To my delight, the New York Times has launched a new tool called WordleBot that can score your performance on any previous Wordle.
If you recently played Wordle, the tool will remember your predictions, but you may also upload a screenshot of a Wordle you previously played. If you don’t want to see spoilers, either stop reading or go solve today’s puzzle, number 293, immediately.
How to make the most of WordleBot
Solve today’s problem first, and then click this link. It will walk you through the entire procedure. Take a screenshot if you wish to analyze an earlier puzzle or one created by another user. (The Wordle stats screen will appear when you answer a puzzle. Simply click the “x” and then take a snapshot of the grid. Additionally, you can return to that page at any time after solving, as long as the new problem has not yet been released, and it will retain your guesses.)
Then, swiping through the result screens is a good idea. The first one assigns you an overall score based on your ability and luck. The remainder of the steps walk you through your predictions, and the bot compares your choices to what it would have done.
How adept at Wordle is WordleBot?
By the way, the bot is a cheater. It determines how well you are restricting the choices, but it also knows which 2,309 words are included in the official answer list. The English language has more than 13,000 five-letter words; many were simply omitted to keep the game interesting, such as plurals of four-letter words that are not frequently included. Additionally, several obscure terms were omitted, as were a number of words that I do not think to be really obscure: FUTON, the bot informed me, will never be a solution, nor will PORKY, but both are permitted as guesses.
On the other side, the bot is unaware of previously published riddles. When I’m playing, if one of the alternatives is SNOUT but I recall that that was the solution last week, I can mentally exclude it. This direction has not been sent to the bot.
How to get knowledge from WordleBot
The most fascinating aspect of WordleBot is not that it can tell you how skilled you are, but that it can also provide you with advice on how to improve.
Today’s problem was quite simple (I and two of my coworkers all solved it in three guesses.) The bot, cocky as it is, claims it would have solved it in two minutes. (It begins with the letter CRANE.)
I received a talent rating of 93 overall. It loved my opener (STONE) and stated that my second prediction of CHAMP down the field to only two options. Thus, my third estimate was fortunate.
Following that, I ran Sarah Showfety’s solution through the bot. It resulted in an overall skill score of just 85, owing to the fact that her second estimate of SNARE reduced the solution pool to three rather than two choices. Deputy editor Joel Cunningham earned an 84 on the talent test, and he appears to have been penalized for being too fortunate in this situation. His opening reduced the number of feasible answers to nine, and after CLAMP, he was left with only one. (We all outperform the average Wordle player, according to the bot: the average skill score is 79.)
What, then, did we learn? We evaluated our starters, among other things. The bot chose CRANE; it also liked Joel’s STERN and Sarah’s ROUTE. All of them will, on average, reduce the pool of options to between 90 and 100. (However, this number may be greater or lower based on the day’s actual answer.) Occasionally, you are simply unfortunate.)
Therefore, feel free to inquire as to how the bot believes you performed. You may be enraged by the answers, but you will almost certainly learn something.