aja8888
Moderator Emeritus
Thanks!
Daily WORDLE - 2022
- Thread starter Lsbcal
- Start date
- Status
Thanks!
SecondCor521
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
In regular mode, each new guess merely needs to be a word.
In hard mode, not only does each new guess need to be a word, it also must contain each yellow letter, and it also must contain each green letter in the correct location.
Thus in hard mode, you have fewer and fewer words to choose from. Which is helpful in one sense - you're choosing a new word from among fewer options, so technically your guess has a better chance of being the final right answer. It's also easier in the sense that the game will not let you make a guess that isn't at least as "right" as the facts that you know so far.
But it's harder in the sense that if your vocabulary isn't vast and the answer hasn't yet occurred to you, it may be harder to think of a word for your next (most likely intermediate) guess at all. It's also harder in that it generally rules out the three-word "GLYPH" approach discussed earlier.
RunningBum
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
- Joined
- Jun 18, 2007
- Messages
- 13,273
Suppose you get 3 letters correct, but have no idea what the word might be. In hard mode you only get two more guesses for new letters since you have to use the 3 you found. In "easy" mode you could pick 5 new letters, giving you a greater chance to find them. So you might use your first three guesses trying to find all the letters, then the last three trying to unscramble them and figure out where they go. Is that easier? I dunno.
Stroke of luck today - got this one on the second try!
Wordle 216 2/6
?
??
Wordle 216 2/6
???ERD50
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
Right, and I do think "hard mode" is a poor description.
It's really a trade-off. You obviously give up one turn of guessing the right answer, but by ignoring the correct letters and trying other letters, you gain some knowledge which could help on the turn after that, and might keep you from losing altogether. But it's always going to cost you a turn. It's for the 'delayed gratification' crowd.
To me "hard mode" is a better description for what some are doing, using the previous solution as the starting word. That normally makes it tougher, as those words aren't chosen for maximum elimination. I'll need to play a little longer without failure before I try that, but it's an interesting take.
-ERD50
- Joined
- Nov 17, 2015
- Messages
- 14,152
Well, considering the dude wrote this for GF and family, and not bazillions of people to critique everyday worldwide, I think I can forgive some of the terminology.
ERD50
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
Geez, I didn't mean it as a personal insult to the developer. I was just commiserating with others that were confused by the term, as was I. There's nothing to forgive, not a big deal, I just felt a little explanation was in order.
-ERD50
I'm below average at word games, but above average at spoiling games with too much analysis
AUDIO
RENTS
GLYPH
I didn't spend too much time on this, but the above 3 words gives me 84.2% coverage. Can anyone think of three words higher than 84%? A puzzle within a puzzle!!
Letter|English Expected
E |0.111607|
A |0.084966|
R |0.075809|
I |0.075448|
O |0.071635|
T |0.069509|
N |0.066544|
S |0.057351|
L |0.054893|
C |0.045388|
U |0.036308|
D |0.033844|
P |0.031671|
M |0.030129|
H |0.030034|
G |0.024705|
B |0.02072|
F |0.018121|
Y |0.017779|
W |0.012899|
K |0.011016|
V |0.010074|
X |0.002902|
Z |0.002722|
J |0.001965|
Q |0.001962|
I'm gonna throw a bit of a curve ball into this analysis, which I've seen replicated elsewhere. My thinking is that since Wordle only uses five-letter words that that would bias these letter frequency statistics somewhat. So, I started with the list of English language words in the "/usr/share/dict/words" file on my Mac (other *nix systems have a similar word list, by the way), which apparently has 235886 words in it, including some proper names. I then extracted just the five-letter words with a simple "egrep" command (I'm kind of a *nix command line guy), made all the letters lower-case, and then eliminated any duplicates (there were a surprising number of those). The resulting list has only 9972 words, some of which are names but those appear to be valid in Wordle, which was convenient. I then whipped up a quick little Python script to count how many times each letter appears in that list and print the results, along with the percentage of the total number of letters in the list (49860, which is 5 times the number of words, as expected). Here's the frequency table I got:
A|5627|0.112856|
E|4801|0.096290|
R|3460|0.069394|
I|3238|0.064942|
O|3237|0.064922|
S|2814|0.056438|
L|2719|0.054533|
N|2691|0.053971|
T|2662|0.053389|
U|2264|0.045407|
C|1792|0.035941|
Y|1789|0.035880|
D|1640|0.032892|
M|1580|0.031689|
H|1550|0.031087|
P|1450|0.029081|
B|1336|0.026795|
G|1261|0.025291|
K|1074|0.021540|
W|740|0.014842|
F|724|0.014521|
V|541|0.010850|
Z|306|0.006137|
X|233|0.004673|
J|233|0.004673|
Q|98|0.001966|
Hope others find this useful, and sorry if it contributes to any analytical spoiling of the game.
SecondCor521
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
^ A real Unix person would do it all in one command line ;-)
I wonder if the Wordle creator was clever enough to realize that people would do this type of analysis. If he was, he could do that math, then calculate the weight of any given five letter word in his list of words, then probabilistically choose words based on the inverse of those probabilities, so that the distribution of letters in the daily words chosen would be approximately equal.
That would be fairly clever and also fairly easy to do. And I suspect he may have done something like that - after playing the game for about five or six days, the number of A's and E's in the answers appears to be too low to a statistically suspicious (*) degree.
(*) Right after finishing my college statistics class I might have been able to work out the math on that. Something about the T-distribution and degrees of freedom? At this point it just feels off, but not provably so.
I wonder if the Wordle creator was clever enough to realize that people would do this type of analysis. If he was, he could do that math, then calculate the weight of any given five letter word in his list of words, then probabilistically choose words based on the inverse of those probabilities, so that the distribution of letters in the daily words chosen would be approximately equal.
That would be fairly clever and also fairly easy to do. And I suspect he may have done something like that - after playing the game for about five or six days, the number of A's and E's in the answers appears to be too low to a statistically suspicious (*) degree.
(*) Right after finishing my college statistics class I might have been able to work out the math on that. Something about the T-distribution and degrees of freedom? At this point it just feels off, but not provably so.
Out-to-Lunch
Thinks s/he gets paid by the post
So, based on this reanalysis, "solar" might be a better starting word than "irate"?
RAISE perhaps, based on that
JoeWras
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
- Joined
- Sep 18, 2012
- Messages
- 11,775
I found today's difficult. I literally couldn't think of a word after round 2, the way my letters fell. It boxed in my brain.
I needed DW to help. She guessed one real close, but it was wrong, although the next choice was then obvious (just one letter).
Ended up a combined 4.
I needed DW to help. She guessed one real close, but it was wrong, although the next choice was then obvious (just one letter).
Ended up a combined 4.
Does your first guess actually need to be a word, or can it be random (common) letters?
PaunchyPirate
Thinks s/he gets paid by the post
All guesses must be a word in the database that is used. It will tell you if it is not and force you to try again.
Last edited:
PaunchyPirate
Thinks s/he gets paid by the post
In 4 tries today
Wordle 217 4/6*
??
???
?
??
Wordle 217 4/6*
???
?????According to the guy who developed a variant called Absurdle (https://qntm.org/wordle):
Might be interesting to see if the 2315 words follow this pattern. I've been using "ouija" as a starter, but maybe "raise" is a better option.
ERD50
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
Got it in 4 today, but I had it down to just 2 equally likely words, so 50/50 I could have been at 5. I'll take it!
- so I'm 4,4,3,3,5,4 so far.
Are proper names or place names allowed? In a private window, I tried the name of a river, that does not appear to have a common word meaning, and it accepted it, so I guess so?
edit/add: Based on below, I am going to change my 1st word. I replayed today's game in a private window, and I think it would have helped me knock off one guess.
-ERD50
- so I'm 4,4,3,3,5,4 so far.
Are proper names or place names allowed? In a private window, I tried the name of a river, that does not appear to have a common word meaning, and it accepted it, so I guess so?
edit/add: Based on below, I am going to change my 1st word. I replayed today's game in a private window, and I think it would have helped me knock off one guess.
-ERD50
Last edited:
ERD50
Give me a museum and I'll fill it. (Picasso) Give me a forum ...
Nice. Now how about the frequency of double letters in that list? Those are tricky.
-ERD50
- Status
Similar threads
