Wordle 246 2/6*

:black_large_square::large_green_square::large_yellow_square::black_large_square::black_large_square: :large_green_square::large_green_square::large_green_square::large_green_square::large_green_square:

Wordle 246 3/6

:black_large_square::large_green_square::large_yellow_square::black_large_square::black_large_square: :large_yellow_square::black_large_square::black_large_square::black_large_square::large_green_square: :large_green_square::large_green_square::large_green_square::large_green_square::large_green_square:

Woah, impressive! How did you come to choose that word as second guess?

And here I was, happy with myself that I found it in 3 :sweat_smile:

Did you make the assumption that one letter would be doubled?

Wordle 246 3/6

:black_large_square::large_green_square::black_large_square::large_green_square::large_yellow_square: :black_large_square::large_green_square::large_green_square::black_large_square::black_large_square: :large_green_square::large_green_square::large_green_square::large_green_square::large_green_square:

I know it’s A_I

where S, E, R are not valid letters.

I can only think of

BASIC BASIL

which contain S

I was focusing on T, so TACIT was the only one that came up to my mind, so I played it.

Honestly, I probably should have played VALID instead.

Wordle 246 3/6*

:white_large_square::large_green_square::large_yellow_square::white_large_square::white_large_square: :white_large_square::large_green_square::large_yellow_square::large_green_square::white_large_square: :large_green_square::large_green_square::large_green_square::large_green_square::large_green_square:

Nice. The fun part is that whenever we get to VALID, you can actually play it and beat the optimal expectation! :smile:

Hm, when calculating the expectation, maybe I should actually do just that: remove the words that have already been played. That would give a better average case than 3.42!

Now, finding the optimal strategy knowing that the data is drawn without replacement is a much harder problem than just assuming i.i.d.

I think that simply amounts to moving words that have already been played from the solution list to the allowable guess list, no?

Yes, but what first words to choose depends on the previous games, so in a sense you need to simulate every ordering of all 2315 games to find the optimal first guess for the first word, which is 2315! games. Such a small number.

Ah no, it’s simpler than that, you’re right. The very first guess should just be SALET. But in the next day you need to resolve everything again, which is not that bad (except that providing optionality took the guy 2 days)