So on ZooEscape, where I already know the opponent dice rolls fail the chi square fit test, my own rolls (331,449) result in 8,766 occurences.

This is 2.64% actual against 2.78% expected.

Observations as to whether these figures fall within the norm?

On 5/13/2021 1:17 PM, Nasti Chestikov wrote:
> So on ZooEscape, where I already know the opponent dice rolls fail the chi square fit test, my own rolls (331,449) result in 8,766 occurences.
>
> This is 2.64% actual against 2.78% expected.
>
> Observations as to whether these figures fall within the norm?
>

Too busy hawking fast cars, sorry.

---
Tim Chow

On Thursday, May 13, 2021 at 6:17:15 PM UTC+1, Nasti Chestikov wrote:
> So on ZooEscape, where I already know the opponent dice rolls fail the chi square fit test, my own rolls (331,449) result in 8,766 occurences.
>
> This is 2.64% actual against 2.78% expected.
>
> Observations as to whether these figures fall within the norm?

No, these results aren't really "within the norm" because there are approx 5 standard deviations too few
66s and the sample is large.

The reason the finding isn't particularly impressive is that a random sample is almost certain to have
something unusual about it. Here, we have surprisingly few 66s.
But we could have had surprisingly many 66s. Or surprisingly few 11s. Or surprisingly many 11s.
Or surprisingly few doubles. Or surprisingly many doubles. Etc. etc. etc.
So the fact that you have found one particular surprise isn't in itself surprising or impressive or interesting.

Paul

On Saturday, 15 May 2021 at 09:40:38 UTC+1, peps...@gmail.com wrote:
> On Thursday, May 13, 2021 at 6:17:15 PM UTC+1, Nasti Chestikov wrote:
> > So on ZooEscape, where I already know the opponent dice rolls fail the chi square fit test, my own rolls (331,449) result in 8,766 occurences.
> >
> > This is 2.64% actual against 2.78% expected.
> >
> > Observations as to whether these figures fall within the norm?
> No, these results aren't really "within the norm" because there are approx 5 standard deviations too few
> 66s and the sample is large.
>
> The reason the finding isn't particularly impressive is that a random sample is almost certain to have
> something unusual about it. Here, we have surprisingly few 66s.
> But we could have had surprisingly many 66s. Or surprisingly few 11s. Or surprisingly many 11s.
> Or surprisingly few doubles. Or surprisingly many doubles. Etc. etc. etc.
> So the fact that you have found one particular surprise isn't in itself surprising or impressive or interesting.
>
> Paul

Thank you for your response - what does "5 standard deviations too few 66s mean"?

For reference, here are my complete rolls on that site:

Roll Occurences Rate Expected Difference
66 8,867 2.65% 2.78% -0.13%
65 18,891 5.64% 5.56% +0.08%
64 18,761 5.60% 5.56% +0.04%
63 18,908 5.65% 5.56% +0.09%
62 18,613 5.56% 5.56% -0.00%
61 18,843 5.63% 5.56% +0.07%
55 9,024 2.70% 2.78% -0.08%
54 18,552 5.54% 5.56% -0.02%
53 18,475 5.52% 5.56% -0.04%
52 18,548 5.54% 5.56% -0.02%
51 18,796 5.61% 5.56% +0.05%
44 9,017 2.69% 2.78% -0.09%
43 18,607 5.56% 5.56% -0.00%
42 18,771 5.61% 5.56% +0.05%
41 18,475 5.52% 5.56% -0.04%
33 9,044 2.70% 2.78% -0.08%
32 18,785 5.61% 5.56% +0.05%
31 18,829 5.62% 5.56% +0.06%
22 9,033 2.70% 2.78% -0.08%
21 18,818 5.62% 5.56% +0.06%
11 9,157 2.73% 2.78% -0.05%

It's only the double sixes that jump out at me as being obviously screwy.

On Friday, 14 May 2021 at 14:40:57 UTC+1, Tim Chow wrote:
> On 5/13/2021 1:17 PM, Nasti Chestikov wrote:
> > So on ZooEscape, where I already know the opponent dice rolls fail the chi square fit test, my own rolls (331,449) result in 8,766 occurences.
> >
> > This is 2.64% actual against 2.78% expected.
> >
> > Observations as to whether these figures fall within the norm?
> >
> Too busy hawking fast cars, sorry.
>
> ---
> Tim Chow

Yeah I always knew you'd show your true colours in the end.

Sorry, "colors" for you lot across the pond.

On 5/15/2021 12:48 PM, Nasti Chestikov wrote:
> On Friday, 14 May 2021 at 14:40:57 UTC+1, Tim Chow wrote:
>> On 5/13/2021 1:17 PM, Nasti Chestikov wrote:
>>> So on ZooEscape, where I already know the opponent dice rolls fail the chi square fit test, my own rolls (331,449) result in 8,766 occurences.
>>>
>>> This is 2.64% actual against 2.78% expected.
>>>
>>> Observations as to whether these figures fall within the norm?
>>>
>> Too busy hawking fast cars, sorry.
>>
>> ---
>> Tim Chow
>
> Yeah I always knew you'd show your true colours in the end.
>
> Sorry, "colors" for you lot across the pond.

Whoops, I misspoke. I meant to say that I was too busy
developing Eraser.

---
Tim Chow

On 5/15/2021 10:47 AM, Clare Richards wrote:
> For reference, here are my complete rolls on that site:
>
> Roll Occurences Rate Expected Difference
> 66 8,867 2.65% 2.78% -0.13%
> 65 18,891 5.64% 5.56% +0.08%
> 64 18,761 5.60% 5.56% +0.04%
> 63 18,908 5.65% 5.56% +0.09%
> 62 18,613 5.56% 5.56% -0.00%
> 61 18,843 5.63% 5.56% +0.07%
> 55 9,024 2.70% 2.78% -0.08%
> 54 18,552 5.54% 5.56% -0.02%
> 53 18,475 5.52% 5.56% -0.04%
> 52 18,548 5.54% 5.56% -0.02%
> 51 18,796 5.61% 5.56% +0.05%
> 44 9,017 2.69% 2.78% -0.09%
> 43 18,607 5.56% 5.56% -0.00%
> 42 18,771 5.61% 5.56% +0.05%
> 41 18,475 5.52% 5.56% -0.04%
> 33 9,044 2.70% 2.78% -0.08%
> 32 18,785 5.61% 5.56% +0.05%
> 31 18,829 5.62% 5.56% +0.06%
> 22 9,033 2.70% 2.78% -0.08%
> 21 18,818 5.62% 5.56% +0.06%
> 11 9,157 2.73% 2.78% -0.05%
>
> It's only the double sixes that jump out at me as being obviously screwy.

Why don't you ask Nasti to run a chi-squared test for you?

---
Tim Chow

On Sunday, May 16, 2021 at 3:22:09 AM UTC+1, Tim Chow wrote:
> On 5/15/2021 10:47 AM, Clare Richards wrote:
> > For reference, here are my complete rolls on that site:
> >
> > Roll Occurences Rate Expected Difference
> > 66 8,867 2.65% 2.78% -0.13%
> > 65 18,891 5.64% 5.56% +0.08%
> > 64 18,761 5.60% 5.56% +0.04%
> > 63 18,908 5.65% 5.56% +0.09%
> > 62 18,613 5.56% 5.56% -0.00%
> > 61 18,843 5.63% 5.56% +0.07%
> > 55 9,024 2.70% 2.78% -0.08%
> > 54 18,552 5.54% 5.56% -0.02%
> > 53 18,475 5.52% 5.56% -0.04%
> > 52 18,548 5.54% 5.56% -0.02%
> > 51 18,796 5.61% 5.56% +0.05%
> > 44 9,017 2.69% 2.78% -0.09%
> > 43 18,607 5.56% 5.56% -0.00%
> > 42 18,771 5.61% 5.56% +0.05%
> > 41 18,475 5.52% 5.56% -0.04%
> > 33 9,044 2.70% 2.78% -0.08%
> > 32 18,785 5.61% 5.56% +0.05%
> > 31 18,829 5.62% 5.56% +0.06%
> > 22 9,033 2.70% 2.78% -0.08%
> > 21 18,818 5.62% 5.56% +0.06%
> > 11 9,157 2.73% 2.78% -0.05%
> >
> > It's only the double sixes that jump out at me as being obviously screwy.
>
> Why don't you ask Nasti to run a chi-squared test for you?

Assuming the sample is indeed (supposed to be) random, there is ***very definitely*** a
too-few-doubles bug in the random number generation.
I would disagree that the problem has anything to do with 6 or 66.
One of the six doubles had to be the most infrequent and it happened to be 66.
The -0.13 discrepancy for 66 isn't wildly different to the -0.09% discrepancy for 44.

The standard deviation for the frequency of each doubles roll is approx
0.028%. So every single doubles roll is far too infrequent.
Even the most frequent double (11) occurs almost two standard deviations
more infrequently than expected.

I admit that it's usually hard to be statistically sure about data when you
don't post your prior hypotheses but the pattern is so glaring here that I'm quite
confident.

For example, if you saw me roll 100 double sixes in a row, you would be sure they
were biased dice without any particular need for a rigorous statistical test.

Paul

Timothy Chow wrote:
> On 5/15/2021 12:48 PM, Nasti Chestikov wrote:
>> On Friday, 14 May 2021 at 14:40:57 UTC+1, Tim Chow wrote:
>>> On 5/13/2021 1:17 PM, Nasti Chestikov wrote:
>>>> So on ZooEscape, where I already know the opponent dice rolls fail
>>>> the chi square fit test, my own rolls (331,449) result in 8,766
>>>> occurences.
>>>>
>>>> This is 2.64% actual against 2.78% expected.
>>>>
>>>> Observations as to whether these figures fall within the norm?
>>>>
>>> Too busy hawking fast cars, sorry.
>>>
>>> ---
>>> Tim Chow
>>
>> Yeah I always knew you'd show your true colours in the end.
>>
>> Sorry, "colors" for you lot across the pond.
>
> Whoops, I misspoke.  I meant to say that I was too busy
> developing Eraser.

Rubber?

>
> ---
> Tim Chow
>

--
Just as 'beautiful' points the way for aesthetics and 'good' for ethics,
so do words like 'true' for logic. All sciences have truth as their
goal; but logic is also concerned with it in a quite different way:
logic has much the same relation to truth as physics has to weight or
heat. Frege in 'Thoughts' (Der Gedanke)

On 5/16/2021 4:10 PM, Peter wrote:
> Timothy Chow wrote:
>> On 5/15/2021 12:48 PM, Nasti Chestikov wrote:
>>> On Friday, 14 May 2021 at 14:40:57 UTC+1, Tim Chow wrote:
>>>> On 5/13/2021 1:17 PM, Nasti Chestikov wrote:
>>>>> So on ZooEscape, where I already know the opponent dice rolls fail
>>>>> the chi square fit test, my own rolls (331,449) result in 8,766
>>>>> occurences.
>>>>>
>>>>> This is 2.64% actual against 2.78% expected.
>>>>>
>>>>> Observations as to whether these figures fall within the norm?
>>>>>
>>>> Too busy hawking fast cars, sorry.

>>>
>>> Yeah I always knew you'd show your true colours in the end.
>>>
>>> Sorry, "colors" for you lot across the pond.
>>
>> Whoops, I misspoke.  I meant to say that I was too busy
>> developing Eraser.
>
> Rubber?

Glue.

Timothy Chow wrote:

>Why don't you ask Nasti to run a chi-squared test for you?

Nasti Chestikov and Clare Richards appear to be the same person. Will
we see either ever again?

From: "Nasti Chestikov" <nasti.chestikov@gmail.com>
Subject: Double six analysis?
Date: Thu, 13 May 2021 10:17:14 -0700 (PDT)
Message-ID: <0ad32e15-6d8f-496b-b50d-5c58aaf98e29n@googlegroups.com>
Newsgroups: rec.games.backgammon
NNTP-Posting-Host: 82.38.158.34
User-Agent: G2/1.0

From: "Clare Richards" <clare.richards201066@gmail.com>
Subject: Re: Double six analysis?
Date: Sat, 15 May 2021 07:47:24 -0700 (PDT)
Message-ID: <52252baa-0ba5-4da8-a861-b704e767ac5cn@googlegroups.com>
Newsgroups: rec.games.backgammon
NNTP-Posting-Host: 82.38.158.34
User-Agent: G2/1.0

On Monday, May 17, 2021 at 12:01:27 PM UTC-7, badgolferman wrote:
> Timothy Chow wrote:
>
> >Why don't you ask Nasti to run a chi-squared test for you?
> Nasti Chestikov and Clare Richards appear to be the same person. Will
> we see either ever again?
>
>
> From: "Nasti Chestikov" <nasti.c...@gmail.com>
> Subject: Double six analysis?
> Date: Thu, 13 May 2021 10:17:14 -0700 (PDT)
> Message-ID: <0ad32e15-6d8f-496b...@googlegroups.com>
> Newsgroups: rec.games.backgammon
> NNTP-Posting-Host: 82.38.158.34
> User-Agent: G2/1.0
>
>
> From: "Clare Richards" <clare.rich...@gmail.com>
> Subject: Re: Double six analysis?
> Date: Sat, 15 May 2021 07:47:24 -0700 (PDT)
> Message-ID: <52252baa-0ba5-4da8...@googlegroups.com>
> Newsgroups: rec.games.backgammon
> NNTP-Posting-Host: 82.38.158.34
> User-Agent: G2/1.0

Aha! This explains why the user "clare201066" exists on ZooEscape, for Nasti Chestikov to juice his own rating. I can't seem to post a screenshot, but their head to head matchup is as follows:

W-L (Nasti-Clare)
210-3

Gammons
63-1

Backgammons
81-0 (including a 192-point doozy where the cube was raised to 64 and Nasti miraculously won!

See for yourself for those who have a ZooEscape account: https://zooescape.com/game-list.pl?v=200&p1=nastichestikov&p2=clare201066

The Clare user has played nowhere near the same number of games against any other opponent, and has not had nearly the astonishingly bad results as against our Nasti! https://zooescape.com/user-rivalry.pl?v=200&usr=clare201066&srt=13

On Monday, May 17, 2021 at 5:02:40 PM UTC-7, Beaver Plant wrote:
> On Monday, May 17, 2021 at 12:01:27 PM UTC-7, badgolferman wrote:
> > Timothy Chow wrote:
> >
> > >Why don't you ask Nasti to run a chi-squared test for you?
> > Nasti Chestikov and Clare Richards appear to be the same person. Will
> > we see either ever again?
> >
> >
> > From: "Nasti Chestikov" <nasti.c...@gmail.com>
> > Subject: Double six analysis?
> > Date: Thu, 13 May 2021 10:17:14 -0700 (PDT)
> > Message-ID: <0ad32e15-6d8f-496b...@googlegroups.com>
> > Newsgroups: rec.games.backgammon
> > NNTP-Posting-Host: 82.38.158.34
> > User-Agent: G2/1.0
> >
> >
> > From: "Clare Richards" <clare.rich...@gmail.com>
> > Subject: Re: Double six analysis?
> > Date: Sat, 15 May 2021 07:47:24 -0700 (PDT)
> > Message-ID: <52252baa-0ba5-4da8...@googlegroups.com>
> > Newsgroups: rec.games.backgammon
> > NNTP-Posting-Host: 82.38.158.34
> > User-Agent: G2/1.0
> Aha! This explains why the user "clare201066" exists on ZooEscape, for Nasti Chestikov to juice his own rating. I can't seem to post a screenshot, but their head to head matchup is as follows:
>
> W-L (Nasti-Clare)
> 210-3
>
> Gammons
> 63-1
>
> Backgammons
> 81-0 (including a 192-point doozy where the cube was raised to 64 and Nasti miraculously won!
>
> See for yourself for those who have a ZooEscape account: https://zooescape.com/game-list.pl?v=200&p1=nastichestikov&p2=clare201066
>
> The Clare user has played nowhere near the same number of games against any other opponent, and has not had nearly the astonishingly bad results as against our Nasti! https://zooescape.com/user-rivalry.pl?v=200&usr=clare201066&srt=13

Also worth noting that the reason "Clare" deleted that message is because "she" accidentally posted Nasti's dice stats! He was logged into the wrong account and, when his error was discovered, deleted the telltale post!

Compare https://zooescape.com/backgammon-stats.pl?usr=clare201066 to https://zooescape.com/backgammon-stats.pl?usr=nastichestikov