Re: Wordle
Posted: Sun Nov 26, 2023 6:30 am
Wordle 890 3/6















The Unofficial and Independent Leyton Orient Message Board
https://lofcforum.com/forum1/phpBB3/
No you won't, because you know you can't. You'll continue to argue that you haven't cheated and wouldn't even know how to but you'll find a way not to submit to a test.Daily Express bot wrote: βSat Nov 25, 2023 10:40 am
Iβll think about it, let you know but if I win you can give the money to charity
All very interesting. I am slightly intrigued as to why you are actually digging me out though. I can (for what it is worth) confirm that I am not cheating. I am in several other Wordle groups also and it is blatantly obvious when people are sometimes cheating as it is odd when someone does so well when in public they find it difficult to articulate or string a sentence together in ordinary conversation. Cheating at this is pointless , it is like playing alone , cheating at golf. I am not here to impress anyone. Just something to do like a crossword.Mick McQuaid wrote: βSun Nov 26, 2023 10:17 amNo you won't, because you know you can't. You'll continue to argue that you haven't cheated and wouldn't even know how to but you'll find a way not to submit to a test.Daily Express bot wrote: βSat Nov 25, 2023 10:40 am
Iβll think about it, let you know but if I win you can give the money to charity
As with others before you have made the mistake of thinking that solving wordle in 3 is a question purely of skill rather than chance. There is an optimal strategy for wordle which solves in three or less 53% of the time. You can use the brute force of a computer to complete about 1.5 million calculations per guess to achieve this, there are more efficient but complex ways to code this but fundamentally there is no way a human can compete).
Although a human can get lucky in any individual game, you can apply a binomial probability calculation over a series of games to see what the chamce was of that number of scores of three or less. Being a question of probability I can't say you are cheating, only that after scraping the details of 72 posts that there is a 99.998% chance you are cheating.
I've gone out and left the computer running a slightly more advanced analysis taking into account your use of the first word ADIEU. I'll let you know the results later.
Very interesting, I still do not understand the following though, iβm fine with the maths incidentally.Mick McQuaid wrote: βSun Nov 26, 2023 1:10 pm I have no idea whether you cheated today.
If I gave you a die and challenged you to roll a 6 there would be a 16.67% chance of you doing so.
If I asked you to do it 72 times (the number of games you have posted) there would be a 12.5% chance of you rolling exactly 12 6's, a 45% chance you would roll less than 12 and a 42.5% chance of rolling more than 12.
These results follow a bell curve centred around 12. There is a more than 99% probability that the number of 6's would be in the range 5-20, and more than 99.9% probability the range would be between 4 and 22.
Your rclaimed scores are roughly the equivalent of rolling 26 6's out of 72.
Please let me know if you have a problem with the maths.
As there is an optimal strategy for wordle you can discount that and look at it entirely as a game of chance, just assuming everyone is playing the optimal strategy.Daily Express bot wrote: βSun Nov 26, 2023 2:17 pm
Very interesting, I still do not understand the following though, iβm fine with the maths incidentally.
The chances and βluckβ are mathematical probabilities.
There is a certain amount of skill and thought involved in these. How do you measure that?
If I was certainly cheating as your percentages imply , why would I and why would I not return an even greater percentage score?
Thanks
Leo
I want to reply to you last post later today. I have found your analysis very intriguing. I achieved another three today incidentally without cheating.Mick McQuaid wrote: βSun Nov 26, 2023 2:58 pmAs there is an optimal strategy for wordle you can discount that and look at it entirely as a game of chance, just assuming everyone is playing the optimal strategy.Daily Express bot wrote: βSun Nov 26, 2023 2:17 pm
Very interesting, I still do not understand the following though, iβm fine with the maths incidentally.
The chances and βluckβ are mathematical probabilities.
There is a certain amount of skill and thought involved in these. How do you measure that?
If I was certainly cheating as your percentages imply , why would I and why would I not return an even greater percentage score?
Thanks
Leo
Solving wordle in 5 is 100% skill, if by skill you mean the ability to hold in your head every possible wordle answer and every possible allowed input word then calculate, taking into account complicated stuff to do with decision trees and information theory, the words you have remaining and the best possible guess in that scenario.
Solving worDLE IN 4 Iis possible 96% of the time will skill alone. Solving wordle in 3 will only happen 53% of the time. Every dataset I have looked at converges around 26-31% of games solved in three or less for humans, you are claiming to be almost as good as a computer while using a far less than optimal starting word. This is in my view impossible. Look at your scores compared to Admin, StillSpike and mine, ours all follow an almost identical curve if plotted, yours are a mad spike and a downward tail.
As for why you don't just go wild and post a 2 or 3 every day, I am assuming that is because you know that no one would believe you, instead I think the most likely thing you are doing is occasionally editing the score you paste by removing a row now and again, you know where you were really close to getting a three so it's not really cheating? As for why you are cheating at all, I assume it is a wish to appear to be good at a word puzzle.
I will of course eat plenty of humble pie if you take on the challenge above and prove me wrong.
Mick McQuaid wrote: βTue Nov 28, 2023 7:48 am Here's the more detailed analysis that I've been running on and off for the last few days:
First I wrote some code which takes the previous answers list then calculates the list of possible words that are left (using ADIEU for express bot and BRAND for StillSpike) it then runs through all 14,000+ allowed input words to find the word that reduces the possible answers to the smallest list. This gives the absolute minimum number of words a guess in three could have been chosen from.
I did this for the last 52 words for Express Bot and 33 for StillSpike (it was taking ages and I think my point is proved).
But what about luck....maybe Express Bot is just er...consistently lucky. So I simulated the series of games a million times for each person and looked at how many times there was a score of three or less, remember given a guess from the least possible amount of options.
Express Bots claim of 24 scores of 3 or less beats the computer 944,822 times out of a million.
Spike beats the computer precisely 0 times.
Of course, you'll still claim you've never cheated but I'd love to know the flaw in my method.
Code and results below.
from listy import Ans
with open('C:/Users/1/Documents/Alw.text','r') as file:
lines = [line.strip() for line in file.readlines()]
Alw = list(lines)
def scoreit(firstguess,answer):
score = [""] * len(answer)
for i in range(len(answer)):
if firstguess == answer:
score = "G"
elif firstguess in answer:
score = "O"
else:
score = "X"
return score
def processit(score,cAns,firstguess):
for i in range(len(score)):
if score == "G":
cAns = [x for x in cAns if x == firstguess]
elif score == "O":
cAns = [x for x in cAns if firstguess[i] in x]
else:
cAns = [x for x in cAns if firstguess[i] not in x]
return(cAns)
def get_best_word(Alw,cAns,score):
bestword=""
total=0
bestscore=2000
for i in range(len(Alw)):
total=0
for j in range(len(cAns)):
score=scoreit(Alw[i],cAns[j])
ctAns=processit(score,cAns,Alw[i])
total=total+len(ctAns)
if total/len(cAns)<bestscore:
bestword=Alw[i]
bestscore=total/len(cAns)
return bestword
with open('C:/Users/1/Documents/answersh.txt','r') as file:
lines = [line.strip() for line in file.readlines()]
past_answers = [line.split('\t')[-1].upper() for line in lines]
guess1=[]
guess2=[]
guess3=[]
for x in past_answers:
firstguess="adieu"
answer=x.lower()
cAns=Ans
##guess1
score=scoreit(firstguess,answer)
cAns=processit(score,cAns, firstguess)
guess1.append(len(cAns))
print(len(cAns))
#guess2
bestword=get_best_word(Alw,cAns,score)
score=scoreit(bestword,answer)
cAns=processit(score,cAns,bestword)
guess2.append(len(cAns))
print(bestword)
print(cAns)
#guess3
bestword=get_best_word(Alw,cAns,score)
score=scoreit(bestword,answer)
cAns=processit(score,cAns,bestword)
guess3.append(len(cAns))
print(bestword)
print(cAns)
print(guess1)
print(guess2)
print(guess3)
Expressbot:
[162, 23, 47, 94, 47, 82, 102, 47, 188, 216, 327, 126, 275, 43, 275, 275, 94, 216, 46, 275, 94, 5, 34, 61, 275, 94, 162, 327, 323, 16, 323, 19, 323, 47, 102, 188, 327, 323, 1, 327, 275, 35, 82, 323, 188, 94, 145, 94, 61, 188, 188, 61]
[5, 2, 3, 4, 1, 5, 4, 2, 15, 11, 16, 18, 3, 3, 5, 7, 14, 6, 1, 3, 5, 1, 5, 1, 14, 5, 8, 7, 20, 1, 3, 4, 9, 5, 7, 12, 6, 35, 1, 8, 13, 1, 3, 1, 10, 7, 23, 13, 5, 5, 8, 5]
[1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, 2, 1, 1, 1, 2, 3, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 2, 1, 1, 2, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 1]
StillSpike:
[319, 206, 434, 20, 16, 434, 70, 33, 88, 434, 321, 20, 20, 208, 124, 124, 129, 434, 8, 73, 129, 13, 206, 206, 24, 35, 70, 129, 19, 434, 319, 321, 434, 206, 319, 319]
[7, 15, 8, 1, 3, 9, 4, 2, 5, 9, 19, 1, 1, 6, 2, 9, 2, 7, 1, 4, 9, 3, 14, 1, 3, 3, 5, 18, 1, 7, 17, 3, 19, 19, 6, 28]
[1, 1, 1, 1, 1, 1, 1, 1, 2, 3, 5, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 2, 1, 2, 2, 1, 1]
import random
from collections import Counter
after1=[319, 206, 434, 20, 16, 434, 70, 33, 88, 434, 321, 20, 20, 208, 124, 124, 129, 434, 8, 73, 129, 13, 206, 206, 24, 35, 70, 129, 19, 434, 319, 321, 434, 206, 319, 319]
after2=[7, 15, 8, 1, 3, 9, 4, 2, 5, 9, 19, 1, 1, 6, 2, 9, 2, 7, 1, 4, 9, 3, 14, 1, 3, 3, 5, 18, 1, 7, 17, 3, 19, 19, 6, 28]
after3=[1, 1, 1, 1, 1, 1, 1, 1, 2, 3, 5, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 2, 1, 2, 2, 1, 1]
print(len(after3))
score3l=[]
for i in range(1000000):
count=-1
total_score = 0 # Initialize total_score
score2=0
score3=0
score4=0
score5=0
for x in after1:
count=count+1
if x==1:
score=2
else:
y = random.randrange(1,x)
if y == 1:
score = 2
else:
if after2[count]==1:
score=3
else:
y=random.randrange(1,after2[count])
if y==1:
score=3
else:
if after3[count]==1:
score=4
else:
y=random.randrange(1,after3[count])
if y ==1:
score=4
else:
score=5
if score<=3:
score3=score3+1
score3l.append(score3)
frequency_counter = Counter(score3l)
sorted_frequencies = sorted(frequency_counter.items(), key=lambda x: x[0], reverse=True)
print(sorted_frequencies)
expressbot:
[(32, 3), (31, 11), (30, 87), (29, 268), (28, 893), (27, 2581), (26, 6312), (25, 14768), (24, 30255), (23, 54936), (22, 86914), (21, 121931), (20, 148627), (19, 155889), (18, 141814), (17, 108523), (16, 68906), (15, 36182), (14, 15014), (13, 4841), (12, 1074), (11, 162), (10, 9)]
stillspike:
[(25, 4), (24, 22), (23, 139), (22, 690), (21, 2609), (20, 9104), (19, 25423), (18, 57062), (17, 108143), (16, 164621), (15, 199377), (14, 189059), (13, 137546), (12, 72984), (11, 26739), (10, 5885), (9, 593)]
I know my starting word is not the best word to begin with but as I have been using it more less from the beginning I decided to continue as it is βsods lawβ that if I changed it, I could miss out on a mind blowing β1β. Like changing your regular lottery numbers.Mick McQuaid wrote: βTue Nov 28, 2023 9:41 am As Admin doesn't post his score daily it means having to look up and enter the games he's played into a list manually. If I can be arsed I'll do it and run his scores overnight.
It's not just computer says no though, it's computer saying I have rigorously calculated the minimum possible set of words you could have guessed from on each of your last 52 games and tested that a million times to show you are, beyond any possible doubt, regularly posting enhanced scores on here.
By the way. adieu is demonstrably a really bad starting word.
Who is Tuffers? I donβt mind. I donβt cheat for sure. Let me know how we go about it.Long slender neck wrote: βTue Nov 28, 2023 11:47 am I'll challenge you just like I challenged Tuffers. We can play Wordle live on this very forum. Post your first guess and I'll post back
Yes but i think you wanted cameras and stuff? I donβt have a PCMick McQuaid wrote: βTue Nov 28, 2023 12:03 pm Oi. I've already set a challenge and even suggested charity donations.
Mick McQuaid wrote: βSun Nov 26, 2023 2:27 pm I've had a quick look at the results of your scores taking into account your starting word and I'm fairly confident that you won't be able to replicate the results you have posted on here even if you use an online wordle solver.
Here's the updated challenge:
I will post 25 games on a thread on the oh what rubbish forum along with an encrypted string of the games chosen. I will provide the key when all games are concluded to show that they were valid wordle games.
These games will be posted as the feedback of the first guess as if your starting word was ADIEU (X for black, O for Orange, G for green)
You will post your guesses identifying which game you are referring to in any order you like.
I will commit to checking and updating at least every 72 hours. You will commit to at least one answer every 72 hours until the game is complete. If there are circumstances that prevent this from happening it is expected a reasonable explanation will be given as soon as possible.
You have posted that you have scored a three or less 12 times in your last 25 games. For every score of three or less over 12 I will pay Β£25 to a charity of your choice, for every score of 4 or more above 13 you will pay Β£25 to a charity of my choice.
If you get exactly 12 at three or less we will both pay Β£50 to a charity of our choice.
If you get less than 8 in three or less you will post a screenshot of you posting a picture of your bare arse with the words "I am a big fat cheat" printed on it to your whatsapp wordle groups.