Nobody knows better than the New York Times the constantly shifting sands of the English language -- and yesterday saw those shifts cause changes in the behavior of WordleBot, the AI used to evaluate play in NYT's Wordle game (hereafter referred to as "The Bot"). If it seems like I think this is important news, let me apologize for misleading you.
It's not.
On the other hand, if you're a complete Wordle Nerd (as I am), it was a *bit* interesting.
I'm sharing my feedback on the New York Time's Wordle Review comments yesterday regarding this weighty matter --
Wordle 936 3/6
⬜⬜⬜⬜🟨 SHONE 496 WL
⬜🟩⬜🟨⬜ TRAIL 10 WL
🟩🟩🟩🟩🟩 BRIEF
WordleBot
Skill 99/99
Luck 57/99
Streak 687
The first change I noticed concerning yesterday’s updates to The Bot’s Dictionary was this: Identifying E out of position as Letter 5 for Guess #1 left me 496 possible solutions instead of 492. I stuck with my trusty TRIAL for Guess #2, finding R (in position) and I (out of position). Based on The Bot’s feedback from Game #699 (GRIEF on 5/19/2023), I strongly suspected there were 3 high-probability solutions remaining (rather than the 10 later claimed by The Bot). I decided to roll the dice and my suspense was BRIEF, as all the letters quickly turned green. I tied The Bot and bested the (current) NYT avg by more than a full guess today (3 vs. 4.2).
My hope for matching my 3-Guess and 4-Guess win totals at 300 still springs eternal (they currently stand at 297 vs 300). If I can retain a “hot hand” through the weekend, this goal may yet be achieved.
And now I see that The Bot has adjusted its play for Game #936 since I played early this morning. With the result being that I BEAT The Bot (3 vs 4). At first, I thought I had just erred while posting comments during the middle of the night. But The Bot’s original solution was TRACE—DEFER—BRIEF, later changed to TRACE—DEFER—GRIEF—BRIEF. Go figure.
Hope you are having a fine Friday.
Take care and stay safe.
No comments:
Post a Comment