Wordle 1,048 4/6*
Connections
Puzzle #326
1 Like
Wordle 1,049 4/6*
That came together pretty quickly and smoothly.
Connections
Puzzle #327
I thought we were going for music genres at first for the yellow
1 Like
Wordle 1,049 3/6
Ha, had a feeling.
Daily Octordle #830
Score: 57
Got lucky on the last word as it was a blind 1-in-4 shot.
Connections
Puzzle #327
WTF is a sample sale?
3 Likes
Wordle 1,049 3/6
Daily Octordle #830
Score: 68
This game takes some getting used-to I feel like! After this I thought about strategy for a little while, then tried playing yesterday’s (#829) and did only slightly better:
Score: 67
I have more strategic ideas to try in the future so we’ll see. I’ve been focused on clearing out common letters early on, but maybe there’s also some advantage in checking uncommon letters, since they might give you a less ambigious idea of what the word is. Honestly, maybe “mid-common” letters are best in the early game. I’m realizing that you really don’t want to have to repeatedly guess to rule out possibilities on a single word in this game, also—any time you’re intentionally going for a word in one of the columns, you want to be as sure as you can that it’s the right guess, I think. If you’re unsure, I think probably it’s better to perform “tests” to narrow down the possibilities in all the columns, like trying words with none of the letters you’ve used so far. At least, that’s what I suspect right now.
Connections
Puzzle #327
Like username I’ve never heard of nor encountered a sample sale—apparently they’re more common in NYC so maybe the NYT has an outsized impression of how well-known they are. Anyway, I found this vexing, and I kind of imagine that 90% of the other players did too…it might have seemed like cute puzzle design to me if I’d had any chance of catching on.
Connections
Puzzle #327
Like everyone else, I’m outraged.
1 Like
Well, okay, I got deeply distracted by Octordle for a while and managed to significantly improve my play through statistical analysis of the five-letter Scrabble words and repeated experimentation. Before I put this aside and get back to work, here are my conclusions at present. I’ll put them behind a dropdown in case anyone doesn’t want their strategy to be contaminated by another player’s. 
meow meow
I’m almost certain that it’s best to guess the most common letters first at this point. The information you gain from letter hits seems a lot more significant than the information you gain from letter misses, and what you do learn from letter misses tends to be greater if the letter is commonplace.
Because the scoring system rewards guessing words early more than it penalizes guessing them late, I think it’s probably best to spend only two guesses on “test words” before starting your attack. Sometimes you end up with a word or two that has unimpeachably ambiguous solutions towards the end this way, but if you’re lucky you can pare down the possibilities for one word with your guess for another, so I think on the whole it’s still probably more advantageous to do two tests than it is to do three.
Based on my computer analysis of what the most common letters are in each position for all the five-letter Scrabble words, the assumption that it’s best to guess the most common letters first, and the idea that it’s more important to guess as many letters as possible when testing as opposed to checking commonly-repeated letters like “s” and “e” (which I feel like is true based on some limited experimentation), I’ve found that “party” and “borty” (a low grade of diamond) are good first guesses. “Borty” ranks slightly higher on average, but “party” gives you more leeway with what vowels you might want to target in your second guess (plus there’s a remote chance it might be the right word, whereas I kind of doubt they’ll ever use “borty”).
Here are some high-ranking second guesses with “borty”:
- canes (16753)
- panes (16688)
- pales (16581)
- manes (16524)
- males (16417)
- dales (16407)
- gales (16361)
and with “party”:
- cones (16705)
- bones (16688)
- coles (16598)
- boles (16581)
- moles (16369)
- doles (16359)
- goles (16313).
Also, some other strong starting guesses based on this methodology are “canty pores” (16705), “banes porty” (16688), “carns poley” (16492), and “corny tales” (16452).
You might note that none of these guesses include the vowels “i” or “u”, because they’re not quite common enough. However, often a few words in the set do include them, so, if after your first guess, you decide you want to check for one, here are some "i"s for borty:
- saine (13525)
- paise (12549)
- maise (12385)
some "u"s for borty:
- aunes (15501)
- cause (12234)
- pause (12169)
some "i"s for party:
- cines (15994)
- bines (15977)
- biles (15870)
- mines (15765)
- dines (15755)
- miles (15658)
- clies (15410)
and some "u"s for party:
- dunes (15563)
- mules (15466)
- dules (15456)
- gules (15410)
- clues (15030)
- blues (15013).
I kind of doubt that they make use of every single five-letter Scrabble word since some of them are really obscure, so drawing conclusions from the Scrabble words is perhaps not perfect, but I think it at least gets close enough that this is valuable data. I will say that I haven’t seen them disallow a Scrabble word yet as a guess, although I’m very far from having rigorously tested that.
After you do your starting test, it’s worth carefully considering each column, because sometimes, even though it looks like you might have scanty information in a certain column, if you think about it a bit you realize that there’s only one or two words that could reasonably go there. If you’re not 100% sure, it’s still better not to guess just for a single column early on, but you can use that information to guide your other moves, like if there’s another column where you have two guesses and you could eliminate a possibility for one column with your guess for the other.
In the later game, if you still only have a few matched letters for a given column, there’s a distinct possibility that those are the only letters in the word, so it’s worth considering what all words you can make from those letters (including if you repeat some). If that’s not right, it’s quite possible that you repeat a matched letter and add a single missing one to get the right answer instead; look for vowels or common consonants you might not have guessed yet for the missing one.
Sometimes, even with great play, there are situations where there are several possible words and no good way to rule them out, so you just have to guess. At the very least, you can consider what the more common letter(s) is/are for the ambiguous positions (again this is based on the five-letter Scrabble words; the numbers are what percentage of the five-letter Scrabble words have that letter in that position):
s => 12.07, c => 7.12, b => 6.98, p => 6.61, t => 6.30, a => 5.70, m => 5.34, d => 5.27, g => 4.91, r => 4.85, f => 4.62, l => 4.42, h => 3.76, w => 3.19, k => 2.90, n => 2.49, e => 2.35, o => 2.02, v => 1.87, j => 1.56, u => 1.46, y => 1.40, i => 1.28, z => 0.81, q => 0.60, x => 0.12a => 17.43, o => 16.14, e => 12.54, i => 10.64, u => 9.15, r => 7.27, l => 5.41, h => 4.22, n => 2.67, y => 2.08, t => 1.85, p => 1.77, m => 1.46, c => 1.36, w => 1.26, k => 0.73, s => 0.71, d => 0.65, b => 0.63, g => 0.59, x => 0.44, v => 0.40, z => 0.22, f => 0.18, q => 0.11, j => 0.09a => 9.55, r => 9.26, i => 8.10, o => 7.63, n => 7.39, e => 6.83, l => 6.57, u => 5.16, t => 4.77, s => 4.11, m => 3.95, d => 3.02, c => 3.02, p => 2.80, g => 2.77, b => 2.58, w => 2.09, k => 2.06, v => 1.86, y => 1.64, f => 1.36, z => 1.08, x => 1.03, h => 0.93, j => 0.36, q => 0.10e => 17.96, a => 8.29, t => 6.91, i => 6.78, n => 6.09, l => 5.95, r => 5.54, o => 5.40, s => 4.00, k => 3.86, d => 3.64, g => 3.25, p => 3.21, c => 3.16, u => 3.10, m => 3.10, b => 1.86, h => 1.81, f => 1.78, v => 1.19, w => 0.99, z => 0.94, y => 0.83, j => 0.22, x => 0.09, q => 0.02s => 30.54, e => 11.75, y => 9.98, d => 6.36, t => 5.63, a => 5.23, r => 5.20, n => 4.10, l => 3.69, o => 2.94, h => 2.86, i => 2.16, k => 2.01, m => 1.40, p => 1.14, g => 1.10, c => 0.98, f => 0.63, x => 0.54, u => 0.50, w => 0.49, b => 0.46, z => 0.24, q => 0.03, v => 0.03, j => 0.02
As far as game score goes, a score in the 50s appears to be very good, and a score in the low 50s seems excellent. Without remarkable luck, it’s hardly possible to do much better, as far as I can tell.
2 Likes
your posting prowess is awe-producing.
1 Like
You’ve put a lot more thought into this than I have, but I will share my personal having played this a few hundred times thoughts:
-
52 is what I consider a clean run and the best “reasonable” score, it is what you get with two initial info gathering clues and then getting guesses 3 through 10 correct.
-
My personal best is somewhere in this topic, I believe it is at worst 49 or at best… maybe 44? That’s basically flukish.
-
In case you are curious my starting words are noisy & later, which take care of most of the letters I want in roughly the spots I want to test for them while not being so uncommon that they’d never come up.
-
In the sequence variant you only get info for one column at a time, I play that as my personal hard mode as I give myself only one initial info gathering guess; I have on rare occasions gotten a few 52s but generally my score is much worse there.
-
As I’ve said several times in here I consider 60 as my cutoff between success and not, although TBF some sets of words are much trickier than others.
1 Like
Wordle 1,050 3/6
That felt like the best option, and it actually was.
Daily Octordle #831
Score: 69
Touching on the above post, this is a fairly iffy score.
Daily Sequence Octordle #831
Score: 57
…While this is a fairly good one especially with my self-imposed limitations (even with me brainfarting my last guess and costing myself a point).
Connections
Puzzle #328
Oops. Also I didn’t even get to it but that purple category is beyond a stretch.
1 Like
Wordle 1,050 6/6*
Connections
Puzzle #328
2 Likes
Wordle 1,050 4/6
cares
mange
table
value
Daily Octordle #831
Score: 55
My gambles mostly paid off this time, especially on my second turn.
tried: borty buyer store ether metro chime cloak clown vegan dodgy fluid
Connections
Puzzle #328
I was helped by the fact that I had a phase around the age of 12 where I got really into lockpicking. Also, the sample sale thing from last time got me to pay more attention to which words are kind of ambiguous vs. only seem to fit in one category at first. Going forward it’s good to know that homophones are something they might do…I was a little suspicious at first about something vaguely like that with carrot and jewel (although it didn’t occur to me then that maybe jewel = joule or I would’ve gotten that one right away ).
2 Likes
Awww thank you!! ^^ I’m glad the products of my distraction could be useful to someone else.
That makes sense, I think that’s more-or-less the same conclusion I came to.
Wow…yeah I guess you would have to play really well and get really lucky too.
That makes sense; based both on my quantitative analysis and my intuition so far, I think the letters s, e, a, o, r, i, l, t, and n are all good ones to identify early (in fact those are the most common letters, in order, in the five-letter Scrabble words ) and it’s also good to check for y in the last place because it’s the third-most-common letter there.
I’ll have to try that…sounds very hard.
That makes sense. Yeah, honestly I’m coming to realize that luck does play a fairly significant role—there’s always a certain amount of gambling and guesswork involved, and some words just seem easier to think of than others too for some reason. I guess, one potential advantage of playing this game for a long time is it might help you think of rhyming or alliterative words more easily.
1 Like
I was really happy with getting this one perfect. The trick really is to spot the ambiguous words and avoid them until you’ve closed off their alternate possibilities.
Connections
Puzzle #328
2 Likes
Congrats, yeah, me too 
And yes, I agree. The last game made me realize that they like to be tricky that way. 
It’s interesting to me to think about the differences between the Wordle-type games and Connections. The Wordle-types have a strong quantitative streak—they’re quite amenable to mathematical analysis and I think you could write bots to play them that would be competitive with the best human players probably (although you would need to track down a good corpus to determine common five-letter words from—the Scrabble dictionary has too much obscure stuff I estimate that around 30–40% of the five-letter words in it are actually in the Wordle pool). Connections is more of a game of social intution—you have to kind of get inside the head of the person who designed the puzzle that day and try to retrace their steps. It involves a lot of free-associative lateral thinking, trivia knowledge comes in handy, etc. I think it would be very hard to make a computer program that could play that game well consistently.
At the same time, Wordle-types do involve a certain amount of social intution, in terms of things like what words you would guess the puzzle designer would actually pick to use and (for the games like Octordle) what they would think a good spread would be (I’ve noticed that they seem to set it up so that the words don’t all share too many letters, for instance). Connections does have certain “purely mechanical” strategic aspects, also, like how you know each category has only 4 words and can use that to eliminate ambiguous options.
One thing about games more like Connections is that it can be tricky for the designer to create a good puzzle—it’s a lot more subjective what will work well, and some players will always have an easier time with some puzzles than others just because of their general background knowledge and so on. With Wordle-type games, the biggest place the puzzle designer can misstep, I think, is picking words that are too obscure for the audience, and although that is a real concern of course it seems easier to avoid on average. At the same time, though, the experience of playing any particular Wordle-type puzzle is going to be more similar to the experience of playing any other; with a Connections-type game each puzzle can be more unique and surprising, for better or for worse.
I don’t think these traits make one type of game superior to the other, they’re both pretty fun I think although I can imagine that some people would more favor one type over the other depending on what sort of problem solving they like better. It’s interesting to reflect on for me for game design reasons; I suppose all puzzle games, or puzzles within games, involve some amount of both types of design, even though in some ways the problems that come up in one type of design vs. the other are very different to work through.
2 Likes
Wordle 1,051 4/6
Dang, that third one felt very good too.
Daily Octordle #832
Score: 76
Oof, just a series of rough breaks there.
Connections
Puzzle #329
These are much easier when you figure out the purples early.
3 Likes
Wordle 1,051 5/6*
good thing I eliminated f early
Connections
Puzzle #329
Of course I got purple instantly
3 Likes
Wordle 1,052 6/6*
One of those
Connections
Puzzle #330
Another instant purple.
3 Likes
Wordle 1,052 3/6
That could have gone wrong, but got a bit of luck there.
Daily Octordle #833
Score: 58
Some tricky words in there, I swear khaki pops up more than one would assume.
Connections
Puzzle #330
I managed to pull back from the brink of stubborness and figure it out.
2 Likes
Wordle 1,053 3/6
I’m really glad I couldn’t think of a better word >_>
Daily Octordle #834
Score: 74
Up and down few days here.
Connections
Puzzle #331
I was more baffled by green than purple this time.
2 Likes