Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > sci.physics > #603887 > unrolled thread

Simple Probability Question

Back to article view | Back to sci.physics

Contents

  Simple Probability Question Bob <rgsros@notme.invalid> - 2016-11-06 08:34 -0500
    Re: Simple Probability Question Poutnik <poutnik4nntp@gmail.com> - 2016-11-06 14:50 +0100
      Re: Simple Probability Question Poutnik <poutnik4nntp@gmail.com> - 2016-11-06 15:06 +0100
      Re: Simple Probability Question Wally W. <ww84wa@aim.com> - 2016-11-06 10:37 -0500
        Re: Simple Probability Question Poutnik <poutnik4nntp@gmail.com> - 2016-11-06 16:47 +0100
    Re: Simple Probability Question Helmut Wabnig <hwabnig@.- --- -.dotat> - 2016-11-06 17:44 +0100
    Re: Simple Probability Question Yuri Kreaton <invlaid@invalid.com> - 2016-11-06 12:50 -0600
      Re: Simple Probability Question Poutnik <poutnik4nntp@gmail.com> - 2016-11-06 20:04 +0100
        Re: Simple Probability Question Yuri Kreaton <invlaid@invalid.com> - 2016-11-06 14:51 -0600
          Re: Simple Probability Question Poutnik <poutnik4nntp@gmail.com> - 2016-11-06 22:24 +0100
    Re: Simple Probability Question Double-A <double-a3@hush.com> - 2016-11-06 13:24 -0800
      Re: Simple Probability Question "reber g=emc^2" <herbertglazier0@gmail.com> - 2016-11-07 13:08 -0800
    Re: Simple Probability Question Odd Bodkin <bodkinodd@gmail.com> - 2016-11-07 09:08 -0600
      Re: Simple Probability Question jimp@specsol.spam.sux.com - 2016-11-07 18:42 +0000
      Re: Simple Probability Question Yuri Kreaton <invlaid@invalid.com> - 2016-11-07 15:44 -0600
        Re: Simple Probability Question jimp@specsol.spam.sux.com - 2016-11-07 21:54 +0000
          Re: Simple Probability Question benj <benj@nobody.net> - 2016-11-07 18:20 -0500
            Re: Simple Probability Question Yuri Kreaton <invlaid@invalid.com> - 2016-11-08 12:13 -0600

#603887 — Simple Probability Question

Hello,

Could someone explain this for me.
Know it's a simple question, but Probability has always bothered me.
You can't learn without asking, so let me ask, please:

Let's say I have a few thousand coins.
They are 100 % identical, in Every respect.

I also have a mechanical gadget that flips them
in absolutely the identical way for each coin.
Same spin, momentum, height, etc.

Each coins gets and experiences the exact same input.
100 % identical.

So, will I see the typical 50% Heads, 50% Tails split ?
Why ?

If not (?), for the split to occur, must there must be "some"
incremental difference between them ?

Or,...?

Thanks,
Bob

[toc] | [next] | [standalone]

#603888

Dne 06/11/2016 v 14:34 Bob napsal(a):
> Hello,
> 
> Could someone explain this for me.
> Know it's a simple question, but Probability has always bothered me.
> You can't learn without asking, so let me ask, please:
> 
> Let's say I have a few thousand coins.
> They are 100 % identical, in Every respect.
> 
> I also have a mechanical gadget that flips them
> in absolutely the identical way for each coin.
> Same spin, momentum, height, etc.
> 
> Each coins gets and experiences the exact same input.
> 100 % identical.
> 
> So, will I see the typical 50% Heads, 50% Tails split ?
> Why ?
> 
> If not (?), for the split to occur, must there must be "some"
> incremental difference between them ?
> 

The 50/50 head/tail applies for events independent on each other.

But your scenario is almost the opposite.
From previous head/tailyou can predict next head/tail.
There would be near 100% heads or tails.

I say near, because there will be
some residual random influences you have not eliminated.

-- 
Poutnik ( The Pilgrim, Der Wanderer )
Knowledge makes great men humble, but small men arrogant.

[toc] | [prev] | [next] | [standalone]

#603890

Dne 06/11/2016 v 14:50 Poutnik napsal(a):

> The 50/50 head/tail applies for events independent on each other.
> 
> But your scenario is almost the opposite.
> From previous head/tailyou can predict next head/tail.
> There would be near 100% heads or tails.
> 
> I say near, because there will be
> some residual random influences you have not eliminated.
> 
P.S.> Real result will be somewhere on scale
from 50/50 for fully random events

   ( this is true as limit case for infinite number for trials,
   otherwise it would follow binomical statistical distribution
   with 50/50 as centre.)

till 100/0 for fully deterministic events.

-- 
Poutnik ( The Pilgrim, Der Wanderer )
Knowledge makes great men humble, but small men arrogant.

[toc] | [prev] | [next] | [standalone]

#603894

On Sun, 6 Nov 2016 14:50:51 +0100, Poutnik wrote:

>Dne 06/11/2016 v 14:34 Bob napsal(a):
>> Hello,
>> 
>> Could someone explain this for me.
>> Know it's a simple question, but Probability has always bothered me.
>> You can't learn without asking, so let me ask, please:
>> 
>> Let's say I have a few thousand coins.
>> They are 100 % identical, in Every respect.
>> 
>> I also have a mechanical gadget that flips them
>> in absolutely the identical way for each coin.
>> Same spin, momentum, height, etc.
>> 
>> Each coins gets and experiences the exact same input.
>> 100 % identical.
>> 
>> So, will I see the typical 50% Heads, 50% Tails split ?
>> Why ?
>> 
>> If not (?), for the split to occur, must there must be "some"
>> incremental difference between them ?
>> 
>
>The 50/50 head/tail applies for events independent on each other.
>
>But your scenario is almost the opposite.
>From previous head/tailyou can predict next head/tail.
>There would be near 100% heads or tails.
>
>I say near, because there will be
>some residual random influences you have not eliminated.

And non-random influences.

The gravitational force of the Moon on the coin (as it is flipped)
would vary in a predictable way throughout the experiment.

If the system could be defined well enough, the influence of the
Moon's gravity on the coin's trajectory would not be random.

[toc] | [prev] | [next] | [standalone]

#603895

Dne 06/11/2016 v 16:37 Wally W. napsal(a):

> And non-random influences.
> 
> The gravitational force of the Moon on the coin (as it is flipped)
> would vary in a predictable way throughout the experiment.
> 
> If the system could be defined well enough, the influence of the
> Moon's gravity on the coin's trajectory would not be random.
> 

Sure, non-random biases as well.

Moon and Sun position and distance,
temperature, humidity, local air convection,
heterogenity of impact surface....

But, if rate of their change is low enough
wrt rate of repeating the event,
they can be often be considered
as constants within experiment timescope.


-- 
Poutnik ( The Pilgrim, Der Wanderer )
Knowledge makes great men humble, but small men arrogant.

[toc] | [prev] | [next] | [standalone]

#603898

On Sun, 6 Nov 2016 08:34:04 -0500, Bob <rgsros@notme.invalid> wrote:

>Hello,
>
>Could someone explain this for me.
>Know it's a simple question, but Probability has always bothered me.
>You can't learn without asking, so let me ask, please:
>
>Let's say I have a few thousand coins.
>They are 100 % identical, in Every respect.
>
>I also have a mechanical gadget that flips them
>in absolutely the identical way for each coin.
>Same spin, momentum, height, etc.
>
>Each coins gets and experiences the exact same input.
>100 % identical.
>
>So, will I see the typical 50% Heads, 50% Tails split ?
>Why ?
>
>If not (?), for the split to occur, must there must be "some"
>incremental difference between them ?
>
>Or,...?
>
>Thanks,
>Bob

Same stupid line of thought as appears here:

On ancient pottery the fixed stars
have exact the same configuration as on today's night sky.
This proves "fixed stars" don't move

Can you find the flaw?
If yes, then you should be able to solve your "problem".

w.

[toc] | [prev] | [next] | [standalone]

#603902

On 11/6/2016 7:34 AM, Bob wrote:
> Hello,
>
> Could someone explain this for me.
> Know it's a simple question, but Probability has always bothered me.
> You can't learn without asking, so let me ask, please:
>
> Let's say I have a few thousand coins.
> They are 100 % identical, in Every respect.

how thick are they ?

there is a small probability they can land on their edge.


>
> I also have a mechanical gadget that flips them
> in absolutely the identical way for each coin.
> Same spin, momentum, height, etc.
>
> Each coins gets and experiences the exact same input.
> 100 % identical.

just make it a thought experiment.

>
> So, will I see the typical 50% Heads, 50% Tails split ?
> Why ?

you assume your coins have 2 sides only.  you assume your flippper 
introduces no biases, nor atmospheric pressure, nor...  => goto thought 
experiment.  Random states, two values, what is probability of either ?

how thick is the coin?  if thickness > coin diameter it lands on the 
edge a lot.

and you won't see exactly 50/50,  it will be off from that.

>
> If not (?), for the split to occur, must there must be "some"
> incremental difference between them ?

nope.

>
> Or,...?
>
> Thanks,
> Bob

[toc] | [prev] | [next] | [standalone]

#603903

Dne 06/11/2016 v 19:50 Yuri Kreaton napsal(a):
> 
> and you won't see exactly 50/50,  it will be off from that.
> 
Exactly 50/50 is a limit case for N -> infinity

E.g. for 4 coin trials with 50/50 probability
the probability there are

2 heads and 2 tails    is   3/8.
1 or 3 heads are       each 1/4,
all heads or all tails each 1/16.

-- 
Poutnik ( The Pilgrim, Der Wanderer )
Knowledge makes great men humble, but small men arrogant.

[toc] | [prev] | [next] | [standalone]

#603912

On 11/6/2016 1:04 PM, Poutnik wrote:
> Dne 06/11/2016 v 19:50 Yuri Kreaton napsal(a):
>>
>> and you won't see exactly 50/50,  it will be off from that.
>>
> Exactly 50/50 is a limit case for N -> infinity
>

if edge thicnkess is 0

but the OP, he was using machines, and machines ware out with time,
so he must have ment a finite time

or a trial, which dosen't have to go to 50/50

> E.g. for 4 coin trials with 50/50 probability
> the probability there are
>
> 2 heads and 2 tails    is   3/8.
> 1 or 3 heads are       each 1/4,
> all heads or all tails each 1/16.
>

[toc] | [prev] | [next] | [standalone]

#603915

Dne 06/11/2016 v 21:51 Yuri Kreaton napsal(a):
> On 11/6/2016 1:04 PM, Poutnik wrote:
>> Dne 06/11/2016 v 19:50 Yuri Kreaton napsal(a):
>>>
>>> and you won't see exactly 50/50,  it will be off from that.
>>>
>> Exactly 50/50 is a limit case for N -> infinity
>>
> 
> if edge thicnkess is 0
> 
> but the OP, he was using machines, and machines ware out with time,
> so he must have ment a finite time
> 
> or a trial, which dosen't have to go to 50/50
> 

Sure, for 2 state scenario.

My last note was for the random scenario,
as you mentioned "you won't see exactly 50/50",
not for his case.


-- 
Poutnik ( The Pilgrim, Der Wanderer )
Knowledge makes great men humble, but small men arrogant.

[toc] | [prev] | [next] | [standalone]

#603916

On Sunday, November 6, 2016 at 5:34:04 AM UTC-8, Bob wrote:
> Hello,
> 
> Could someone explain this for me.
> Know it's a simple question, but Probability has always bothered me.
> You can't learn without asking, so let me ask, please:
> 
> Let's say I have a few thousand coins.
> They are 100 % identical, in Every respect.
> 
> I also have a mechanical gadget that flips them
> in absolutely the identical way for each coin.
> Same spin, momentum, height, etc.
> 
> Each coins gets and experiences the exact same input.
> 100 % identical.
> 
> So, will I see the typical 50% Heads, 50% Tails split ?
> Why ?
> 
> If not (?), for the split to occur, must there must be "some"
> incremental difference between them ?
> 
> Or,...?
> 
> Thanks,
> Bob


Depends on whether the wind kicks up.

Double-A

[toc] | [prev] | [next] | [standalone]

#604026

On Sunday, November 6, 2016 at 1:24:51 PM UTC-8, Double-A wrote:
> On Sunday, November 6, 2016 at 5:34:04 AM UTC-8, Bob wrote:
> > Hello,
> > 
> > Could someone explain this for me.
> > Know it's a simple question, but Probability has always bothered me.
> > You can't learn without asking, so let me ask, please:
> > 
> > Let's say I have a few thousand coins.
> > They are 100 % identical, in Every respect.
> > 
> > I also have a mechanical gadget that flips them
> > in absolutely the identical way for each coin.
> > Same spin, momentum, height, etc.
> > 
> > Each coins gets and experiences the exact same input.
> > 100 % identical.
> > 
> > So, will I see the typical 50% Heads, 50% Tails split ?
> > Why ?
> > 
> > If not (?), for the split to occur, must there must be "some"
> > incremental difference between them ?
> > 
> > Or,...?
> > 
> > Thanks,
> > Bob
> 
> 
> Depends on whether the wind kicks up.
> 
> Double-A

Depends on how heavy its platted with nickel Trebert

[toc] | [prev] | [next] | [standalone]

#603977

On 11/6/2016 7:34 AM, Bob wrote:
> Each coins gets and experiences the exact same input.
> 100 % identical.

Can't do that. Violates laws of physics.

-- 
Odd Bodkin --- maker of fine toys, tools, tables

[toc] | [prev] | [next] | [standalone]

#603995

Odd Bodkin <bodkinodd@gmail.com> wrote:
> On 11/6/2016 7:34 AM, Bob wrote:
>> Each coins gets and experiences the exact same input.
>> 100 % identical.
> 
> Can't do that. Violates laws of physics.

It is fairly trivial to do down to the limits of measuring capability.


-- 
Jim Pennino

[toc] | [prev] | [next] | [standalone]

#604035

On 11/7/2016 9:08 AM, Odd Bodkin wrote:
> On 11/6/2016 7:34 AM, Bob wrote:
>> Each coins gets and experiences the exact same input.
>> 100 % identical.
>
> Can't do that. Violates laws of physics.
>

yup.  they cannot weigh the same, if real coins, there is at least a 5% 
variance.


I weighed 50 aspirins once, 325mg, however each one did not weigh 325mg, 
they were all off, but within +-30mg.


ALSO, if it is a coin, it has been stamped on both sides, with different 
impressions, therefore the center of gravity is NOT exactly in the 
middle of the coin, creating a bias.

[toc] | [prev] | [next] | [standalone]

#604041

Yuri Kreaton <invlaid@invalid.com> wrote:
> On 11/7/2016 9:08 AM, Odd Bodkin wrote:
>> On 11/6/2016 7:34 AM, Bob wrote:
>>> Each coins gets and experiences the exact same input.
>>> 100 % identical.
>>
>> Can't do that. Violates laws of physics.
>>
> 
> yup.  they cannot weigh the same, if real coins, there is at least a 5% 
> variance.
> 
> 
> I weighed 50 aspirins once, 325mg, however each one did not weigh 325mg, 
> they were all off, but within +-30mg.
> 
> 
> ALSO, if it is a coin, it has been stamped on both sides, with different 
> impressions, therefore the center of gravity is NOT exactly in the 
> middle of the coin, creating a bias.

You people need to decide whether you're talking about real coins out
of pocket or 'special' experimental, controlled coins made for the
purpose of the experiment.

FYI the front to back resistration on real coin is pretty consistant,
so any bias should be the same for a given lot of real coins.
 

-- 
Jim Pennino

[toc] | [prev] | [next] | [standalone]

#604048

On 11/7/2016 4:54 PM, jimp@specsol.spam.sux.com wrote:
> Yuri Kreaton <invlaid@invalid.com> wrote:
>> On 11/7/2016 9:08 AM, Odd Bodkin wrote:
>>> On 11/6/2016 7:34 AM, Bob wrote:
>>>> Each coins gets and experiences the exact same input.
>>>> 100 % identical.
>>>
>>> Can't do that. Violates laws of physics.
>>>
>>
>> yup.  they cannot weigh the same, if real coins, there is at least a 5%
>> variance.
>>
>>
>> I weighed 50 aspirins once, 325mg, however each one did not weigh 325mg,
>> they were all off, but within +-30mg.
>>
>>
>> ALSO, if it is a coin, it has been stamped on both sides, with different
>> impressions, therefore the center of gravity is NOT exactly in the
>> middle of the coin, creating a bias.
>
> You people need to decide whether you're talking about real coins out
> of pocket or 'special' experimental, controlled coins made for the
> purpose of the experiment.
>
> FYI the front to back resistration on real coin is pretty consistant,
> so any bias should be the same for a given lot of real coins.
>
I have heard that American coins have a very slight bias toward heads 
but it takes an enormous number of flips for it to show up.

This paper says coins then to come up with a bias toward how they were 
when flipped. http://statweb.stanford.edu/~susan/papers/headswithJ.pdf

[toc] | [prev] | [next] | [standalone]

#604136

On 11/7/2016 5:20 PM, benj wrote:
> On 11/7/2016 4:54 PM, jimp@specsol.spam.sux.com wrote:
>> Yuri Kreaton <invlaid@invalid.com> wrote:
>>> On 11/7/2016 9:08 AM, Odd Bodkin wrote:
>>>> On 11/6/2016 7:34 AM, Bob wrote:
>>>>> Each coins gets and experiences the exact same input.
>>>>> 100 % identical.
>>>>
>>>> Can't do that. Violates laws of physics.
>>>>
>>>
>>> yup.  they cannot weigh the same, if real coins, there is at least a 5%
>>> variance.
>>>
>>>
>>> I weighed 50 aspirins once, 325mg, however each one did not weigh 325mg,
>>> they were all off, but within +-30mg.
>>>

>>> ALSO, if it is a coin, it has been stamped on both sides, with different
>>> impressions, therefore the center of gravity is NOT exactly in the
>>> middle of the coin, creating a bias.
>>
>> You people need to decide whether you're talking about real coins out
>> of pocket or 'special' experimental, controlled coins made for the
>> purpose of the experiment.
>>
>> FYI the front to back resistration on real coin is pretty consistant,
>> so any bias should be the same for a given lot of real coins.
>>
> I have heard that American coins have a very slight bias toward heads
> but it takes an enormous number of flips for it to show up.
>
> This paper says coins then to come up with a bias toward how they were
> when flipped. http://statweb.stanford.edu/~susan/papers/headswithJ.pdf
>

interesting paper, but if the tossing machine was perfect, it would flip 
each coin identically each time, and land the same way all the time....

[toc] | [prev] | [standalone]

Back to top | Article view | sci.physics

csiph-web