I didn’t choose the highest skill 2nd guess, but I had the highest skill among players with my starting word. My choice was the 5th best word. The next highest skill player chose the 19th best word. However, luck was too low to solve in 3.
Wordle 340 4/6
WordleBot
Skill 98/100
Luck 39/100
.
2 of the puzzles had especially lucky guesses. One had an unlucky guess.
In the “How our guesses divided solutions” section Wordlebot provides a supporting analysis that compares the results with your actual guess vs other words including the ones Wordlebot rates highest. While sometimes counterintuitive, I generally agree with Wordlebot’s conclusion. The differences I find are minor and often relate to differences in opinions about what should be optimized or possible solutions.
Regarding testing out all the vowels, Wordlebot is unlikely to favor this type of strategy unless all of the vowels are among the most common letters in the remaining solutions, which is rarely the case. However, if seeing the vowels is far more likely for you to recognize the solution compared to seeing hit consonants, then it may be a good strategy for personal play, even if Wordlebot doesn’t like it.
I’ll use the May 23rd puzzle as an example in case people haven’t finished May 24th. My first guess was ROATE. Only the E hit with green, leading to:
####E
In this example, I knew there was likely a 2nd vowel and it was not an ‘O’ or ‘A’. ‘I’ is the most likely option. ‘U’ and a 2nd ‘E’ are also likely. In general ‘U’ isn’t a common letter among possible solutions, but with the "A’ and "O’ eliminated from the first guess and another vowel remaining, this is a rare puzzle where ‘U’ is likely and guessing all of remaining vowels (not counting 2nd E) may be near optimal. Other common letters among possible solutions include '‘G’, ‘L’, ‘N’, ‘S’, and to a lesser extent ‘C’ and ‘D’,
There are a variety of words that can be made with the first 6 letters (IUGLNS), but few with both ‘I’ and ‘U’. Ignoring the words that have ‘I’ and ‘U’ in start/end due to bad positioning, I believe that just leaves SUING. I instead chose SLING, which Wordlebot rated as 99. The other guesses Wordlebot rated highest were:
SLING – 99, Ave. guesses = 1.57, Ave. solutions = 3.0, Ave. group size = 2.0, Largest group = 6
GUILD – 95, Ave. guesses = 1.62, Ave. solutions = 4.2, Ave. group size = 2.1, Largest group = 10
DINGS – 94, Ave. guesses = 1.63, Ave. solutions = 3.9, Ave. group size = 2.2, Largest group = 9
GILDS – 93, Ave. guesses = 1.64, Ave. solutions = 4.1, Ave. group size = 2.2, Largest group = 8
SUING – 93, Ave. guesses = 1.64, Ave. solutions = 3.5, Ave. group size = 2.2, Largest group = 7
…
SLUNG – 86, Ave. guesses = 1.72, Ave. solutions = 4.1, Ave. group size = 2.6, Largest group = 8
SLING and SUING differ in only the 2nd letter – ‘L’ vs ‘U’. I suspect SLING rated slightly higher because of the inverse correlation between ‘U’ and ‘I’. If the ‘I’ doesn’t hit, then in this particular instance there is a high probability of a ‘U’ being somewhere in the word. You know there is a likely ‘U’ without guessing ‘U’. Guessing ‘U’ to learn the positioning of the ‘U’ helps, but you gain more information with the ‘L’ of SLING since the ‘L’ is less corelated with the ‘I’ and other letters.