New constructor's question

[Toast]

Hi, all. I'm just getting into puzzle constructing and I'm working through my first 15x15 grid, using Crossfire. My experience so far has been similar to what is described in the "Sage Advice" articles: I put in what I think are interesting answers and then get stumped putting in respectable fill around them. After flailing around a bit, I rip out some of the answers I wanted to keep and start fresh, with interesting words and phrases I've encountered along the way.

What I have a question about is a point that Nancy Salomon made in "Some thoughts on word lists" -- she wrote:

"Coming up with a good fill [...] involves identifying the most constrained sector(s) of the fill and starting there to make sure one has acceptable alternatives."

This is surely sound advice, but I'm having a difficult time picturing what this means in a practical sense. How do I identify constrained sectors in order to start there? Are there clues that you look for in the grid design or is it letter combinations that might originate from some of the longer or themed entries?

[mmcbs]

Sorry, I'm not familiar with Crossfire, but I'm told it's pretty feature rich, so perhaps that is built in. In Crossword Compiler, the system immediately figures out which individual space has the fewest possibilities and starts there. But it may be a combination of entries in a particular sector that cause the problem. You can eyeball this by considering the number of entries that have multiple letters already forced by theme entries or others - this increases the odds of problems. Another factor is the nature of the letters - that is if there are a lot of high score scrabble letters in a section, it's likely to be tough to fill.

Another consideration is the word list itself, because if it isn't robust enough, many more sections will be tough to fill. There are good word lists on this site for members, and perhaps an even better one on xwordinfo.com for a fee.

And, finally, if you have a sector that you just can't seem to conquer, it may be time to add a "cheater" blank square, re-arrange the theme entries (if possible), or even redesign the grid. We all have to do some of all of those things once in a while.

[Toast]

Thanks for the tips, Mark. I think I'll use part of this snow day to work on the puzzle.

[RichP]

I started using the xwordinfo wordlist in Crossfire with a "min fill word score" of 50, and it made finding a clean, fillable grid much easier.

[kwanjul]

Before you finalize your grid, you need to make sure that every down entry (assuming your themers are all across) has multiple decent options. A constrained sector would be an entry that has very few options for fill, such as E??H or A?F?E. When I'm doing my fill, I like to start with the long non theme entries, while also keeping an eye on the shorter fill that may be tough because I already have 2-3 letters committed to a 4-6 letter entry. For the 3, 4, and 5 letter entries, I take what I can get while minimizing obscurity and crosswordese.

[Glenn9999]

This is surely sound advice, but I'm having a difficult time picturing what this means in a practical sense. How do I identify constrained sectors in order to start there? Are there clues that you look for in the grid design or is it letter combinations that might originate from some of the longer or themed entries?

Once you get your theme entries laid out in your grid, you want to look for parts of the grid where there's odd combinations of letters that might be difficult to find words to match with.  For example, in my (only) grid attempt (so far), I had ??W??F as a possible word with the W and F being parts of my theme entries.  When you do it, you should have some idea from solving puzzles as to how words are constructed and what letters naturally fit together.

More or less, you want to identify something suitable for those areas first and then work the rest of the grid fill from there.