That’s where I am with today’s game (in hard mode). I’ve played two guesses and have two green letters, but the number of possible words that remain is so minimal that I can’t think of one that fits right now. Once I figure it out, and I will, I will solve today’s in three tries.
As for the OP, there is a limited universe of answers and a limited universe of possible guesses, limited even more by hard mode, and you are somehow surprised that 4 shows up as the median? I bet 3 and 5 are the next highest guesses for most people, followed by 2 and 6. If only there were a phrase for this tendency for the number of guesses to be found centrally. I wonder if other distributions end up looking like that normally. Next thing you’ll be telling me that the sum of rolling two dice is more likely to end up 7 than 6 or 5. Scam, I say!
To RitterSport’s point, I’m obviously assuming that the people who play Wordle WANT to play Wordle, they like it, they enjoy it, they find it a challenge to their wits. And while “scam” and “pure luck” may be overstating my case, I don’t think it overstates by far. I see people moaning about how they did, or beaming with pride, depending on whether they got it in 5 today or 2, and I think there’s some comfort in knowing that it’s far more random than they think.
I don’t think so. You’d have a better case if the answer were often a rare or obscure word. But whenever I think of two possible choices, I always choose the less rare one, and I’m always right, so far. Not much skill there, or much breadth of vocabulary.
I suppose that could be a way to “win” my anti-Wordle game, in which you try not to win in the fewest moves but rather to delay winning the longest: always play the more obscure word when you get a choice of two possible moves…
I can certainly counter it. As noted, after two guesses I stepped away for a couple of minutes as I couldn’t immediately come up with a valid word. I then came up with two while away from my computer. I picked the one that I consider more common. Oops, it was the other one. I then went and checked and I was correct about which is more commonly used (hovering over the link shows spoilers for today’s Wordle scam. You have been warned).
Wordle 240 4/6*
RATIO
PENIS
MANIC - More commonly used than CYNIC
CYNIC
I’m sure you’re right mathematically (and interesting site, btw) but I consider both choices in the same range of commonness–neither is a highly specialized word that rarely gets used in conversation (I’ve probably used both in the week past) and neither is a basic building block word either.
I’m glad that you consider them the same as it proves my point. In 2019, when “manic” was at its lowest usage in over a decade, it was still almost 4 times more frequently found in their corpus. Remember, I was responding directly to this statement (bolding mine):
Well, I didn’t take that point quite that mathematically literally. If I have a choice between a common word and a borderline scrabble word (or a plural or verb form ending in “s”), the common word will almost certainly be the correct choice. If my choice is between WOVEN and WOXEN, well, I guarantee it’s WOVEN, even though both words are allowable as guesses. Remember, the answer set is only around two thousand something words, and if words like WOXEN and ETUIS and VITAE show up, players will be pissed.
Well, that’s me finding the two choices in the same general range–maybe it costs me a move on the occasions that I’m wrong in my guesses. But I still usually get it in 4 moves, as does everyone who plays a bunch of games. I’m obviously working off the same small sample size as everybody else.
What I’ve seen with computer algorithms is that given perfect play, a computer can average about a 3.6 over the full set of known Wordle answer words. If it is allowed to remember previous answers and eliminate them, then you can get it down to 3.3 over the entire set. So averaging 4 for human play sounds about right. (And, yes, one part of the algorithm does involve a word frequency chart.)
This is the second thread in which you call Wordle a “scam” and write a long rant that basically claims everyone who plays it is an idiot for doing so.
I really don’t understand why you have such animosity towards an innocuous little word game. It’s really bizarre.
And just what is the “scam” here? It’s free and you don’t even sign in to anything to play. So they are getting nothing from you. It’s no different than calling newspaper puzzles and jumbles “scams.”
I call myself an idiot? I don’t remember doing that, but then again if I’m an idiot maybe I have.
I like playing the game. It’s harmless. But I do think it’s silly to feel good or bad depending on how you did that day. Just as rolling dice would be, and feeling great if you rolled a 7 on your first try, and bummed out if you rolled six times and never got a 7.
If it truly is luck, go play it in a language you don’t speak and report back to us your 4.0 average. As someone else already said, I can play a dice in any language.
Your third guess could not have been correct because your first guess showed A wasn’t in the answer.
Wordle can be solved, using an optimal strategy, in 3.42 guesses on average, and can (on normal mode) always be solved in 5 guesses. On hard mode, it takes 3.51 guesses on average, and a solution can always be found in 6 guesses. That’s if you use the 2315 words known to be possible solutions.
If you don’t, I’m not sure there’s an optimality result for the average case, but you can always solve it in 6 moves on normal mode, 14 moves on hard. The latter isn’t known to be optimal. (See here for the numbers).
This video by 3Blue1Brown explains very nicely how different strategies lead to different average numbers of guesses, by optimizing the possible reduction in entropy for each new guess.
(Answer: no, I didn’t. Just mostly luck. Of course you need to speak the language, and to have a normal breadth of vocabulary, and to follow my few simple rules of playing, just like in dice you have to know simple addition, and to have some facility for throwing small objects.)
That’s where I get my 3.6 estimate from (that uses word frequency data but not the Wordle list itself). His optimal 3.43 comes from using the set of known Wordle answer words and also eliminating words that have been used in previous puzzles. Is your 3.42 from the same source? If not, what are the constraints on that.
(Note that he also has an update to that video, where he realizes he screwed up a little bit regarding repeated letters, but it doesn’t change the end analysis much, except the “best” starting word for a computer employing optimum strategy changed.)
Good catch and I thought hard mode was supposed to stop that. It turns out it only requires you to use “revealed hints”, but doesn’t stop you from using letters marked invalid. See, I should have gotten it in three.
Edit: My average is 4.021276596, so I’m definitely worse than a computer program, in case the snafu above wasn’t already proof.
