Anton Hodgetts 🇬🇧 (scaredycat)and the statistics i am experiencing. more negative today. This is not psychology - it's fact! Numbers don't lie...

"There are lies, damned lies, and statistics" Benjamin Disraeli (probably)

Numbers absolutely do "lie" when they are based on random processes, there is an implicit selection bias in the evidence in favour of a preferred conclusion, and you lack the necessary expertise to correctly interpret their meaning. I am a statistics student at university. I suggest you Google the "Dunning Kruger effect".

I will be sure to share my published results with you when I am finished, and then you can see how to go about this correctly and the amount of work necessary to reach reliable results.

Anton Hodgetts 🇬🇧 (scaredycat)I assume you are still only an undergraduate since you seem to think you know everything about your subject and show youthful arrogance regarding my knowledge when you know nothing of me. I have a PhD in history and I can tell you it was the Duke of Wellington who you quote, furthermore when I see a run of numbers which are mostly negative, it means they are mostly negative, and that is beyond doubt. You clearly have much to learn.

Indeed I am a an undergraduate with much to learn! As we all! My knowledge of statistics is more than sufficient to explain where you are going wrong though, your ignorance on statistics speaks for itself. A PHD in history has absolutely nothing to do with statistics or Backgammon.

Anton Hodgetts 🇬🇧 (scaredycat)I see you deleted your initial reply where you attempted to correct me on my own subject by quoting wikipedia! That was very funny, you're going to struggle with that approach to academia. Read what I wrote, not what you interpret from your reading of it. My numbers do not lie and do indicate a reliable pattern within their own sample. think more... say less... I'll be around for a game some time when you are free

Well all of my replies are still there so I don't understand what you mean about me deleting them? Could you explain in what way you are right and the Wikipedia page is wrong? Give a source explaining who really coined the phrase? I would be interested as it is relevant to my field of study but I lack the historical research skills to check beyond just Googling Wikipedia.

FinnDoris [gb-eng] (finndoris)I will now proceed to explain to you using school level statistical maths why your sample is not statistically significant. First, let us consider the hypothesis test for the number of times you had positive/negative luck. The number of matches out of your 25 matches in which you had positive luck can be modelled using a binomial distribution as follows:

X~Bin(25, p)

H0: p=0.5, H1: p<0.5

P(X<=9|H0) = 0.11476147

0.11476147 > 0.05

the p-value is not lower than 0.05 which is a conventional significance level to use to balance type 1 and type 2 errors.

Therefore we fail to reject H0.

So we conclude that the frequency with which you had negative luck is not statistically significant. If we consider the critical region of such a test, we find that x<=8 is the critical region, so unfortunately you are just shy of the necessary threshold in order for us to change our mind on a statistical basis. The mean luck might be of some interest, but you did not provide the data with sufficient precision to be able to determine significance. This is also not even getting into the issues with which you are collecting the data, you would not be collecting the data in the first place if there weren't some extreme results, and this is just a subset of a much longer sequence of matches you have played on Backgammon Hub. If you would like to read a paper by a prominent statistics professor debunking similar statistical claims made against GM Hikaru Nakamura by GM Vladimir Kramnik claiming that his winning streaks in online Chess is statistically dubious, refer to this link: https://www.chess.com/news/view/nakamura-winning-streaks-statistically-normal-professor-says. I would say this case highlights how knowledge and expertise is domain specific, and just how much experts in one field can overestimate their understanding in another, which is a funny form of the Dunning-Kruger effect.

FinnDoris [gb-eng] (finndoris)Not to mention, even if the results of a test like this was significantly different, do you even have a qualitative explanation to explain it? Alfie has already explained many times to people like you that the dice are pseudorandom predetermined, we actually KNOW how the randomness and luck is coming about so I'm not sure what any statistical inference on your games can tell you. We already know how the system works, there isn't really anything to infer.

Searching for games I have played with you, we have played 12 one point games against each other and I won 8/12. I also had an average PR of 4.3 compared with your 7.3 so I played better. Playing one more game proves nothing. Besides, I have no interest in playing Backgammon with a sore loser who can't accept a bad luck streak in a dice game. I myself am in a bad luck streak but do not complain about it because I understand that if I play Backgammon for long enough I will eventually have a bad luck streak just as I have good luck streaks.

I rest my case, I have better things to do.

I had to break up my replies because the website would not accept messages above a given size

Anton Hodgetts 🇬🇧 (scaredycat) I merely scanned your rant briefly; it looks like you've been drinking. As you are unable to prove me wrong on my subject through wikipedia, allow me to prove you wrong through statistics and factually.

2a. Your average luck was 0.21, mine was 0.10.

2b. All bar one game was won by the player with the better luck. You won that outlier, but that doesn't mean you are a better player than me, quite the opposite since...

2c. Your PR multiplied by your luck gives you an adjusted PR of 0.89, while mine results at 0.71, thus proving I am the better player.

But this argument is not the point, you didn't read what I wrote again, I have never complained about losing, only a run of bad luck, so why did you write that I am a 'sore loser'? In fact If you read what I wrote in this thread you would see that I declared 'I'm happy to lose to anyone better or worse than me'.

You put yourself up as an expert; you talk of frequency and conventional significance etc, refer me to statisticians, post formulae etc., and yet you can't even recognise that you have 6 repeated entries in a 12 datum list! You are embarrassing yourself.

This is a case of the hare and the tortoise. I advised you before to think more, say less. You personify that a little bit of knowledge is a dangerous thing which is not unusual for overexcited undergrads who think they know it all. Think before you pronounce otherwise you will fail your course.

Good morning Finn,

I don't appreciate the aggression and ad hominem. I also encourage you to actually read my replies just as you are telling me to do for you.

Ah yes well done for spotting the repeated data points, I simply missed this as it was late and in my experience this glitch does not happen all of the time.

I am still waiting for your source on who came up with the phrase, I asked ChatGPT (god forbid) and it generally gave me back sources confirming what I said, that we don't know who coined the phrase but that it is generally (and probably wrongly) attributed to Benjamin Disraeli. it also mentioned nothing about the Duke of Wellington...

You can see what it generated here if you want to check sources, I honestly couldn't be bothered by this point:

I am unfamiliar with your definition of luck adjusted PR, (PR multiplied by luck?), I tried googling for it but couldn't find it. Could you kindly direct me to a resource where I can learn more about it?

FinnDoris [gb-eng] (finndoris) I am undoubtedly the better player than you. Looking at the our statistics on our profile pages, you primarily play DMP games so let's compare that. For DMP games, I have an average PR of 4. You have an average PR of 8.8

I did not compare across all match lengths because you almost exclusively play DMP games, and PR is artificially halved in DMP because of the lack of take points. Looking at the few longer matches you have played, your PR is practically that of a beginner at ~18. My median PR in cube matches is ~7 based on my Backgammon Galaxy stats which I believe is more accurate as I play on there more often. My Hub stats are not that different.

There is nothing more to say.

Anton Hodgetts 🇬🇧 (scaredycat) I can't be bothered with you any longer. You made a statistical error that a junior school child wouldn't, and yet you still pretend to be a knowledgeable statistician. You clearly have problems beyond the bounds of Backgammon. I wish your supervisors the best of luck handling your fragile ego and delusions of grandeur.

FinnDoris [gb-eng] (finndoris) If you can't be bothered why are you replying? I'm still waiting for your sources, I'm eager to learn. Surely a historian would cite their sources no?

Alfie Kirkpatrick [Developer] Hi Alfie, yes I read the dice thing a while ago. It was only a request to find other peoples experiences; which dice do they prefer etc., (I use E but will change now). I knew the PR etc. I'd be interested to see what my average luck score is however life's too short to go back that far. There isn't much else to add except that I remain suspicious. In addition to luck in the wider game, too many games are changed where I am well ahead and only one possible combination of dice could swing the game against me (e.g. off the bar to blot my only undefended piece in my home board) and it happens again and again.Too frequently for 'real' dice. I did suspect situational AI on the dice rolls but maybe that's going too far. I even (correctly) predict it happening now and laugh when it does. Oh well, life goes on...

FinnDoris [gb-eng] (finndoris) your suspicion of the dice and "gut feel" that things are wrong is not uncommon at all! But it's hard to argue with if you prefer to listen to your gut over hard stats.

Two things:

For info, here is a breakdown of your average luck/PR for the players you've played most... a positive skill diff means you played worse on average.

Can you spot a pattern? My observation is you are generally coming up against players who have a higher skill level, which is inevitably frustrating and often causes players to doubt their dice!

Alfie Kirkpatrick [Developer] I wouldn't waste any more time on it Alfie. I do always try to nab the E dice quickly and usually get it so they are a fairly good 'norm'. Skill has nothing to do with the very unusual occurrences which change the game too often for my liking. I'm happy to lose to anyone better or worse than me, but I remain suspicious. End of...

just re-read this post and realised it could be interpreted as rude. it was only meant to draw a rather bored line under the subject. Apologies if it appeared otherwise

FinnDthoris [gb-eng] (finndoris)

That happens in real life regularly. It's not called the cruellest game for no reason

FinnDoris [gb-eng] (finndoris) go on Galaxy, play 2000 DMP games, then go on Heroes and play another 2000 DMP games (like you did here). Then come back to let us know if your stats are significantly different in any of the two. Just keep track of your winning percentage, your average PR, and opponent average PR (this part is not easy). Would be interesting.

But when you will complete that probably you will have long time forgot about this thread. Here the rolls are verifiable while on other sites not.

Denis-George Constantin (playhunter)OK that's a challenge. thank you I'll take a look

Denis-George Constantin (playhunter)Hi Playhunter, apparently I need to be invited by a current usewr to play at Heroes. Please can you invite me so I can follow your advice? Thanks