Grrrr…working on the new one after it reset. I have a 994, 995 and 998/1000 and can’t come up with the damn word.
You all must have brains wired differently from me cause I quit about 20 words into my first game and the descriptions of what you all are going through sounds absolutely miserable.
5 Likes
Took me forever to get a top 1000 word but pretty quick after that. 75 guesses
:harold:
:harold:
:harold:
So many bad “hits” that have nothing to do with the word.
+1
I tried it once. That was enough.
I had all of those too. You’re gonna be real mad!
They’re talking about Wordle now on Olympics broadcast while Shuster sweeps.
Wat. Today’s will be the last of these I do. We got #998/1000, and guessed for a long time, and finally gave up. After we revealed the result, we decided never again.
Give up, look at the answer, and delete your browser history.
Argh. I got it eventually by branching back out to other animal body parts but that sucked. Would have been more likely to land on it randomly based on some other top 1000 guesses if I’d never gotten the 990s ones.
I still think I hate the game but I’ve gotten two in a row now so I’ll probably keep trying like an asshole.
My early hits were curved, blade, bird, and spear in that order. I was so confused.
At least I’ve learned that if I’ve only gotten one or two words in the top 1000, I shouldn’t get too attached to them and still need to scattershot. But I’d expect 950+ stuff to really help zero in on it in a logical way. With the previous day’s word I really feel like this held true, but maybe the part of speech has a lot to do with it. It may just not be great at comparing/contrasting nouns, or there’s too much “noise” from phrases that contain them as mentioned earlier.
I’m not sure why you guys struggle so much with Semantle…
Just kidding, semantle is not fun. I started with one of the words mentioned a few posts up and luckboxed this one, but gave up yesterday after never getting higher than 16.xx
1 Like
Daily Dordle #0019 4&3/7
The distribution of words in jargon heavy text - eg legal documents, medical journals can be quite different as opposed to regular text. As a result you’ll get differing notions of semantic similarity between individual words.
Also words not found in a particular vocabulary generate random representations in word2vec. Imagine all the common words probably never seen in legal texts.
(There are other models that solve this in interesting ways - each with pros and cons)
My best guess is that they are using the pre-trained word2vec model by Google which is trained on the Google news dataset. This isn’t a bad dataset to model general vocab.
Its actually not very expensive (time and $$ wise) to train the model on large dataset but the biggest issue is obtaining such a dataset in the first place.
Wordle 238 5/6*
Killing it on Dordle.
Daily Dordle #0020 3&4/7
On Wordle, I typed in RONIN but thought better of it at the last second before pressing enter.
Wordle 239 4/6*
Went super saiyan on Nerdle too.
nerdlegame 25 3/6
Pleasantly surprised to get this in 3 after the first two guesses didn’t give much info.
Wordle 239 3/6*
I was just trying to check vowels after coming up with no AEU on the first two guesses
What were 994, 995 and 998 if you still have a record?
I was throwing darts at animal parts for a while before I remembered that obvious thing they have.
Two in a row!
Wordle 239 2/6*
1 Like