On a vacation trip, this was an extra, “Even More of a Challenge” in the Delta Sky magazine of June 2008.
Line marking was tough, and no UR showed up.
Nothing on the bv map, but a 345-wing . . .
removes 4 r2c2 for the above naked triple in the NW box.
Quite certain that no XY ANL were available, I allowed Beeby UR, regular fish, and and simple ALS_XZ before AIC building. The result was this ALS_41, before a string of AIC, mostly 1-ways. The first one starts well from 1r4c4, with possible ANL and 1-way victim.
Then another 1-way, from rr3c9, attacks the 4’s.
And a third 1-way is followed by an AIC ANL.
As I add a healthy cluster, the first XY ANL appears. Now, instead of 11 link between the terminals, I can cut out 8 in a short cut through the cluster. A convincing demo for early coloring.
Next is a remarkable pair of 1-ways that share the same AIC. Start at 4r1c4. If it is true, 4r9c4 is false. If it is false, then follow the AIC to show 4r1c4 to be false. So 4r1c4 is false, regardless of 4r1d4. Now start at 7 r9c4. If it is true, 7r1c4 is false. If it is false, then follow the AIC back to show 7r1c4 to be false. So 7r1c4 is false, regardless of 7r7c4.
The 1-way removals leave a naked quad in c4.
This gets me a rare demonstration of a grouped cluster expansion. A 1-group clinks with blue 1r7c6 and blue r1c6 to expand the blue/green cluster in r1.
A simple AIC ANL wraps blue.
The wrap leaves a stunning naked quint, but no collapse.
An added bv allows the ALS_ 92. The new 261-wing at r7c8 has no victim.
We finally get XY ANL from this growing bv field.
Adding the overdue cluster, it not only duplicates a swordfish removal with a trap, but row clinks force two orange 1’s in the N box, wrapping orange.
Here’s the colorful solution.
This post replaces an obsolete post on this date, so as to leave the Sysudoku basic blog ending undisturbed.
Systematic Sudoku 
Hi Mr. Welch,
I just wanted to thank you for your contribution to the Sudoku world. I discovered your “Order of Battle” yesterday and have found it very helpful in tackling the more challenging puzzles.
I’m looking forward to reading and learning more from your blog.
With great appreciation,
Geoff Corbett
Thanks, Geoff. You can help by commenting directly on any post where I could be more clear or more correct, especially in the earlier posts that introduce concepts. I do minor editing on past posts.
Thanks for being so diligent, Richard. The comment “238 has 23 and 28” was in filtering for possible XYZ wings, before examining links. Your deeper questions inspired me to review definitions, and I discovered that I had overspecified the XYZ-wing, needlessly excluding some. The Z’s don’t have to see each other. I have revised this post, and the next. I think your questions are answered by these revisions.