Saturday, April 23, 2022

Starter Words

Oh right.  I'm supposed to be providing a service here.  I played the new Wordle-derived games intensively the first few weeks, now more intermittently.  When they came out, I immediately attacked the strategy questions, which interested me much more than actually playing the games themselves.  The games are passively interesting now, but...don't really need 'em after the first couple of weeks. Mild entertainment.  But working out what would be the optimal starter strategies?  That intrigued me for those two weeks.


1. I start Wordle with ORATE.  I get the point of starting with ADIEU, but I find it more comfortable to use that or IRATE, or with two vowels CRATE, TALES, RATES, TILES whatever with that one more consonant and then for the second guess seek to construct a word out of whatever was a hit and some combo of the common letters NCILSU. I get more 3's than 4s that way, but only by a little. By guess 6, the problem is often trying to find any word that fits the pattern of what you have got. It's nice to get the vowels nailed down, but consonants provide so much more information.

1A. Sometimes you get one of those patterns where there are more pure-guess possibilities than you have left, like S_INE where it could be SPINE, SHINE, SWINE with only two left. "Waste" a guess on WHIPS, which will identify the missing letter. Once there were seven possibilities and I had to waste two guesses using three each of the possible letters to get it right on guess #6.  That's very rare, but it happens.

1B. If an S, C, G, or T is a hit, then H is much more likely.  That is true for W as well, but you are going to consider H before W anyway.

2. I start Dordle with ALIEN STORY. I very seldom lose. I think it is the toughest game of the four, but if it had one more guess it would be the easiest. My wife starts with ADIEU STORY and also seldom loses. I am okay with trading N and L for U and D because if no vowels are a hit with my method you know it has to be a U. But it is nice to see the colored square.  I get that.

3. For both Quordle and Octordle I start with LATER, COINS, PUDGY.* I never lose. It's just an exercise at that point, not a real problem. I don't even play those most days anymore.

4. WORLDLE continues to both intrigue and irritate. When I do not recognise the shape or its probable location (the NE-SW slant of the -stans, the EW slant of Europe, the verticality of SE Asia, sensing whether something is a coastline versus a river boundary - you can sometimes tell) I guess near the equator or just below because so many of the answers are going to be there, and because the shape of the continents, when your first guess is wildly off, informs what the arrow and mileage should narrow it down quickly.  Also, just fold your cards and go to the map when you are in the smaller Caribbean islands, the South Pacific, or those ridiculous tiny islands you never even heard of in the Indian Ocean or the South Atlantic. Even the people who live there might not recognise those outlines. Sometimes you even need to search a very specific ocean map because a normal map doesn't provide enough detail.** However, there is a smoothness to the equatorial islands and a jaggedness to the colder ones that sometimes gives a clue.

* Be alert to the priors in the YouTubes that purport to tell you what the most common letters and best strategies are. You don't need to know the most common letters overall, but the most common letters in five-letter words in which the game does not include both four letter verbs that have an S at the end or four-letter nouns that have an S at the end, because the game is artificially designed that way.  That pushes S from #6 down to #10, or something like that.

** You discover in a week or two that Africa and the South Pacific are much, much larger than you ever knew.  All that discussion about the Eurocentrism of mapping which still persists into the modern era is true. The Europeans did the first real world maps and centered them well above the equator. Our perspective has been skewed ever since. Greenland is not that big.


Douglas2 said...

I find myself too disturbed by everything under the ** footnote. "Surprised at the size and distances of Africa and the South Pacific" is not my experience. But I was still a child when something twigged -- either noticing distances that I measured with a ruler on a Mercator map didn't scale to the distance tables between world cities, or wondering why the air routes were arcs not straight lines. It then took me all of a few minutes to work out that the flat rectangular world map did not distort cardinal direction (North being straight up across the whole page) but did tremendously distort distance, whereas the segmented world maps like the Goode homolosine projection (that seemed to be in every classroom I had in primary school) preserved area but made it really hard to work out whether Stockholm was further north than Anchorage. At that point I hadn't even heard of Tissot's indicatrix, but worked out that using a globe took away the risk of being fooled by the inevitable map distortions. So I buy globes for all my little nephews and nieces.

The phrase "Eurocentrism of mapping" bugged me, and I see it is a term of art in critiquing how non-cartographers use maps, particularly in teaching and visual display of quantitative information, rather than in how maps are created -- but it is also used by people who are disciples of the surprisingly successful fraud and map crank Arno Peters. In college I started praying for world evangelism using the Ralph D. Winters book and wanted a world wall-map as a visual aid; it bugged me that I had such difficulty finding one that didn't split Asia in half because of being centered on the USA. But I don't see any corresponding effort by map printers to prioritize Europe, if it is 'in the middle' or 'too large' that is a side effect of very logical choices made for good reasons to make a map useful for a particular purpose by minimizing or eliminating one or more of the inevitable 4 areas of distortion at the expense of increasing another.

"Greenland is not that big" appears to me in this context as a non-sequitur -- it is big precisely because of the use of cylindrical projection, which is accurate at the equator, not because the map is centered north of the equator.

Sorry for the tedious essay in response to a footnote. It bugged me. I think I did wordle twice, to verify that my theories were correct. Of course the first time I accessed the site my curiosity compelled me to right click and 'view source' so I contaminated myself by seeing the entire word-list in date-order. I was happy to check back a fortnight later and see that they had fixed the HTML to prevent me from doing that.

Assistant Village Idiot said...

When I look at the offered images, many of them seem centered on Europe, with corresponding swollenness of Greenland. Not all.

Douglas2 said...

And on USA maps, if you find the mid point vertically and horizontally it is usually Kansas:

Donna B. said...

I play Wordle sporadically. I changed my start word and then discovered that if I'd kept to the original, I'd have got the word on the 2nd try. Oh well, it's a nice diversion sometimes and doesn't generally tax my brain too much as it seems the other versions you write about would do.

Grim said...

The Mercator projection actually had a very important navigation feature. It wasn’t drawn that way to make European places seem bigger, but so you could use it like so:

“ This map, with its Mercator projection, was designed to help sailors navigate around the globe. They could use latitude and longitude lines to plot a straight route. Mercator’s projection laid out the globe as a flattened version of a cylinder. All the latitude and longitude lines intersected at 90-degree angles. Because the projection was intended to be a reference for navigation and not land geography, the landmasses on the map are not necessarily proportional to their actual size; at higher latitudes, landmasses appear larger than their actual size.”

My sister asked me to join her and mom in Wordle. I open with LASER then TONIC. This last week I have had to make a fourth guess only once. Otherwise it’s been 3s. The game is mostly strategic, as you say; I think of those opening two shots as akin to bracketing fire with mortars. The information gained from the two brackets usually lets me hit the target with the third round.

Thos. said...

With a good set of three starting words (there are lots of sets that will work), I can pretty much guarantee that I'll get Quordle - with about 75% of those wins happening in 7 guesses.

But I have never yet been able to get it in fewer than 7 guesses. A good two-word starting combo might help me guess one of the words on the third guess, but it's never been good enough to help me get all four by the sixth guess.