grid fill, which I'm pretty bad at, and that makes it even more of a challenge. The final puzzle is the clue-writing, which I find somewhat tedious at times, but when its done really well (i.e. Al Sisti, Dr.Tom) it adds a whole new element to the solve experience.

This forum is the best thing that ever happened for amateur constructors. Just having a place to get your random creations out makes it well worth the effort of constructing. There are way too many new puzzles every week to keep up with, but every one seems to have something unique about it.

Looking forward to your creations!

[Cap'n Rick]

I'm also trying to figure out, as I try the larger grids, how do I decide where the black squares should go.

Choosing grid patterns is definitely a challenge for me. I usually start with a standard 15x15 pattern (see https://crosswordgrids.com for samples) that best accommodates my theme/meta entries and then tweak it until everything fits. Sometimes, when I'm lucky, it's a relatively straightforward process, but oftentimes it becomes a bit of a wrestling match.

Crosshare has a nice feature that let's you "analyze" your grid's word lengths and word counts. 78 total entries appears to be the magic number that we try not to exceed for 15x15 grids. Crosshare also provides several options for grid symmetry (including no symmetry).

I'm hoping that some of the more-experienced constructors will also chime in here and provide some of their tips and tricks.

[benchen71]

I love Crosshare!

Grids can be tricky to make from scratch, so I use https://ugleh.com/gridsearch to search for a grid with the right bits for the meta I'm constructing. I then recreate the grid on Crosshare, modifying it as needed. I put my meta mechanic in place and then hope that the automated fill comes up with the rest of the grid. I never just accept what appears, however. It's just good to see that a completed grid is possible! Then comes the hack work of selecting the best fill. Sometimes I end up in a dead end, and have to backtrack a little (making sure not to delete the meta mechanic in the process!)

If the automated fill is struggling to come up with something, use the "!" key to take you to the entry that is causing issues. Find something you are able to clue and keep doing that until the automated fill kicks in.

[Cap'n Rick]

Thanks for the info, @benchen71! Very helpful.

[AnswerPfinder]

Woo! Hooray for more constructors!!

Here are my two cents, sorry for the long post - hopefully one or two things in here might be useful!

I use a combination of Crosshare and Crossfire. In my personal opinion, Crosshare is far superior for nearly every aspect of creating and filling grids (plus it's free!), but I like the flexibility that Crossfire provides in terms of saving, editing, and exporting puzzles to different formats. It's worth mentioning on this page that the creation and maintenance of Crosshare was/is no small feat. @miked has done a really, really good job with it and deserves a ton of credit!

In terms of finding valid black square placements, I started out using the pre-generated designs from http://www.keiranking.com/apps/phil/. Now, I mostly do the placements by hand, but the ugleh site that @benchen71 mentioned looks awesome - I will definitely take a look at that the next time I get stumped!

If you're trying to come up with theme entries, I agree with @vandono - there is nothing better than OneLook. From time to time I will also use the Regex Search extension through Google Chrome, which serves as a soup-ed up version of CTRL + F (https://chrome.google.com/webstore/deta ... n?hl=en-US). I think it is more powerful in its search capabilities than OneLook, but you can only search one page at a time, which is why I don't use it nearly as often as OneLook (which will search all of the pages in the deep depths of the internet).

When I started making puzzles I started with 15x15s and I found it pretty manageable. I actually found it easier to make 15x15s than midis (say a 7x7, for instance) because the vast majority of a 15x15 puzzle will be broken down into smaller subsections/corners which often end up smaller than 7x7.

I have never felt the need to pay for any wordlists. There are enough free options out there that I have always felt pretty happy using whatever other people came up with. Tailoring these pre-exiting lists to your preferences by removing words you don't like and adding in some phrases that are missing is an excellent thing to do, and it is totally worth the time investment if you see yourself making lots of puzzles. I think Mike D. is working on enabling this feature for Crosshare in the near future, too, which is very exciting!

My only other advice would be to make puzzles that YOU like. We all lose a lot of creativity, ingenuity, and joy if we only try to fit the prescriptions created by the big publishers. I've been trying to make puzzles that I think the NYT would like for several years, and those puzzles never bring me nearly as much joy as the ones I make for Pandora's Blocks. And, as @TMart said, this forum is really wonderful at making sure all of our wacky brainchildren find loving homes in the minds of appreciative and enthusiastic solvers.

Can't wait to see what you come up with!

[madhatter5]

My only other advice would be to make puzzles that YOU like. We all lose a lot of creativity, ingenuity, and joy if we only try to fit the prescriptions created by the big publishers.

This would definitely be my main advice as well. Also two-letter words are very important and have every right to be placed in a crossword, thank you very much.

[Abide]

Also two-letter words are very important and have every right to be placed in a crossword, thank you very much.

And some of us have every right to say $@%&! when we see them

[madhatter5]

You know what guys, just try expressing one single sentence using only words comprising three (optionally more) letters. It's ... okay well actually it's less difficult than was presumed before typing out this rebuttal ... never mind then haha

[MajordomoTom]

Once I decided to get serious, I tried both and found CrossFire much more user-friendly. I also joined Cruciverb ($40/year), which has an excellent and easily searchable database for looking up past puzzles and theme entries. It also has a helpful page on constructing tips (you do not need to be member to access the tips). Finally, I ponied up about $200 for Jeff Chen's word list. The Patrick Berry Book is also invaluable.

would this be the 'for Dummies' book? If so, I just ordered a copy from Amazon.

[MarkWoychick]

Thanks @BarbaraK for posting this question - I’ve wondered about this myself - and thanks to everyone who has replied - the advice is really helpful. As a result, I’ve created a puzzle on Crosshare, and I’m hoping to recruit some test solvers. If you are willing, please let me know and I’ll send the link (or PDF, .puz file - let me know your preference). Thanks again to all - this really is a terrific community!

[BrianMac]

Once I decided to get serious, I tried both and found CrossFire much more user-friendly. I also joined Cruciverb ($40/year), which has an excellent and easily searchable database for looking up past puzzles and theme entries. It also has a helpful page on constructing tips (you do not need to be member to access the tips). Finally, I ponied up about $200 for Jeff Chen's word list. The Patrick Berry Book is also invaluable.

would this be the 'for Dummies' book? If so, I just ordered a copy from Amazon.

Yes, I think so: https://www.amazon.com/Crossword-Puzzle ... 0764556223

It's officially out of print, but you can still get it on Amazon as above and directly from Patrick on his website: https://aframegames.com/store/?download=21

[BarbaraK]

Thanks @BarbaraK for posting this question - I’ve wondered about this myself - and thanks to everyone who has replied - the advice is really helpful. As a result, I’ve created a puzzle on Crosshare, and I’m hoping to recruit some test solvers. If you are willing, please let me know and I’ll send the link (or PDF, .puz file - let me know your preference). Thanks again to all - this really is a terrific community!

I don't think I'm a good test solver, because I don't really know what kind of feedback to give. But I'd love to solve your puzzle if you wouldn't expect much useful in response. I'd prefer the .puz version.

In my puzzle creation adventures, I was playing with filling the grid, and amazed at how much putting in one entry constrains so much around it. Then I got distracted with other stuff and haven't been back to it in a few weeks. I intend to though, just no idea when.

[MarkWoychick]

I don't think I'm a good test solver, because I don't really know what kind of feedback to give. But I'd love to solve your puzzle if you wouldn't expect much useful in response. I'd prefer the .puz version.

In my puzzle creation adventures, I was playing with filling the grid, and amazed at how much putting in one entry constrains so much around it. Then I got distracted with other stuff and haven't been back to it in a few weeks. I intend to though, just no idea when.

Thanks, BarbaraK - I will send it to you. As for creating the grid, I ran into the same issue with word constraints at first. Using the site that benchen mentions above helped a lot in making reducing constraints and allowing Crosshare to suggest more entries.

[benchen71]

I don't think I'm a good test solver, because I don't really know what kind of feedback to give. But I'd love to solve your puzzle if you wouldn't expect much useful in response. I'd prefer the .puz version.

In my puzzle creation adventures, I was playing with filling the grid, and amazed at how much putting in one entry constrains so much around it. Then I got distracted with other stuff and haven't been back to it in a few weeks. I intend to though, just no idea when.

Thanks, BarbaraK - I will send it to you. As for creating the grid, I ran into the same issue with word constraints at first. Using the site that benchen mentions above helped a lot in making reducing constraints and allowing Crosshare to suggest more entries.

I missed your request for test solvers. I would be interested, too!