I must first note that corn maze is always a delightful event, and I am immensely grateful to the organizers and the maze designers for any chance to race in the corn.
That said, this year's maze looks somewhat uninteresting. There are two classes of corn maze:
- Maze Type 1: There is essentially only one path between two points, and the challenge is to spot it and avoid dead-ends. Example from SLOC 2011
- Type 2: there are many paths, and the challenge is to determine which is fastest to execute (not necessarily the shortest); numerous examples, standouts include 2011 Mike's, 2013 Mike's, 2014 Mike's.
There are some borderline cases, of course - last year's Mike's Maze
leaned heavily to Type 1; for most controls, there's only 1 or 2 plausible routes, and the main challenge is to avoid traps.
In my view, the pinnacle of corn maze orienteering is a good type 2 maze - lots of options for routes between two points, many chances to get confused, extremely high frequency of map reading. Long periods of required dead running during which you can't get lost are thoroughly uninteresting.
The 2017 maze looks to continue the tradition of 2016 - most legs will almost exclusively be Type 1, with only one reasonable route. Consider any leg going from anywhere in the "T" to anywhere in the "H" has to pass three choke points - the "RT" boundary, the "pale", and "dot". Transitioning between those 3 choke points is dead running with basically no decisions.
Compare this, for example, to the medium length leg 10-11 on the 2013 classic
- there are about four plausible variants of two main options, straight and around to the right. On the most direct route, there are about ten decisions that have to be made at 3-5 second intervals. Consider the first leg on the 2011 Maze
- it's diabolical, with perhaps five major variants of route. So many decisions, so many opportunities for errors of all sizes.
I think there are objective measures of quality - for instance, error rate. On the type 1-ish 2016 Maze
, consider controls 7-11 (sample) on the classic course. On those 5 legs, AP's splits
indicate 21 mistakes among the 44 competitors with splits, i.e. 21 errors in 220 legs for a 9.5% error rate. On legs 7-11 of the 2013 classic, there were 42 errors for 35 competitors with splits, i.e 42 in 175 legs or a 24% error rate. Also, (without analysis), the errors in 2013 tended to be much larger than those on the 2016 course; there was more risk of catastrophe. Type 1 mazes are often runner's courses with comparatively few navigational choices, whereas type 2 mazes are navigationally interesting. Of course, it's possible to design a weaving complex type 1 maze and a trivial runner's type 2 maze; but type 1 mazes are about running between choke points, while type 2 mazes are about evaluating and executing many possible routes.
This year's maze only uses a small fraction of its area (say 25%) for interesting maze. The rest is dead space. One suggestion going forward: increase the complexity by overlaying a pattern behind whatever picture is used. The 2010 and 2011 mazes executed this very well, with a complex network behind the main image. Curved paths work best for this, as they limit visibility. Imagine if behind the "EARTH" logo, we had the block maze from the bottom bit of the 2010 Maze
or the grid of paths from 2011
(even without the letter courtyards). That would be much more interesting.
I am very excited about this event, but I hope in the future, the corn mazes are more maze-like and Type 2.