|   Statistics & Proof | Jon Risch
Apr 17, 2001 7:47 PM | | Recently, there are those who have tried to make claims concerning the absence of cable sonics, and hide behind the "burden of proof" argument so they could avoid providing any evidence for their claims.
Their reasoning is more than a bit circular.
It goes something like this:
"Others have made claims concerning cable sonics. There have been no gold-standard* positive results from controlled listening tests. Therefore, (within certain reasonable limits of similar resistance, etc), all cables sound the same.
* (Gold-standard scientific evidence is that which has been published in a peer-reviewed, professional journal. Technically, all else are anecdotes)
In order to justify this kind of conclusion WITHOUT citing their own results/tests/evidence, they cite the burden of proof argument, and also seem to imply that a lack of evidence IS evidence of a negative.
I believe that the first one is a matter of semantics as much as it is _when_ to invoke the 'burden of proof' for counter-claims to an original claim.
If there is no gold-standard evidence for cable sonics, then the correct way of stating this is to say that: "There is no widely accepted scientific evidence for the existence of cable sonics." This is quite different from saying "There ARE no cable sonics."
The first statement is based on factual and well accepted ways of looking at things. The second now invokes a need to provide evidence for what has now become a counter-claim, one that states a definitive viewpoint. If all we had was uncertainty, and this is what is stated, then the statement does not need it's own proof. Once we go into the realm of making definitive statements, now the burden of proof applies to this statement just as much as any other claim.
As for equating a lack of any widely accepted (gold standard) scientific evidence, to that of a certainty that the cables sonics do not exist, this is also a poor way of reasoning through the scientific methods that apply to Statistics and scientific reasoning.
This is sometimes stated as"Absence of evidence is not evidence of absence!" and has been discused in scientific circles elsewhere.
See:
http://www.tufts.edu/%7Egdallal/pval.htm
and
http://www.bmj.com/cgi/content/full/311/7003/485
for one example from medicine.
Then there is the issue of nulls being equated with negatives. Several here have argued that acceptance of the null hypothesis means that we equate it to a negative. I have argued that this is just not the case, and provided some references regarding this issue.
I now have some pretty definitive references that ANYONE reading this can check for themselves.
"A null hypothesis is not accepted just because it is not rejected."
See:
http://davidmlane.com/hyperstat/B33945.html
AND
"Null hypothesis are never accepted. We either reject them or fail to reject them."
See:
http://www.tufts.edu/%7Egdallal/sigtest.htm
These are both from highly regarded online Statistics References, the Hyperstat is often used to teach Statistics courses!
There are other statistical references online, as well as textbooks that can be perused, and they say pretty much the same things.
I have posted this kind of information before, see:
http://forums.consumerreview.com/crforum?13@@.ee7a593/31
I think that all these are pretty unequivocal, and should leave no doubt.
A null DOES NOT equal a negative.
You can whine, you can cry, you can cry foul, but these actions will not change the facts, and will not change the reality of how statistical anaylsis works according to scientific tenants.
Anecdotal Evidence
Now, some folks want to go outside the realm of the gold-standard for scientific evidence, they want to accept anecdotal evidence, such as posted at web sites (ABX website), or as told on the newsgroups, and count this in the ledger as failures to acheive positives.
However, rather one sidedly, they are willing to embrace the anecdotal null results |
| re: Statistics & Proof, Part 2 | Jon Risch
Apr 17, 2001 7:48 PM | | Part 2
However, rather one sidedly, they are willing to embrace the anecdotal null results, but will not consider the anecdotal positives at all.
Now does this seem a bit unfair and unscientific to anyone else?
The fact is, there are no gold-standard listening tests AGAINST cable sonics. None. Only what would be strictly speaking, anecdotal tests published in popular press magazines. Only three of these, and only for speaker cables.
I have been over this ground before, see:
http://www.audioreview.com/message/DCForumID6/2574.html#31
So the bottom line is, there is not reams and piles of listening tests with null results, not even any that would be considered as real scientific evidence. The fact is, you can not equate a null with a negative.
When people post that audio cables do not have any audible differences, they don't have a scientific leg to stand on. They refuse to supply corraborating references or evidence, or to even try and attempt to reason through the basis for their statements, despiote having made an actual definitive claim. Just belief in an outdated methodology called ABX, just faith in a few DBT high priests who have been mouthing the same tired old buzzphrases such as "20 years of ABX/DBT testing have shown that", or other such loaded and misleading phrases and 'sound bytes'.
No proof, no reasoning, nothing based on science or statistics or even opinions based on personal experiences in almost all cases.
Jon Risch |
| re: Statistics & Proof, Part 2 | Tom N.
Apr 18, 2001 6:59 AM | | Hi Jon,
I read your entire statement and have to ask...
what is your point?
Attempting to move the burden of proof from one claimant to another still does not resolve the core issues.
All you have done is said... You can't prove anything and neither can I.
Gee how helpful. Perhaps a little less wind in your sails would lead youto see that you could have said the same thing by just placing your tongue between your lips and blowing...
Tom N. |
| re: Statistics & Proof, Part 2 | Jon Risch
Apr 18, 2001 7:57 PM | | And YOUR contribution IS?
I just feel that it is not legitimate for naysayers to duck out over the issue of new claims that are made in the name of dismissing other claims.
Many of the hard core naysayers firmly believe that the anecdotal nulls equal hard core evidence for negatives. They post as if it is a forgone conclusion that there are no possible differences between cables, all the while refusing to provide ANY backup for these kinds of claims, other than the very illegitimate anecdotal nulls they happen to believe in.
In terms of my own contributions, I have performed many listening tests (just not published in peer-reviewed professional journals), made my refined and improved methods available publicly via an AES paper presentation (AES preprint #3178), and then made the results of my listening tests available in the form of DIY recipes for audio cables that perform as well as very expensive retail cables, for a fraction of the cost.
My simple DIY Belden 89259 IC has been compared sonically to retail cables costing hundreds, while the more advanced twisted pair recipe has been favorably compared to IC's costing thousands. These cost about $20 and $30 respectively to make, even less if more than just one pair is made at one time.
My Cross-Connected 89259 speaker cable recipe has also been compared favorably to some very expensive retail cables.
Cost of a ten foot pair is approx. $50-60, and they can easily be used in lengths up to 30-48 feet depending on the exact application before significant audible losses become a problem. They are very low in inductance, without the typical penalty of very high capacitance.
The digital IC recipes, using 89259, 1506A, and 1695A, have all been favorably compared to retail digital IC's costing hundreds to thousands, and they cost around $15-20 each to make.
I don't sell them, nor do I sell the raw materials, nor do I obtain any benefit or renumeration from the dissemination of this information.
By sharing the fruits of my personal research and listening tests, anyone who wants to find out if there is anything to this cable sonics business can do so for very little cash outlay, and in so many cases, they find that, HEY, these things DO make a difference, and they didn't have to pay mega-bucks to find out.
If you are curious, you can find link info on my DIY recipes at this post:
http://www.AudioAsylum.com/audio/cables/messages/19902.html
down toward the bottom.
Jon Risch |
| re: Statistics & Proof, Part 2 | Eyespy
Apr 18, 2001 8:22 PM | | I believe you meant "remuneration". Being renumerated is no picnic at all! Here's a few brackets for your trouble:
((((((((((((<<<<<<<{{{{{{[[[[||||||]]]]}}}}}}>>>>>>>))))))))))) |
| "...renumeration from the dissemination of this information". | Richard Greene
Apr 18, 2001 8:54 PM | | "Risch the Rapper" has obviously attended the
Al Sharpton School of Internet Posting!
Risch-man, I've got $2,000 cash in my hand that I'll give to you if you can hear any differences among the various materials used as
wire insulation in 10-foot speaker wires, under double-blind conditions, with volumes matched, while listening to music
in a normally reverberant listening room (hopefully that description will match your own stereo and listening room).
Since you have claimed (on the internet) to be able to do this under double-blind conditions and have even published rankings of various wire insulation materials based on sound quality, I am merely asking you to repeat a "test" you have already done. This may take less than one hour and I'll pay you $2,000 if you succeed.
If $2,000 is not enough for your time, make it $3,000.
Of course if you do not succeed, and you won't, I get the same amount
of money from you. I will donate all of the money I win from you (beyond the cost of airfare to your home) to my favorite charity,
The Salvation Army.
Put up some money.
Or shut up ... about 10 foot speaker wires.
They all should be assumed to sound the same unless at least ONE person in the world can hear a difference under blind conditions.
I'm still waiting for you to provide that ONE name.
All I hear is "silence" from you. |
| I'll think about taking that test | gonefishin
Apr 19, 2001 6:27 AM | | |
| to hear........ | gonefishin
Apr 19, 2001 6:37 AM | | .....silence from you ;)
enjoy the music
(damn you tinhere....now look what you have me doing!...hehe) |
| to hear....... | TinHere
Apr 21, 2001 6:33 PM | | "now look what you have me doing!"
Please assume responsiblity for your own statements and don't trivialize this board with statements you can't support. Am I your puppet master?
I try to keep my lack of knowledge a secret so if yer gonna be here don't ask me anything technical. [remember me if ya get the payday because I can't help but think ya owe me one] :) |
| "...renumeration from the dissemination of this information". | Bruce Burke
Apr 20, 2001 7:09 AM | | Actually, to do the test correctly, it should be conducted on multiple systems as well.
I've measured some materials that could be used as interconnect wires and found them to have no "sonics" of their own. (In other words, they carry the signal without altering it, other than really high capacitance cables actually improving the the distortion of high frequency signal(20kHz) because their cut-off frequency is low enough to lop off the some of the out of band harmonics(greater than 20kHz).)
If there are sonic differences, it may well be how the individual pieces of equipment react to the cables as part of the load they are driving, which is why I suggested the methodology of using multiple systems. If the cables truely do introduce their own "sonics" they will do so exactly the same on every system combination.
Further, there is work being done by others in the area of Power Amps that investigates how they respond to various real-world loads, which does include the speaker cables. Although the cables themselves do not have any sonic qualities, the way in which they alter the load the amplifier sees could affect how the amplifiers perform, which I have suspected all along. But the snake-oil salesmen will continue to peddle their over-priced product to the unsuspecting consumer.
Bottom line is, keep your vallues of R, L, and C at a reasonable minimum and you will get good results from your equipment. Beyond that, there is no magic bullet or wire formulation that will help.
Of course I expect to get flammed for this position.
-Bruce |
| "...renumeration from the dissemination of this information". | Dupper
Apr 25, 2001 8:53 PM | | You can hear these words typed here? I want proof that you can :> |
| re: Statistics & Proof, Part 2 | mtrycrafts
Apr 18, 2001 9:33 PM | | >>>In terms of my own contributions, I have performed many listening tests (just not published in peer-reviewed professional journals), <<<
You never will Jon, it will not stand up to scrutiny. That simple.
>>>), made my refined and improved methods available publicly via an AES paper presentation (AES preprint #3178), <<<<
A conference paper, Jon,not peer quality, and in 1991, when all the yeasayers presented. Must have been a lax reviewer. Besides, what does this demonstrate? That you presented a paper at an AES conference? Anything more?
>>>>My simple DIY Belden 89259 IC has been compared sonically to retail cables costing hundreds, <<<<
And the results Jon, can be found where? Who has duplicated the results? Reminds you of cold fussion yet, Jon?
>>>By sharing the fruits of my personal research and listening tests, <<<<
Nothing wrong with sharing but people must know a little more about those comparisons, how reliable are they? Has it been duplicated by anyone in a credible manner? So, all your effort may be for not if they are flawed. |
| Jon why do you waste your time trying to teach a new song | To the one note choir here...
Apr 20, 2001 9:13 AM | | either you hear the difference are you don`t....simple.... |
| Jon why do you waste your time trying to teach a new song | mtrycrafts
Apr 20, 2001 11:46 AM | | >>>either you hear the difference are you don`t....simple.... <<<
Yep, that is simple enough. What is a bit more difficult though is to demonstrate claims that you can and do hear them. Anyone can make empty claims.
If it is so simple, why does he has such a difficult time to demosntrate his claim to fame of 'golden ear' status? |
| Jon why do you waste your time trying to teach a new song | Dupper
Apr 21, 2001 8:30 AM | | I guess basing opinion on claims is no worse than basing opinion on flawed DBT testing. |
| Jon why do you waste your time trying to teach a new song | MajorDump
Apr 24, 2001 7:24 PM | | I am confused to why you think people have to demostrate claims to YOU. Who are you? Just another Joe Blow in the forums. Unless you think you are more than that? Maybe the Cable Savior? I have my own opinion on cables but thats not important in this post. What is important is what you come off as. |
| "What is important is what you come off as" | Eyespy
Apr 25, 2001 12:31 AM | | Why is that so important to you? |
| "What is important is what you come off as" | MajorDump
Apr 25, 2001 1:36 PM | | Oh, I wasn't saying it as it was important to me. It may be important to him that he is coming off as then know all and end all. If he wants people to think that when he is not, its not fooling me |
| re: Statistics & Proof | CE
Apr 18, 2001 7:56 PM | | WHAAAAAAAAAAAAAAAAAAAAAAAAAT? Do you have anything to say? |
| Jon, Jon, Jon, Jon, Jon, Jon, Jon... | GCM
Apr 20, 2001 10:12 AM | | "A null DOES NOT equal a negative"
I don't believe we're saying that. It's been stated over and over again that the purpose of hypothesis testing is to disprove the null. Untill that happens, "the null hypothesis is a statement that will be accepted if our sample data fail to provide us with convincing evidence that is is false". That quote being from Statistical Techniques in Business and Economics" Eighth Edition by Mason and Lind.
This is the point you conveniently overlook. Nobody is saying that there is proof that all cables are the same. We are saying there is no proof there isn't. Absent that proof, the null hypothesis is accepted. This is basic statistics. You can whine, you can cry, you can cry foul, but these actions will not change the facts, and will not change the reality of how statistical anaylsis works according to scientific tenants.
Let's set up a hypothesis test. Every month or so I get an email about how for a small fee, I'll be let in on the secret of how to make millions in the international monetary fund. This includes anecdotal quips from people who have supposedly performed this feat. So, we have the alternate hypothesis that states I can make millions every day converting international currency, the null being that I won't. There is no documented evidence anywhere that this works, and lots of documented complaints that it doesn't (null results). Armed with the knowledge that lots of people have had null results and nobody that anyone can name has made any money, what do you do with this email? Do you assume that just because the null hasn't been disproven, there is no reason to assume you won't make money? Or do you see a pattern in the repeated nulls and the failure to disprove the null, and hang on to your money?
I don't view cables differently. There is no conclusive evidence of audible differences. The null is not rejected, therefore following the advice from the text quoted above, I'm accepting it. |
| GCM, (repeat as needed) | Jon Risch
Apr 20, 2001 10:47 AM | | The reference you cite is clearly the opposite of the references I have posted. However, the title of the book gives us a clue that this is the expedient view of the null hypothesis, that is, business and management are often taught this way. It doe snot make it correct or more scientific, just that they can not allow for the open ended lack of results that failure to acheive a positive result creates.
Hence they accept the null AS IF IT WERE THE ACTUAL CASE, knowing that it really is not, but nothaving time or funds to make a better determination.
Since we are discussing mor of an absolute truth here, and the arguments are often aimed at exact and precise exposition of the basic tenants of statistics and scientific validity, the expedient view will not do.
One thig the expedient view depends on is the fact that the experiments or tersts MUST be considered totally above reproach, no possibility of flawed procedures, etc. The mere handful of popular press articles hardly meet these requirements.
[ So, we have the alternate hypothesis that states I can make
millions every day converting international currency, the null being that I won't. There is no documented evidence anywhere that this works, and lots of documented complaints that it doesn't (null results). Armed with the knowledge that lots of people have had null results and nobody that anyone can name has made any money, what do you do with this email? ]
So you are saying that there is LOTS of evidence, null results, for cables? Where? By whom? What professional journal? Even what popular press magazine?
I would love to see this list of all of this massive amount of evidence.
The fact is, there is NOT a lot of null evidence against audio cable sonics. Not ANY goden standard type evidence, and very very little of even anecdotal evidence.
That's the rub, there just isn't anything solid to base the POV that cables have no sonic character, because there ARE NOT a lot of null results to cite or to use as evidence. mnay of the anecdotes (ABX web site, Arny's posts, etc.) have obvious and fatal flaws in them, they just aren't very good tests at all, very weak and unsubtantiated.
So when you have that list of ALL that 'evidence', I will be waiting to see it.
Jon Risch |
| GCM, (repeat as needed) | mtrycrafts
Apr 20, 2001 11:53 AM | | >>>The reference you cite is clearly the opposite of the references I have posted. <<<<<
Maybe, maybe not. Are you just conveniently picking one sentence out of your text?
Have you read:
"Accepting the Null Hypothesis," by Robert Frick, Memory and Cognition, 1995, 23(1), 132-138? Perhaps you should. |
| GCM, (repeat as needed) | Jon Risch
Apr 21, 2001 7:35 AM | | GCM,
>>The reference you cite is clearly the opposite of the references
I have posted. <<<<<
[ Maybe, maybe not. Are you just conveniently picking one sentence out of your text? ]
That 'one sentence' seems pretty darn specific and unequivocal, and it is repeated in varying forms in more than just one of the the references I provide.
[ Have you read: "Accepting the Null Hypothesis," by Robert Frick, Memory and Cognition, 1995, 23(1), 132-138? ]
Have you?
This seems to be an isolated instance of ONE person going against the established and accepted scientific practices for statistical analysis, and is certainly NOT the mainstream view of statisticians.
I have spoken with college professors in the field of statistics, more than one, and NONE of them disagree with what I have stated or been posting:
"Null hypothesis are never accepted." "
"A null hypothesis is not accepted just because it is not rejected."
I have provided approx. 4 text references from mainstream statistics books, some of them used for teaching statistics courses, and 4 solid web references )and there are undoubtedly more, I just have not taken the time to find more. How many does it take), all of them very definite and unequivocal in how they state the issue. ALL of these back up what I have been saying. If my position was the one that was off the beaten path, I doubt that there would be so many clearly worded instances of my POV.
Since I have seen the one book you referenced above (not the Frick book), it is even possible that it contains a direct statement along those lines too, and the quote you posted was taken out of context.
It is also very likely that the Frick book discussers the issue of expediancy, and if one IS going to take the step of accepting the null as if it were actually true, then I am sure that he would spell out that to accept test/experiment results in that manner can ONLY be done when one has some assurance that the test/eperimental methods and procuedures were performed according to accepted scientific practice, and that the tests/experiments demonstrated a defined level of sensitivity and competence, by showing they could acheive a certain minimum level of detection, etc.
Accepted scientific practice does NOT mean that only one aspect of such testing was addressed. Naysayers love to point to blind (or double blind) test conditions, as if they were ALL that was required to provide a valid test. This is ludicrous, as there are so many OTHER aspects that also must be addressed, any one of which would disqualify the tests as being considered valid evidence.
Jon Risch |
| GCM, (repeat as needed) | GCM
Apr 23, 2001 9:01 AM | | "So you are saying that there is LOTS of evidence, null results, for cables? Where? By whom? What professional journal? Even what popular press magazine? "
Where did I say that?
I'm using is as an argument against your stance that nulls don't count as evidence. |
| You never answered me by the way.. | GCM
Apr 23, 2001 12:21 PM | | You didn't answer my question. Instead you made up something I didn't say, then attacked me based upon what you made up. So, I'll ask my question about the currency exchange scheme again.
Do you assume that just because the null hasn't been disproven, there is no reason to assume you won't make money? Or do you see a pattern in the repeated nulls and the failure to disprove the null, and hang on to your money? |
| You never answered me by the way.. | Markw*
Apr 23, 2001 1:15 PM | | A quote from one of his responses to me:
"According to science, if I perform a dozen different trials, and they came up with nulls, and then perform one more, and it comes up with statistically significant positive results, and then can verify this with another positive result, not necessarily immediately or sequentially, all the nulls are overturned, and considered a moot point."
Question 1 ) What is a "statistically significant positive result"?
Question 2) If he preforms dozens of tests and gets null (I can't hear a difference) results and then preforms one test and gets his "statistically significant positive result", all previous nulls are thrown out. Now, he again preforms his tests UNTIL he gets another "statistically significant positive result", no matter how many null (I can't hear a difference but it doesn't count anyway) results pop up in between, it still counts as a positive?
That's better house odds than the casinos have here in Joisey. Perhaps he works offshore in Biloxi?
That's like saying we'll flip a coin to settle a bet. I call heads. If it doesn't come up heads, we'll keep flipping until it does. See, I win! Told ya so! |
| You never answered me by the way.. | Jon Risch
Apr 23, 2001 6:15 PM | | [ Question 1 ) What is a "statistically significant positive result"? ]
Most of the naysayers state it is immutably the 95% confidence level. As far as I can tell, this is not based on anything in particular, as it is not written in stone that one MUST use the 95% value for statistical significance.
This level equates to:
12 correct forced choices out of 16, the classic ABX criteria. It happens that 16 trials is a a lot to do, and hard to maintain focus and concentration for that long.
OR
9 out of 10
7 out of 8
5 out of 5
Others feel that a 90% confidence level is fine for less than life-and-death, innocent or guilty circumstances, where one wishes to actually get some useful data from such difficult and tiring tests.
This equates to approx.:
10 out of 14
7 out of 9
5 out of 6 (close enough)
I prefer to use the 7 out of 9 criteria, as that keeps the number of trials down to avoid fatigue, and is still a good solid 90% or better confidence level.
BTW, one is supposed to select the confidence level BEFORE the experiment, to avoid fitting the data results to match the statistical outcome. However, it is considered legitimate to use the data analysis to garner a further understanding of the results, and reaching higher confidence levels is usualy considered a more solid result.
[ Question 2) If he preforms dozens of tests and gets null (I can't hear a difference) results and then preforms one test and gets his "statistically significant positive result", all previous nulls are thrown out. Now, he again preforms his tests UNTIL he gets another "statistically significant positive result", no matter how many null (I can't hear a difference but it doesn't count anyway) results pop up in between, it still counts as a positive? ]
Yes, this IS how statistical analysis of listening test results work. The nulls are a complete lack of data or results, the positives are considered as results. Lots of nothing vs. something.
It technically DOES NOT MATTER how many times a null might be generated from any given test, or from any given series or runs of tests, the positive results are overriding, and it technically only takes one.
Now I would agree, based on the principle of scientific verification, before any result could be considered valid, that there would have to be more than just one single positive result before we could say that there was a valid positive for sure. After all, there is a possibility that due to 'random chance' or 'guessing', that one could achieve what seem to be positive results.
Tying this into the confidence level selected, lets say 90%, then, ON AVERAGE, in ten separate tests, in is theoretically possible that one would come up as positive, even if the listener was actually guessing or tossing coins. So in order to beat these 'odds', the positives would have to either: come up more often than every tenth time, AND/OR the results would have to consistently exceed the 90% confidence levels.
Interestingly enough, when the listeneres admit they are guessing, and don't have a clue, the socres are often below the 50% correct ratio, that is, for 9 trials, they might only get 3 or 4 'correct'. Seldom do truly random guesses come out right at the 50% ratio of correct answers to incorrect ones, especially with the smaller numbers of trials. This is just an observation over the years. Again, according to theory, the results should be 'random' and at a 50% ratio if the subject were actually guessing, but this is a whole other subject, and is even more convoluted than the basics that we are discussing, and con not agree on.
As I have stated, my positive results for any given test parameter have come closer to having statistically significant positive results about half the time for completed tests, rather than only every tenth time. I did experimnent with some 83% confidence levels early on, but quickly switched to 90%, so as to keep the results at a decent level of c |
| REPLY, Part 2 | Jon Risch
Apr 23, 2001 6:17 PM | | PART 2
As I have stated, my positive results for any given test parameter have come closer to having statistically
significant positive results about half the time for completed tests, rather than only every tenth time. I did
experimnent with some 83% confidence levels early on, but quickly switched to 90%, so as to keep the
results at a decent level of confidence, and have been able to reach those levels on a regular basis.
I fail to see how or WHY one would be required to acheive positive results almost EVERY time, as this is
just not the way such difficult subjective tests go. It is not that uncommon for testing that is at the limits of
what can be reliably detected using that test methodology, to have rather more spurious seeming results
than positives every time.
Since my methods are a refinement of the classic ABX/DBT forced choice techniques, they tend to do
better than when such classic methods are used, and even so, I still have to struggle to keep the test
sensitivity high, and the listeners focussed and concentrating.
Even Penicillin will not reach a 100% track record, and we KNOW it works!
I think that ANY forced choice test is NOT similar in nature to how we listen to music for pleasure, and
therefore, the listening tests that use this method are not going to exhibit the most sensitivity, and will not
be capable of providing positives for some real phenomenon that are truly audible but yet are subtle in
nature, due to the inherent problems and limitations of such testing.
Perhaps the problem lies in how poorly understood statistics is, and the common example of having an
'honest' coin come up heads three times in a row has most people thinking that the odds are in favor of it
coming up tails on the fourth toss, while statistics and theory says the odds are STILL 50/50. Many people
have a problem with this, and just do not comprehend what is the basis for such analysis and how the
odds are arrived at.
Jon Risch |
| Still waiting for an answer | GCM
Apr 24, 2001 5:00 AM | | and waiting, and waiting, and waiting... |
| Still waiting for an answer | Jon Risch
Apr 24, 2001 7:37 PM | | I assume you are refering to this silly scenario:
[ Let's set up a hypothesis test. Every month or so I get an email about how for a small fee, I'll be let in on the secret of how to make millions in the international monetary fund. This includes anecdotal quips from people who have supposedly performed this feat. So, we have the alternate hypothesis that states I can make millions every day
converting international currency, the null being that I won't. There is no documented evidence anywhere that this works, and lots of documented complaints that it doesn't (null results). Armed with the knowledge that lots of people have had null results and nobody that anyone can name has made any money, what do you do with this
email? Do you assume that just because the null hasn't been disproven, there is no reason to assume you won't make money? Or do you see a pattern in the repeated nulls and the failure to disprove the null, and hang on to your money? ]
Aside form the loaded and biased connotations of your example, there is the one big (and usual) flaw in the analogy:
there are not "lots of people have had null results" for audio cable listening tests.
There are only three very questionable popular press articles, none of them related or consistent with one another, none of which represent the latest or most refined methodolgy, and all were just for speaker cables.
How many even less distinguished web anecdotes are there? The ABX web site? AND?
I asked mtry for a list, just exactly how many anecdotal nulls are there? Where is this long list of nulls? SHOW ME THE NULLS!
Fact is, there just aren't even that many anecdotal nulls, and if we are keeping score, as some seem to think appropriate, then I would venture that the anecdotal nulls outweigh the nulls by a wide margin.
So, for your example, even with the loaded connotations, lets take a look:
there are many claims that the scheme makes money.
Just a handful of claims that they made no money.
Perhaps one COULD make money with the scheme! He he he!
Of course, by using such an example, you have loaded the discussion with an obvious impossibliity, one of making money by making change (Hey, we make up for it in sheer volume. That's a JOKE, BTW).
It is not a certainty that audio cables do not have a sonic character, this has not been proven, and there are no 'objective' proofs that will stand up to close scrutiny either, so the certainty that you seem to have is a false one, based on your take on things, your POV. This simply does not mean that your POV MUST be correct, just because you hold it.
I hope that others can see through the hype to the root of your problematic attempts to force an argument that is a tilt-a-whirl booby trap of an analogy, one without any real merit, and used as a straw man to throw red herrings onto the mound of naysayer BS obscuring the view to the real truth of the matter.
Yeah. ;-)
Jon Risch |
| Jeeze Jon, stick to the subject already. | GCM
Apr 25, 2001 5:10 AM | | I'm NOT TALKING ABOUT CABLES any more. You claim that you can't form any opinions based on failure to disprove the null. I'm giving an UNRELATED example, that will hopefully illustrate my point. So again, you still won't answer. |
| Jeeze Jon, stick to the subject already. | Jon Rsich
Apr 25, 2001 10:52 AM | | I am really sorry that your reading comprehension and ability to reason are so impared. Tough break.
I see through the straw man you have set up, and for those who can read and comprehend, I have answered your question. You just don't like the answer.
Jon Risch |
| Jeeze Jon, stick to the subject already. | GCM
Apr 25, 2001 1:01 PM | | I asked you a simple question to make a point. I believe my example illustrates why conclusions can be drawn from failure to disprove a null hypothes. If you disagree, say why using only the example I provided. You shouldn't need to make up things I didn't say and refute them. As to your assessment of my reading comprehension, clearly you're an intelligent chap. It's a shame to see you resort to insults. |
| Jeeze Jon, stick to the subject already. | Jon Risch
Apr 25, 2001 5:12 PM | | [ I asked you a simple question to make a point. ]
Now you insult my intelligence. Again.
[ I believe my example illustrates why conclusions can be
drawn from failure to disprove a null hypothes. ]
Of course it would, since you designed to be a one-sided example. This dose not illustrate the issues which were under discussion, as I so inconveniently pointed out.
[ If you disagree, say why using only the example I provided. ]
I did, without allowing you your tautology. Perhaps that is why you are in denial.
Jon Risch |
| Jeeze Jon, stick to the subject already. | GCM
Apr 26, 2001 5:16 AM | | You refuse to answer a simple question that supports my point and I'm in denial? That's a good one. I give up now. |
| Jon, Jon, Jon, Jon, Jon, Jon, Jon... | Dupper
Apr 27, 2001 10:31 PM | | So, we have the alternate hypothesis that states I can make millions every day converting international currency, the null being that I won't. There is no documented evidence anywhere that this works, and lots of documented complaints that it doesn't (null results).
Sorry GCM, contradiction to what you say:
"A null DOES NOT equal a negative"
Bad example. |
| re: Please feel free to correct me if I'm wrong.... | Monstrous Mike
Apr 23, 2001 2:12 PM | | You seem to be saying that a Null does not prove anything in particular about the sonics of cables. It does not equal a negative. It is not conclusive. Well I have to agree with you on that one.
You also say there is no gold-plated evidence proving either sonics in cables or the absence of sonics in cables.
I have to therefore conclude that you are basing your belief on your own anecdotal evidence and the anecdotal evidence of others who have claimed to have heard sonics in cables in controlled listening tests. I presume this to also mean that any anecdotal evidence in controlled tests which show no cable sonics must have had some flaw in the procedure or the equipment or whatever. Otherwise, how could you accept such a result?
I have yet to experience hearing cable differences, other than improper or damaged cables. I have yet to see evidence that there are sonics in cables. I work with many scientific people and have yet to come across one who can explain how sonics are possible in cables.
Pick the proper cable for the job, and anything more is a waste of time and money. That's my opinion, and I'm stickin' to it.
If you think my position is based on very thin evidence, then your contrary position must be based on equally thin evidence. I think that's the main point you brought out in your post. |
| Try This On for Size | Jon Risch
Apr 23, 2001 7:12 PM | | Say that you were the one who had intially heard something, that there seemed to be sonic differences, and that after performing some blind auditions, you were STILL able to detect which cable was which. Then, take that even further, so as to be SURE for yourself, you did some research, did some homework, found out what the necessary conditions were to conduct a listening test that was to the highest standards, and as good as the then current state of such tests were, and STILL came up with positives when everyhting was done right, by the book, to the letter.
Now, Mike, we are all going to ignore your results, call you names, say that what you experienced DOES NOT MATTER, just becuase you did not publish your efforts. Hey, some of us even want you to deny your own experience, and 'admit' that it was a delusion!
Think about it. You KNOW. All the furor, all the clamor, all the BS in the world would not be enough to make you deny the evidence of your own undertaking and results, would they?
Yes, strictly speaking, my experiences are anecdotes to OTHERS, but for me, I _KNOW_ the truth of the matter, and that I am not making any of this up, and that it was real and that I was not the only one who could discern this either.
[ If you are proven wrong, then what will be your excuse for hearing something that wasn't there? ]
As I said, think about it. Put yourself in my shoes, and think about what YOU would do to assure that the tests you poerformed were as right and as correct as you could do. COuld anyone prove you wrong, when you beat the odds, reached 99.9% confidence levels in some tests?
There is no way to prove me wrong, to me. I have the test results, I have seen others do it, it is not some sort of meta-physical guess on my part, it is solid science and test results. Again, despite any number of null results, they still do not constitute a proof of anything, so no one on this earth can 'prove' me wrong.
Jon Risch |
| Try This On for Size | Eyespy
Apr 24, 2001 7:48 AM | | Yes, so you say, in one breath. But in the next one, you refuse to let an identified third party witness your testing procedure, on your system, in your listening environment, with your own music, oh, and did I say, using your very own testing procedure? Then you complain and whine when your claims are not taken on blind faith by others. I fail to see a valid basis to your complaints or confusion as to why your claimed results are not embraced by some. But, faith-based claims can behave in that manner when subject to third party scrutiny. |
| Try This On for Size | Jon Risch
Apr 24, 2001 5:03 PM | | Your supposition is exposed.
I have had third party verification, just not the third party of YOUR choice. I will not subject these other third parties to the kind of mean-spirited, demeaning, rude and ill-conceived onslaught of naysayer zealots just to satisfy a whim.
I am not asking for blind faith. I just refuse to disallow my own personal experiences, and deny that they ever occured.
Jon Risch |
| Try This On for Size | Eyespy
Apr 24, 2001 8:53 PM | | I don't expect Norm to behave or treat you in the manner you described (mean-spirited, demeaning, rude and ill-conceived onslaught...). Do you?
For him to do so would seem very out of character, and would invalidate any failure on your part to produce a positive. Hence, he has no incentive to treat you that way during your testing procedures. I'm open to the possibility you could succeed; I'm actually rooting for you. I'd love to be able to change my stance on cable sonics. |
| Try This On for Size | Jon Rsich
Apr 25, 2001 10:56 AM | | I am not specifically refering to Norm, there have been others who have behaved this way, including the OTHER two 'offers'.
Jon Risch |
| re: Please feel free to correct me if I'm wrong.... | DulceBuzz
Apr 23, 2001 7:30 PM | | I think Mike made some good points here. I have a simpler way of putting it. Until someone comes up with certified test results, we should all make our decisions on what we think is best for ourselves. Bashing and being sarcastic about the subject because it doesn't agree with your POV, is, well childish. |
| re: Please feel free to correct me if I'm wrong....Part 2 | Monstrous Mike
Apr 23, 2001 2:15 PM | | The more I think about it, the more I pity your position. You have claimed to have heard sonic differences and I have claimed the opposite.
If I am proven wrong, then I wasn't listening hard enough or my equipment was inferior, or I had wax in my ears.
If you are proven wrong, then what will be your excuse for hearing something that wasn't there? |
| How about.... | Eyespy
Apr 24, 2001 7:53 AM | | ...he was listening TOO hard, his equipment was inferior, or the sun was in his eyes? |
| |
