tuffers#1 wrote: ↑Wed Aug 03, 2022 5:28 pm
Huh ? youre clearly not very good at understanding stats are you. The 10 most used letters show that random probability of 1 in trillions as opposed to 1 in 26 especially as 1 in 10 are the most frequently used of the stated 26 . I'll waist no more time on year 9 & 10 mathematics to show your ludicrous comparable statistic .
Let's ignore the fact that you're not consistently using the most common letters (e.g. 391 where you used the same vowel twice in the opening guess and got them both correct). The starting point for all my calculations is using the probability as if you were playing the perfect game for what I am measuring.
For game length I used the best bot result from here,
https://freshman.dev/wordle/#/leaderboard and took the probability of winning a game in three or less (0.692)
For chance of a correctly placed letter on the first go I use the information here
https://artofproblemsolving.com/blog/ar ... ing-wordle (0.546)
I then use a binomial probability calculator
https://stattrek.com/online-calculator/binomial to calculate against results you have posted.
For game length I use games where you achieved three or less as the success measure (x). For first letter it's 0.546/5 as the probability for an individual letter to be correct, and then the number of correct letters in the sample size as x, number of games is also multiplied by 5. Has to be for individual letters rather than the game as a whole as you can't have more successes than trials.
The thing about being a bit better than is probable is that from a small sample it's impossible to state anything significant, so posting a 6 and saying 'Look I'm not cheating' doesn't really work.
For example rolling a 5 or better in a dice game has a probability of 1/3. The chance of rolling 5 or better 3 times out of 5 is about 20% so It will happen fairly often. Keep up that same ratio though for 30 times out of 50 and the probability of doing that well drops to 0.01%.
This is how I worked out your first 63 posts had a probability less than thst of winning the lottery - I think you had one or two where you got a 4 or more rather than the 19 or so expected. Doing that well over 10 or so games is plausible, doing it over 60+ games means you have an idea of what the word is before you start. I think your previous claim that the words aren't random and you have seen the pattern is more plausible than claiming you are especially skilled at guessing a random word in a logical way.
Where the numbers get truly astronomical is looking at frequency of individual letters guessed correctly taking into account their position. For reasons I don't understand the first letter is the least likely to be guessed on the first couple of attempts, second and third easiest and last two in the middle. A quick look at your results shows your guesses are on average pretty much exactly the wrong way round, again as if you know the word you're aiming for before you start. I might get round to doing the calculation but will have to do it manually, even the online calculator I found that gives at least 25 decimal places was giving a big fat 0 chance on the 20 game sample I put in it.
Obviously you're not cheating, and these calculations are only valid if you consider the word to guess as being randomly chosen, which as youve pointed out, it isnt.. I was doing this for my own slightly nerdy satisfaction of seeing whether it was possible to quantify how lucky someone is at wordle, I may well have the maths wrong or not considered something. I did consider Gary's suggestion that you only post the good scores but you seem to post regularly for periods then have a break of a few days, I think that pattern of doing less well in discrete chunks before returning to form is even less likely than the scores you have posted.