"Paul Miller" wrote in message
om...

Well, I would suggest that you specify the conditions of your
experiment that would predict the largest deviations from SR.

That wasn't the agreement.

As far as I know, we don't have an agreement yet. I thought we were
trying to work towards one. Like I said, that may take time.

Before one can come to total agreement, one must come to preliminary steps.
I'm still awaiting your first contribution.

In terms
of accuracy, as I'm offering 100-1 I would need accuracy at least one
order of magnitude larger, i.e. p 0.001 where p is prob. of getting
a wrongly falsifying result. I know I promised to get back to you with
more detail, but I really will.

I presume a repeatable measurement of a value of the variation outside
the current 'best' value will be sufficient? That's all I'm asking for.
Give me a specific value of 'c' and a range of values around that main
value. Within that range you (and SR) will be declared the winner (and
you will receive $1,000 US from me). Outside of which, I will be
declared the winner (and receive $100,000 US from you). This is a
trivial excercise for you and your "pal."

I will then design an experiment (and you and your "pal" will approve)
such that the experimental precision will be tight enough to meet your
"challenge."

As I understand it, we are trying to distinguish between two possible
worlds, one in which SR is correct, and one in which it is wrong.

Actually, I'm only concerned with whether SR is correct (or not) in this
world. I'm not interested in other possible worlds.

The
best way to distinguish between them is to find the conditions under
which they look as different as possible from each other.

I don't agree. I'm simply looking for a single experimental deviation from
basic SR assumptions (in this case, the second postulate).

Setting the
range of acceptable values of c depends on many factors, such as: the
value of c that you would predict that differs from SR, the nature of
experimental noise etc. Without investigating these issues, it's not
'a trivial exercise' for us.

Certainly it is. Because the level of noise required cannot be determined
until you identify what you consider the current "SR" value of c.

I did say in advance that it would take
some time to sort this out, did't I?

You explicitly stated that it would take no significant time after you
returned from vacation on Sept. 10.

Given that I'm risking a
considerable amount of money, is it unreasonable for me to ask for
some discussion about these issues? Why don't you try telling me your
thoughts on this, rather than always trying to pick a fight?

Because it's the only way to get you to respond, it seems.

In the meantime, please show me some
reason to think that you have the resources to actually carry this
experiment out.

What a pathetic attempt at evasion. If I had no resources, you would
not have to risk any money. And how can I provide you with an
estimate of the resources required, when you won't provide even the
minimum to design the experiment? (Aside from the fact a few locals
agreed to assist with equipment scrounging and purchase.)

What I'm talking about here is our time, not money. If this experiment
can't actually be done for real, then we're wasting our time. How
about you telling me about a couple of designs that could be done? At
least that would give me a start.

I can't do that until you tell me what your pass/fail criterion is. You
know, the one you promised me shortly after Sept. 10th. Because the
pass/fail criterion will drive the specifics of the experiment. I've
already given you the general experimental setup.

{snip stuff on Paul's knowledge}

I know nothing about GR, and
there's nothing in that post to suggest otherwise. My comments in that
post come from a basic understanding of the philosophy of science,
which is a common element in a cog.sci. degree. I'm not sure why I'm
now having to convince you that I'm telling the truth about this.
Firstly, I'm not sure if I care - if you want to think I'm a physicist
with an invisible friend, that's fine by me. Secondly, I don't
understand the relevance of this to the bet.

You are correct, I care not whether your friend is real, is Tom Roberts, or
is Harvey the rabbit. I just enjoy jerking the chain of people who act like
trolls, and have to be chivvied to live up to their commitments. So, I'll
drop the above issue.

After all, you want your money, right? Try again when you can simply
live up to your newsgroup claims. Like with a pass/fail criterion.

Setting a pass/fail criterion depends on lots of factors, such as
experimental noise.

No, the pass/fail criterion comes first. You will also get the opportunity
to discuss experimental noise (beyond the basic criterion) at a later date.

That's one reason why I was asking whether you
have the resources to do this - it would be good to see how accurate
your measurements are able to be.

We won't know that until after you tell us how precise you want it.

Not on paper, but for real. Setting
up experiments like this isn't that straightforward -

That's why you have to tell us what we are shooting at.

maybe we should
try to work together to define the criterion, rather than assuming
that I'm lying about everything.

I don't see a lie, yet. Merely an unwillingness to provide a starting
point.

I would suggest that would involve
you talking to me about a few possible designs,

I already gave you the design. Then you agreed to provide a pass/fail
criterion, so that we can design the equipment.

then we can think
about what predictions SR would make, and what you would predict. Then
we can think about what kind of accuracy we can expect. We might even
have to collect some data before we set the criteria.

No way, Jose! You don't get to modify the criteria after you see whether SR
is losing!

This is how real
world experiments are done, at least in my experience.

I thought you weren't a physicist. No labs, no math.

If you go back to my original, first post,

You mean this one?
http://groups.google.com/groups?selm...ffde%40posting.
google.com

I did say that it was *your* job to
completely specify the experiment, in detail. The idea was that I
would take that experimental design - including criteria for
falsification - and talk to my pal about that.

Nothing in that post about criteria for falsification. At best, the above is
called disingeneous. Because in your later posts you committed to providing
something:
http://www.google.com/groups?selm=46...0posting .goo
gle.com

greywolf42:
"Now. How about that win/lose criterion? Has Tom Roberts / your anonymous
friend had time to consider this very basic and simple question?"

Paul Miller:
"I'll get back to you on this - I'm away until 13th Sept, but I'm sure it
won't take long to specify the details once I return.

Now I can see that it's going to take a bit more effort on our part,
which is going to make demands on our time. That's not a problem,
as long as your willing to work with us a bit. If you would rather
assume that I'm lying about all of that for some inexpicable reason,
then walk away from the challenge.

Oh, I'm not walking away. This is too good an opportunity to miss. For both
of us.

So, how about that promised criterion? It need not be final, but we need a
talking point.

--
greywolf42
ubi dubium ibi libertas
{remove planet for e-mail}

(Paul Miller) wrote in message . com...
It's nice to be friends with yourself.

sorry?

Paul claimed that he was " ... not a physicist, nor a mathematician, nor
do I claim any understanding of relativity." The only problem is that
Paul was somewhat unwilling to commit himself to any substantive
negotiations in that challenge.

Well, I would suggest that you specify the conditions of your
experiment that would predict the largest deviations from SR.

That wasn't the agreement.

As far as I know, we don't have an agreement yet. I thought we were
trying to work towards one. Like I said, that may take time.

In terms
of accuracy, as I'm offering 100-1 I would need accuracy at least one
order of magnitude larger, i.e. p 0.001 where p is prob. of getting
a wrongly falsifying result. I know I promised to get back to you with
more detail, but I really will.

Sure, you will.

I presume a repeatable measurement of a value of the variation outside the
current 'best' value will be sufficient? That's all I'm asking for. Give
me a specific value of 'c' and a range of values around that main value.
Within that range you (and SR) will be declared the winner (and you will
receive $1,000 US from me). Outside of which, I will be declared the winner
(and receive $100,000 US from you). This is a trivial excercise for you and
your "pal."

I will then design an experiment (and you and your "pal" will approve) such
that the experimental precision will be tight enough to meet your
"challenge."

As I understand it, we are trying to distinguish between two possible
worlds, one in which SR is correct, and one in which it is wrong. The
best way to distinguish between them is to find the conditions under
which they look as different as possible from each other. Setting the
range of acceptable values of c depends on many factors, such as: the
value of c that you would predict that differs from SR, the nature of
experimental noise etc. Without investigating these issues, it's not
'a trivial exercise' for us. I did say in advance that it would take
some time to sort this out, did't I? Given that I'm risking a
considerable amount of money, is it unreasonable for me to ask for
some discussion about these issues? Why don't you try telling me your
thoughts on this, rather than always trying to pick a fight?

In the meantime, please show me some
reason to think that you have the resources to actually carry this
experiment out.

What a pathetic attempt at evasion. If I had no resources, you would not
have to risk any money. And how can I provide you with an estimate of the
resources required, when you won't provide even the minimum to design the
experiment? (Aside from the fact a few locals agreed to assist with
equipment scrounging and purchase.)

What I'm talking about here is our time, not money. If this experiment
can't actually be done for real, then we're wasting our time. How
about you telling me about a couple of designs that could be done? At
least that would give me a start.


"Paul Miller" did briefly resurface on September 10th:

http://www.google.com/groups?selm=46...0posting .goo
gle.com

However, this post certainly does not sound like it is written by
someone who " ... is not a physicist, nor a mathematician, nor
... claim(s) any understanding of relativity."
===========================
"All you are saying here is that you could build an adaptive controller
to make the corrections without having any knowledge of GR. So what?
The claim 'GPS wouldn't work without GR' means no more than 'the
controller would land up learning exactly the same corrections as
those predicted by GR'. This puts GR in the same category as any other
well-supported and well-accepted theory. You could train a neural
network to model the predictions of any theory, past or future, so you
'empirical' statement applies to all theories. It tells us nothing
about the status of those theories."
===========================

Sorry, but wrong again.

Wrong that you posted something? You claimed that you did not have "any
understanding of relativity." Yet you posted an authoritative post on GR
and the support for physics theories in general. That appears to contradict
your prior claims.

Wrong about these rather odd 'conspiracy theory'-like comments. As I
said from the start, I'm not a physicist, but my friend is. My
background is cognitive science, which is a branch of psychology, so
it's typically taught without any maths. I know nothing about GR, and
there's nothing in that post to suggest otherwise. My comments in that
post come from a basic understanding of the philosophy of science,
which is a common element in a cog.sci. degree. I'm not sure why I'm
now having to convince you that I'm telling the truth about this.
Firstly, I'm not sure if I care - if you want to think I'm a physicist
with an invisible friend, that's fine by me. Secondly, I don't
understand the relevance of this to the bet.

If you were "real", you would simply provide what you agreed to before.
After all, you want your money, right? Try again when you can simply live up
to your newsgroup claims. Like with a pass/fail criterion.

Setting a pass/fail criterion depends on lots of factors, such as
experimental noise. That's one reason why I was asking whether you
have the resources to do this - it would be good to see how accurate
your measurements are able to be. Not on paper, but for real. Setting
up experiments like this isn't that straightforward - maybe we should
try to work together to define the criterion, rather than assuming
that I'm lying about everything. I would suggest that would involve
you talking to me about a few possible designs, then we can think
about what predictions SR would make, and what you would predict. Then
we can think about what kind of accuracy we can expect. We might even
have to collect some data before we set the criteria. This is how real
world experiments are done, at least in my experience. If you go back
to my original, first post, I did say that it was *your* job to
completely specify the experiment, in detail. The idea was that I
would take that experimental design - including criteria for
falsification - and talk to my pal about that. Now I can see that it's
going to take a bit more effort on our part, which is going to make
demands on our time. That's not a problem, as long as your willing to
work with us a bit. If you would rather assume that I'm lying about
all of that for some inexpicable reason, then walk away from the
challenge.

Paul

xxein: You do not act cognitively.

Paul Miller:
[...]
your measurements are able to be. Not on paper, but for real. Setting
up experiments like this isn't that straightforward - maybe we should
try to work together to define the criterion, rather than assuming
that I'm lying about everything. I would suggest that would involve
you talking to me about a few possible designs, then we can think
about what predictions SR would make, and what you would predict. Then
we can think about what kind of accuracy we can expect. We might even
have to collect some data before we set the criteria. This is how real
world experiments are done, at least in my experience.

Correct. Obtaining the actual data is generally a minute fraction of
the time and effort involved in designing, building and insuring the
apparatus is capable of the desired resolution. In a well conceived
experiment, questions of reproducing settings and any other potential
sources of systematic optimism are identified and quantified before making
measurements in order minimize the interaction of the experimentors with
the experiment while taking real data. A year or more of development
and a data run of a few days isn't unusual. The length of time needed
to take actual data is as short as the desired statistics allow.

[[[[Posted to Paul Miller on 10/25/04 to "What Happened to Paul
Miller"
thread of relativity Newsgroup --- he had made a bet that no one could
experimentally beat SR]]]]
++++++++++++++ ++++++++++ ++++++++++++ ++++++++++++ +++++++

Quick Little Hint to "Paul Miller," et. al.

All that is needed to experimentally disprove SR is one half of
Einstein's train experiment.

The experiment goes as follows:

Take only the right-hand light beam; as Einstein said, it started
equidistant from each observer (per their own personal rulers in
their own personal frames). We can label this distance L. Also as
Einstein implied, the light beam reached the two observers absolutely
differently. (This must be an absolute difference because Einstein
used it to derive his relativity of simultaneity, and no mere
relative difference would be acceptable for that.) (Besides, it
is absolute because all observers in all frames see that the beam
reached the two observers differently; i.e., they see that one of
the observers is *not* there when the beam reaches the other.) We
can label these absolutely different beam arrival times Ta and Tb.
Thus, according to Einstein's own train example, which was used to
derive SR, we have the following simple calculations for light's
one-way speed in each of the two given frames:

Train Frame: Light's one-way speed = L/Ta

Embankment Frame: Light's one-way speed = L/Tb

This simple experiment is performed many times daily whenever
two people walk at different speeds toward an approaching light
ray (e.g., a ray from the sun). (Not to mention the fact that
Einstein used it to derive both the relativity of simultaneity
and SR, so you cannot find a more legit experiment in the context
of SR.)

And in case you missed it, the experiment directly disproves
Einstein's basis for SR, aka his light postulate aka the 2nd
postulate. That is, the experiment proves that light's one-way
speed, contrary to SR, varies directly with frame velocity.

And the experiment is so simple that even a cognitive scientist
type of guy can (or should be able to) comprehend it.

Of course, I don't really believe that anyone out there is
really putting up any $$$, so my above is just for the record.

On Sun, 24 Oct 2004 22:37:54 +0000, greywolf42 wrote:

"Paul Miller" wrote in message
om...

[bunches of words mostly going nowhere]

Just out of curiosity, have you guys considered taking the radical step of
exchanging email addresses, so you can negotiate a little more directly
than via Usenet posts?

Or even, heaven forbid, exchanged phone numbers, for some real
high-bandwidth communication?

I realize letting someone know your email address or (egad) your phone
number requires a certain amount of trust, but compared to the trust
required to make actually proceeding with a bet on this scale a sensible
move, it seems pretty minor.

The show is somewhat entertaining, albeit a tad slow-going, but if there's
to be any more to it than that you'll both eventually have to break down
and let the other party know for sure whether you're really a person, a
sock puppet, an Internet ghost, or what.


--
I can be contacted through http://www.physicsinsights.org

On Fri, 22 Oct 2004 21:01:22 +0000, greywolf42 wrote:

"Paul Miller" wrote in message
om...
Hey greywolf,

sorry, but the day job has taken over for a while, both for me and my
pal.

It's nice to be friends with yourself.

Paul claimed that he was " ... not a physicist, nor a mathematician,
nor do I claim any understanding of relativity." The only problem is
that Paul was somewhat unwilling to commit himself to any substantive
negotiations in that challenge.

Well, I would suggest that you specify the conditions of your experiment
that would predict the largest deviations from SR.

That wasn't the agreement.

In terms
of accuracy, as I'm offering 100-1 I would need accuracy at least one
order of magnitude larger, i.e. p 0.001 where p is prob. of getting a
wrongly falsifying result. I know I promised to get back to you with
more detail, but I really will.

Sure, you will.

Watch your step, Wolfie. He just tossed in an extra "" up there. He
says an order of magnitude better than the odds or no go, which means
p0.001, but that "" is a tad stronger than "" and what's worse it's
not well defined.

[ ... ]

If you must know, my background is in cognitive science, not physics or
maths. Just as I said first time round.

A "scientist" without a math background? :-)

Guess again.

If he's for real, and he's got a cognitive psych degree that's worth
anything, he's sure to know probabilities. (How do you think they decide
in advance how many rats will be needed in the experiment in order to get
a publishable result?) And probability is what gambling's all about.

Ask him a few questions about Fisher's exact test, one and two tailed T
tests, why chi^2 would or wouldn't be useful in deciding who won, or
anything else you can think of about statistics. You may find he's got you
seriously outgunned in exactly the area that's important for deciding who
won.

If a judge is needed to decide the outcome, and if Paul's really a shrink,
you may have more trouble than you expect holding onto your $1000.


--
I can be contacted through http://www.physicsinsights.org

"sal" wrote in message
news : On Fri, 22 Oct 2004 21:01:22 +0000, greywolf42 wrote:
:
: "Paul Miller" wrote in message
: om...
: Hey greywolf,
:
: sorry, but the day job has taken over for a while, both for me
and my
: pal.
:
: It's nice to be friends with yourself.
:
: Paul claimed that he was " ... not a physicist, nor a
mathematician,
: nor do I claim any understanding of relativity." The only
problem is
: that Paul was somewhat unwilling to commit himself to any
substantive
: negotiations in that challenge.
:
: Well, I would suggest that you specify the conditions of your
experiment
: that would predict the largest deviations from SR.
:
: That wasn't the agreement.
:
: In terms
: of accuracy, as I'm offering 100-1 I would need accuracy at least
one
: order of magnitude larger, i.e. p 0.001 where p is prob. of
getting a
: wrongly falsifying result. I know I promised to get back to you
with
: more detail, but I really will.
:
: Sure, you will.
:
: Watch your step, Wolfie. He just tossed in an extra "" up there.
He
: says an order of magnitude better than the odds or no go, which
means
: p0.001, but that "" is a tad stronger than "" and what's worse
it's
: not well defined.
:
: [ ... ]
:
: If you must know, my background is in cognitive science, not
physics or
: maths. Just as I said first time round.
:
: A "scientist" without a math background? :-)
:
: Guess again.
:
: If he's for real, and he's got a cognitive psych degree that's worth
: anything, he's sure to know probabilities. (How do you think they
decide
: in advance how many rats will be needed in the experiment in order
to get
: a publishable result?) And probability is what gambling's all
about.
:
: Ask him a few questions about Fisher's exact test, one and two
tailed T
: tests, why chi^2 would or wouldn't be useful in deciding who won, or
: anything else you can think of about statistics. You may find he's
got you
: seriously outgunned in exactly the area that's important for
deciding who
: won.
:
: If a judge is needed to decide the outcome, and if Paul's really a
shrink,
: you may have more trouble than you expect holding onto your $1000.
:
Heck, I said put your money up front as you would in a casino, before
the cards are dealt or the roulette wheel was spun. He said I didn't
know what I was talking about. Simple enough litmus test, wasn't it?
Miller comes up troll.
Androcles.


: --
: I can be contacted through http://www.physicsinsights.org
:

"sal" wrote in message news
On Fri, 22 Oct 2004 21:01:22 +0000, greywolf42 wrote:

"Paul Miller" wrote in message
om...
Hey greywolf,

sorry, but the day job has taken over for a while, both for me and my
pal.

It's nice to be friends with yourself.

Paul claimed that he was " ... not a physicist, nor a mathematician,
nor do I claim any understanding of relativity." The only problem is
that Paul was somewhat unwilling to commit himself to any substantive
negotiations in that challenge.

Well, I would suggest that you specify the conditions of your experiment
that would predict the largest deviations from SR.

That wasn't the agreement.

In terms
of accuracy, as I'm offering 100-1 I would need accuracy at least one
order of magnitude larger, i.e. p 0.001 where p is prob. of getting a
wrongly falsifying result. I know I promised to get back to you with
more detail, but I really will.

Sure, you will.

Watch your step, Wolfie. He just tossed in an extra "" up there. He
says an order of magnitude better than the odds or no go, which means
p0.001, but that "" is a tad stronger than "" and what's worse it's
not well defined.

[ ... ]

If you must know, my background is in cognitive science, not physics or
maths. Just as I said first time round.

A "scientist" without a math background? :-)

Guess again.

If he's for real, and he's got a cognitive psych degree that's worth
anything, he's sure to know probabilities. (How do you think they decide
in advance how many rats will be needed in the experiment in order to get
a publishable result?) And probability is what gambling's all about.

I have seen statistics courses given to psychology students.
They are worse than a joke.


Ask him a few questions about Fisher's exact test, one and two tailed T
tests, why chi^2 would or wouldn't be useful in deciding who won, or
anything else you can think of about statistics. You may find he's got you
seriously outgunned in exactly the area that's important for deciding who
won.

Mingst doesn't even know the standard definition of a "null result".
Google for
greywolf42 "null result" miller


If a judge is needed to decide the outcome, and if Paul's really a shrink,
you may have more trouble than you expect holding onto your $1000.

Shrinks are clinical psychiatrists or psychologists (See Webster),
not cognitive psychologists.

Dirk Vdm

In terms
of accuracy, as I'm offering 100-1 I would need accuracy at least one
order of magnitude larger, i.e. p 0.001 where p is prob. of getting a
wrongly falsifying result. I know I promised to get back to you with
more detail, but I really will.

Sure, you will.

Watch your step, Wolfie. He just tossed in an extra "" up there. He
says an order of magnitude better than the odds or no go, which means
p0.001, but that "" is a tad stronger than "" and what's worse it's
not well defined.

That's a fair point, but when I originally talked about p values, I
was trying to move the discussion away from setting measurement
criteria, which would be difficult for me to do without knowing
something about the statistics of our measurements. As before, the
criteria is up for negotiation - I'm not insisting anything at this
stage. At the moment, I have very little feel for what kind of
accuracy we can hope for - maybe p0.001 is not achievable, but I
can't judge that until I get a better view of the experiment.

If he's for real, and he's got a cognitive psych degree that's worth
anything, he's sure to know probabilities. (How do you think they decide
in advance how many rats will be needed in the experiment in order to get
a publishable result?) And probability is what gambling's all about.

Ask him a few questions about Fisher's exact test, one and two tailed T
tests, why chi^2 would or wouldn't be useful in deciding who won, or
anything else you can think of about statistics. You may find he's got you
seriously outgunned in exactly the area that's important for deciding who
won.

Absolutely right - I am allowed to have *some* expertise, even if it's
not in SR!


Paul

"sal" wrote in message
news
On Sun, 24 Oct 2004 22:37:54 +0000, greywolf42 wrote:

"Paul Miller" wrote in message
om...

[bunches of words mostly going nowhere]

Just out of curiosity, have you guys considered taking the radical step of
exchanging email addresses, so you can negotiate a little more directly
than via Usenet posts? Or even, heaven forbid, exchanged phone
numbers, for some real high-bandwidth communication?

Sure. But e-mail takes just as much effort as the newsgroup. And since the
challenge was made on the newsgroup, the negotiations may be of interest to
other contenders.

I realize letting someone know your email address or (egad) your phone
number requires a certain amount of trust, but compared to the trust
required to make actually proceeding with a bet on this scale a sensible
move, it seems pretty minor.

The trust won't be needed until the bet is made. Hence, I'm fine for the
moment.

The show is somewhat entertaining, albeit a tad slow-going, but if there's
to be any more to it than that you'll both eventually have to break down
and let the other party know for sure whether you're really a person, a
sock puppet, an Internet ghost, or what.

Frankly, I don't care if Paul Miller is real. So long as the money is
deposited in a mutually-agreed-upon control. Whenever and if ever this
occurs.

--
greywolf42
ubi dubium ibi libertas
{remove planet for e-mail}