The HSCA Acoustical Evidence: Proof of a Second Gunman in the JFK Assassination

[Michael T. Griffith]

It's pretty obvious that at some point in time the bug professor became a fanatic CT advocate and fanboy Griffith is now parroting anything and everything published by Dr. Thomas.

The longer sales pitch contained the same dishonest claim of matching "all five" ignoring, and not even mentioning, the so-called false positives (no, I'm not going to transcribe the BS he's proffering) at around 25:00 onward.

Unfortunately 'parroting' doesn't mean 'understanding' so there will be no explanation of the 1 in 100.000 claim by fanboy Griffith, but it's useful to check Appendix C in the Barger report which explains why even detecting possible gunshots is a statistical exercise in itself: we're looking for spikes in a sea of noise, which BTW is assumed to be random, which we don't actually know.

The quality of the tape was so bad they had to set the threshold at .6 so their 'detector' was more like a 'dredger'.

As for the process in regard to time taken and the amount of patterns checked raised by Joe Elliott, I'll bet they initially spent days dialing in the threshold (lowering it progressively) as they realized how bad that tape was.

Your comments are comically wrong, once again. You still have not answered my question about why you choose the Decker crosstalk as your time indicator over the Fisher crosstalk, the two Curry transmissions made in Dealey Plaza, and the two dispatcher time notations.

Now, of course, we both know that you cling to the Decker crosstalk because otherwise you would have to explain the amazing and intricate correlations between the dictabelt's impulse patterns and the patterns of at least four of the Dealey Plaza test-firing shots.

You'd have to explain why five of the dictabelt impulse patterns match five of the test-firing shots in the correct locational order and in the correct average rate of speed. The odds that five impulses would match in the correct locational order alone are 1 in 125, less than 1%.

You'd have to explain the N-wave and muzzle-blast patterns and the fact that they occur in the correct order and interval in the dictabelt impulse patterns. 

You'd have to explain why windshield distortion occurs when it should and does not occur when it should not in the dictabelt impulse patterns.

You'd have to explain the WA sonar analysis of the grassy knoll impulse pattern, which removed the false-match problem, reduced the acceptance window from 6 milliseconds to 1 millisecond, and reduced the microphone spacing from 18 feet to 1 foot.

You know you can't explain these things (the NRC panel could not explain them either), so you cling to the Decker crosstalk and use it as your excuse for not dealing with the actual acoustical evidence.

You and Joe Elliott will always find some excuse, no matter how lame, to avoid admitting that the acoustical evidence is valid and solid. Just look at Elliott's last sad, labored, and evasive reply about Dr. Thomas's finding that the actual odds that chance caused the grassy knoll shot are 100,000 to 1 against. You guys will continue to deny the acoustical evidence because you don't want to believe it, because if you did, you would have to abandon your entire theory of the assassination.

I'm curious: Most of us in the research community know that Dr. Josiah Thompson's upcoming new book, Last Second in Dallas, will include a detailed defense and confirmation of the acoustical evidence, and that he has been working with Dr. Barger on the acoustics evidence for years. The book will include information on a new acoustical test that was conducted. The book will be released on December 3, less than eight weeks from now. Do you plan on reading it?

[Joe Elliott]

You'd have to explain why five of the dictabelt impulse patterns match five of the test-firing shots in the correct locational order and in the correct average rate of speed. The odds that five impulses would match in the correct locational order alone are 1 in 125, less than 1%.

Back to the class room Mr. 1-in-125. It is 1 in 5*4*3*2*1 or 1-in-120. Go read a basic book on algebra which has a probability section.

You'd have to explain the N-wave and muzzle-blast patterns and the fact that they occur in the correct order and interval in the dictabelt impulse patterns. 

Already explained, many times. Sampling bias.

You'd have to explain why windshield distortion occurs when it should and does not occur when it should not in the dictabelt impulse patterns.

It’s not clear there shouldn’t always be distortion in any of the BBN motorcycle positions, either from the windshield or the torso on the rider.

You and Joe Elliott will always find some excuse, no matter how lame, to avoid admitting that the acoustical evidence is valid and solid. Just look at Elliott's last sad, labored, and evasive reply about Dr. Thomas's finding that the actual odds that chance caused the grassy knoll shot are 100,000 to 1 against. You guys will continue to deny the acoustical evidence because you don't want to believe it, because if you did, you would have to abandon your entire theory of the assassination.

And yet you can find no expert who endorses Dr. Thomas’s 1-in-100,000 claim Not even from his close friends.

I'm curious: Most of us in the research community know that Dr. Josiah Thompson's upcoming new book, Last Second in Dallas, will include a detailed defense and confirmation of the acoustical evidence, and that he has been working with Dr. Barger on the acoustics evidence for years. The book will include information on a new acoustical test that was conducted. The book will be released on December 3, less than eight weeks from now. Do you plan on reading it?

Mr. 1-in-125 is in the research community?

Plan on reading it?

[Michael T. Griffith]

Back to the class room Mr. 1-in-125. It is 1 in 5*4*3*2*1 or 1-in-120. Go read a basic book on algebra which has a probability section.

Oh, dear! Yes, there are 120 ways, not 125 ways, to sequence five events. So let's see what difference this makes in the odds: 1 in 120 is 0.83% vs. 0.8% for 1 in 125, which means the probability that the locational order in the matches is *not* the result of chance is 99.17% instead of 99.20%. I'll take 99.17%. No problem.

So, how do you explain the locational order in the matches, since the odds that such matches could be mere coincidence are 120 to 1 against?

Already explained, many times. Sampling bias.

LOL! What utter hogwash. What "sampling bias"?  There was no "sampling bias." The BBN scientists examined every peak and line of every second of the dictabelt tape. You've been making this bogus claim from the beginning, and I have debunked it several times. This is related to your erroneous claim that N-wave patterns were "scattered throughout" the tape, and that the BBN scientists dismissed the 4-second pattern merely because it was too short, when in fact it failed three of the five screening tests. 

Please explain how "sampling bias" would explain the fact that N-wave and muzzle-blast patterns, plus muzzle-blast echoes, occur, in the correct order and interval, in dictabelt impulse patterns, and that they occur during the correct timespan as established by five separate time indicators on the tape.

We're talking about incredibly tiny timing windows and intervals, and yet they occur in the correct order and in the correct interval--and at the correct timespan on the tape. The patterns are never once out of order. The N-wave always comes before the muzzle blast, just as it should, and it always comes at the correct interval (10-30 milliseconds ahead of the muzzle blast), and the muzzle blast is followed by its own echo, just as it should be. You must be kidding to suggest that such astonishing correlations are due to "sampling bias."

Why don't you go comb through the articles on the acoustical evidence on McAdams' propaganda website and copy and paste any parts from those articles that explain these correlations? We both know that you can't, because no such parts exist in those articles (I've read them). Or, find an explanation for these correlations in the NRC panel's report. Alas, the NRC panel never even mentioned these correlations, much less tried to explain them.

It’s not clear there shouldn’t always be distortion in any of the BBN motorcycle positions, either from the windshield or the torso on the rider.

You know this is false. They tested for and established the effect of windshield distortion. I've pointed this out to you several times now. And, as I've also pointed out to you, McClain's torso would not have intervened between the sound waves and the microphone. The waves were coming from high above him and to his right. At the very most, a very small part of the top part of his right right shoulder might have intervened, hardly enough to cause any substantial distortion.

And yet you can find no expert who endorses Dr. Thomas’s 1-in-100,000 claim Not even from his close friends.

You're lying again. And, uh, math does not care who endorses it. Math is math. Anyone who knows math well enough can check Dr. Thomas's calculations.

If Dr. Barger disagreed with Dr. Thomas about the 1 in 100,000 probability against chance, Dr. Thomas would have noted this. Why would Dr. Barger "disagree" with the 1 in 100,000 calculation unless it were wrong, which it is not?

The only possible question about the 1 in 100,000 calculation would involve the values that Dr. Thomas assigns for a few factors in the calculation, but those values are entirely reasonable, if not a bit on the overly safe side.

And, allow me to again note that Weiss and Aschkenasy noted in their report that their "5% or less" probability of chance calculation was "highly conservative" and that the actual probability of chance was "considerably less."

Speaking of Weiss and Aschkenasy, I notice that you snipped out the observations about their sonar analysis of the grassy knoll impulse pattern.

Mr. 1-in-125 is in the research community?

I guess you're the kind of person who waxes petty and juvenile when they finally catch someone in an error, even if it is a minor error (99.17% vs. 99.2%/0.83% vs. 0.8%). My 120 vs. 125 error pales in comparison to the blunders you have made. Shall I recite the long list of your egregious gaffes again?

Yes, I am in the JFK assassination research community. My research can be found referenced in many books and articles on the case. I have had a book on the case published by JFK Lancer (Compelling Evidence). Many JFK assassination websites include a link to my website, and some of them carry or provide links to a number of my articles. I have been interviewed as a guest expert on the JFK case on BBC Canada and on some local radio stations. And I have been invited to speak at JFK assassination conferences (but could not do so because of work commitments). So, yes, I am in the research community.

Plan on reading it?

Dr. Thompson informed me three weeks ago that he's going to send me a complimentary copy. I had already pre-ordered a copy on Amazon, but I will be grateful to get an autographed copy from him. So, yes, I'll be reading it.

[Joe Elliott]

Oh, dear! Yes, there are 120 ways, not 125 ways, to sequence five events. So, let's see what difference this makes in the odds: 1 in 120 is 0.83% vs. 0.8% for 1 in 125, which means the probability that the locational order in the matches is *not* the result of chance is 99.17% instead of 99.20%. I'll take 99.17%. No problem.

Your error shows a basic lack of mathematical reasoning ability. It doesn’t matter if your answer is 125, 121, 119 and that you are only off by a little bit. An answer of, let’s say 60, would be better, because that might mean you miss punched a “2” as a “1” when you multiplied 5*4*3*2*1 on a calculator. But it is clear you somehow thought that 5 to the third power would give the right answer. How do I get the correct answer? Look up the formula online? Have the formula memorized? No. I figure it out anew each time I run into the problem. If I need to find “how many ways to order 7 members (drawn without replacement) five years from now, I will figure out the formula anew, on my own. This keeps one in good practice. Relying on memory hinders one from becoming good at math.


Why does this all matter? Because it shows lack of basic mathematical ability which makes you a particularly unsuitable member of the “research community” to be the primary advocate of the Dr. Thomas 1-in-100,000 probability argument at this forum. Particularly since you can’t name any experts who publicly support this claim. The best you can do is name an expert in the field, Dr. Barger, who, while he too does not publicly support this claim, is a good friend of Dr. Thomas. And has proof read Dr. Thomas’s article, which does not tell us if he agrees with this claim or not. So, you cannot find support for this claim by Dr. Thomas and clearly you don’t have anything near the ability to evaluate it yourself. And Dr. Thomas’s claim is unbelievable. Dr. Barger, Dr Weiss do not know how to do this calculation and come within 3 orders of magnitude of the correct answer, but the insect expert Dr. Thomas does?

So, how do you explain the locational order in the matches, since the odds that such matches could be mere coincidence are 120 to 1 against?

LOL! What utter hogwash. What "sampling bias"?  There was no "sampling bias." The BBN scientists examined every peak and line of every second of the dictabelt tape. You've been making this bogus claim from the beginning, and I have debunked it several times. This is related to your erroneous claim that N-wave patterns were "scattered throughout" the tape, and that the BBN scientists dismissed the 4-second pattern merely because it was too short, when in fact it failed three of the five screening tests. 

Please explain how "sampling bias" would explain the fact that N-wave and muzzle-blast patterns, plus muzzle-blast echoes, occur, in the correct order and interval, in dictabelt impulse patterns, and that they occur during the correct timespan as established by five separate time indicators on the tape.

First, N-waves can be formed by lots of things, like a human voice. Take a look at a waveform caused by a voice and you will see N-waves. No acoustic expert, not Dr. Barger, not Dr. Weiss, ever claimed N-waves can only be formed by gunfire. Their claims were instead based on the similarity of a waveform with test shots recorded in 1978. The only problem is, that in most cases, the same 1963 waveform that resembled a shot from the TSBD also resembled a shot from the Grassy Knoll. That is a big problem which you seem oblivious to.


What sort of “sampling bias” to I propose. After “finding” the second shot, along a wide stretch of road, from microphone 2 ( 6 ) through 3 ( 5 ), they picked 2 ( 6 ), since it had the strongest correlation. Using it they could estimate where the first shot was, just using the Warren Report’s 11 mph estimate. It would only take one person on the BBN team to be aware of this fact to make use of it. So:

•   They searched for the first shot, 137.70, around 2 ( 4 ) through 2 ( 6 ), and found a random correlation at 2 ( 5 ).

•   They searched for the third shot, 139.27, around 2 ( 10 ) through 2 ( 12 ), and found a random correlation at 2 ( 11 ).

•   They searched for the fourth shot, 145.15, around 3 ( 4 ) through 3 ( 8 ), and found a random correlation at 3 ( 4 ).

•   They searched for the fifth shot, 145.61, around 3 ( 4 ) through 3 ( 8 ), and found a random correlation at 3 ( 5 ).


With a limited search like that, any correlation they do find will match their expectation formed after “finding” the second shot. It wouldn’t require a 1-120 chance at all.

So, once again, I have explained it to you. My theory may be true or false. But I am sure you will again ask me to explain how the correlation between time and the microphone locations could possibly occur, as if I haven’t done so already.


You could point out that I am speculating? Which I am. But if you propose they somehow completed in 10 days a compare/calculate process for all 3,024 combinations of the 7 1963 impulses with the 432 1978 impulses, you would be speculating as well. We are forced to speculate, because Dr. Barger never stated how through a search, they were able to complete in the 10 days. Did they search for the last impulse along microphones 1 ( 1 ) through 1  ( 12 )? We don’t know. But I am pretty sure they didn’t because they should have found “false alarms”, false positives, as they did from 2 ( 5 ) through 3 ( 8 ).

Why do I feel my “speculation” is superior to your speculation? Because mine explains why they had terrible correlations for the location of the shooter (finding correlations for both the TSBD and the KNOLL on the same impulse), terrible correlations for the location of the target (which generally did not match up with the location of the limousine) but good correlation for a plausible location for the motorcycle over time. Your speculation does not explain this at all.

You're lying again. And, uh, math does not care who endorses it. Math is math. Anyone who knows math well enough can check Dr. Thomas's calculations.

The key is not making the calculation. They key is knowing what values to substitute into the equation. And Dr. Barger and De. Weiss would have a better idea of this than the insect expert Dr. Thomas.

If Dr. Barger disagreed with Dr. Thomas about the 1 in 100,000 probability against chance, Dr. Thomas would have noted this. Why would Dr. Barger "disagree" with the 1 in 100,000 calculation unless it were wrong, which it is not?

No. It is far more plausible that Dr. Barger would remain silent if it disagreed with Dr. Thomas, then for him to remain silent if he agreed.

This “discovery”, would vindicate his work form 1979. It would remove a cloud that has been hanging over his head since 1982. He would try to make as many people aware of this “discovery” as possible.

On the other hand, if he disagreed, he might remain silent so as to not embarrass his friend. Besides, Dr. Barger needs all the support he can get.

Yes, I am in the JFK assassination research community. My research can be found referenced in many books and articles on the case. I have had a book on the case published by JFK Lancer (Compelling Evidence). Many JFK assassination websites include a link to my website, and some of them carry or provide links to a number of my articles. I have been interviewed as a guest expert on the JFK case on BBC Canada and on some local radio stations. And I have been invited to speak at JFK assassination conferences (but could not do so because of work commitments). So, yes, I am in the research community.

Wow. You are in real impressive company. 

To be worthy of being a member of such a prestigious group you should brush up on your basic math skills, which are based on reasoning skills.

[Michael T. Griffith]

Your error shows a basic lack of mathematical reasoning ability. It doesn’t matter if your answer is 125, 121, 119 and that you are only off by a little bit. An answer of, let’s say 60, would be better, because that might mean you miss punched a “2” as a “1” when you multiplied 5*4*3*2*1 on a calculator. But it is clear you somehow thought that 5 to the third power would give the right answer. How do I get the correct answer? Look up the formula online? Have the formula memorized? No. I figure it out anew each time I run into the problem. If I need to find “how many ways to order 7 members (drawn without replacement) five years from now, I will figure out the formula anew, on my own. This keeps one in good practice. Relying on memory hinders one from becoming good at math.

I hope someone is paying you to spend so much time posting such dishonest drivel. Dr. Barger himself noted that there was a 95% probability that the locational matches were not the result of chance (he was assuming only four gunshot patterns). The HSCA report also noted this. I've quoted their statements on this fact.

Plus, you can go to any math site that enables you to calculate permutations and confirm that there are 120 permutations of five objects and thus 120 ways to draw a given set of five numbers, such as the onlinestatbook.com:

http://onlinestatbook.com/2/probability/permutations.html

If the dictabelt impulse patterns are not Dealey Plaza echo patterns, then the five locational matches are purely by chance and have nothing to do with the geometry of Dealey Plaza. If so, any spurious match would be as likely to occur at any one microphone as at any other. If the five matches were merely all coincidences, then the sequence of matches in relation to the test microphone locations would have nothing to do with the location of the motorcycle--therefore, they would be in random order. But the matches are not in a random order: they are ordered 1-2-3-4-5, just as they should be if their impulse patterns were recorded by a motorcycle traveling with the motorcade.

Why does this all matter? Because it shows lack of basic mathematical ability which makes you a particularly unsuitable member of the “research community” to be the primary advocate of the Dr. Thomas 1-in-100,000 probability argument at this forum. Particularly since you can’t name any experts who publicly support this claim. The best you can do is name an expert in the field, Dr. Barger, who, while he too does not publicly support this claim, is a good friend of Dr. Thomas. And has proof read Dr. Thomas’s article, which does not tell us if he agrees with this claim or not. So, you cannot find support for this claim by Dr. Thomas and clearly you don’t have anything near the ability to evaluate it yourself. And Dr. Thomas’s claim is unbelievable. Dr. Barger, Dr Weiss do not know how to do this calculation and come within 3 orders of magnitude of the correct answer, but the insect expert Dr. Thomas does?

More of your dishonest, diversionary drivel. The reason for the overestimation has been explained to you twice now, but you just keep ignoring it and posting dishonest doubletalk.

There is no dispute that WA overestimated the probability that chance caused the correlations produced by the sonar analysis of the grassy knoll shot. The NRC panel realized this error was caused by assigning errant numbers for some of the values used in the calculation. Dr. Thomas has documented this fact in several writings, including his book, and I quoted a segment on this issue from his book a few replies ago.

And Dr. Thomas has explained in great detail why the actual odds against the chance explanation are 100,000 to 1. You have not said one word that directly deals with Dr. Thomas's detailed explanation, which includes all the math that produce those odds. As usual, you have done nothing but snipe at the point with dishonest evasions. Perhaps it is time to ask you to put up or shut up. If you have any evidence that Dr. Thomas's calculation is wrong, let's see it. Your failure to do so will be a clear admission that you can't refute it.

And I will again note that you keep ignoring the fact that WA themselves noted that their probability of chance calculation was "highly conservative" and that the actual probability was "considerably less."

First, N-waves can be formed by lots of things, like a human voice.

HUH??? This silly lie again? I've already debunked this nonsense. As I have documented for you, the BBN scientists screened all of the impulse patterns on the tape for N-wave-like characteristics. They did not have the test-firing data yet, but, being acoustical scientists who had dealt with gunfire cases before, they understood the general characteristics of N-waves and how they would appear on a oscillogram/spectrogram. This was one of the five screening tests, and only the six impulse patterns between 136.20 and 146.30 passed this test.

Sheesh, why am I having to explain this to you again? Because you won't deal honestly with the evidence.

Take a look at a waveform caused by a voice and you will see N-waves. No acoustic expert, not Dr. Barger, not Dr. Weiss, ever claimed N-waves can only be formed by gunfire. Their claims were instead based on the similarity of a waveform with test shots recorded in 1978. The only problem is, that in most cases, the same 1963 waveform that resembled a shot from the TSBD also resembled a shot from the Grassy Knoll. That is a big problem which you seem oblivious to.

You're lying again, or else you're proving that you just don't understand the evidence that I have presented to you. I also address this issue in my article, but you have said nothing about the evidence I present therein on this subject.

Let's just say this: I invite you to cite a single source on acoustics that says that actual N-waves--not sound waves that bear a general resemblance to N-waves, but waves that really are N-waves--can be formed by human speech. For a sound wave to be an N-wave, it must come before the muzzle blast, and it must do so in the very small and specific timeframe of 10-30 milliseconds, and then the muzzle blast will come after the N-wave, and then the muzzle blast will be followed by its own echo. The WA report explains how these groups of impulses appear in the graphical representation of one of the test-firing shots:

The first waveform appearing in the graph, the large peak at the left-hand side, corresponds to the supersonic shockwave [N-wave] of the rifle bullet. The second large peak is the waveform of the muzzle blast. Following it, with generally diminishing heights, are the waveforms of the echoes of the muzzle blast. The delay time of each echo was determined by direct measurement of the distance from the leading edge of the muzzle blast waveform to that of the echo. (8 HSCA 24)

And, as I've already documented several times, the grassy knoll impulse pattern contains impulses that match the abovementioned characteristics. But you just keep ignoring this fact.

A loud human yell or a burst of static can cause an impulse pattern that will generally resemble the graphical representation of an N-wave on a spectrogram or oscillogram. However, that yell or static is not going to show the N-wave coming ahead of the muzzle blast and doing so at the correct interval, and it is not going to show the muzzle-blast echo coming after the muzzle blast. Professor Aschkenasy discussed this matter in his HSCA testimony. WA were displaying an oscillogram, and Congressman Edgar thought a certain set of "squiggly lines" were caused by supersonic gunfire. Professor Aschkenasy explained the congressman's misunderstanding and explained that identifying an N-wave required a more detailed analysis:

Mr. EDGAR. How do you know that the squiggly lines you are looking at are really supersonic?

Mr. ASCHKENASY. They are not supersonic. Those are sound waves. Those are presentations of sound waves. The question, what you might want to ask, is about whether we can tell a bullet was there, namely, was it creating a supersonic shock wave? That is what you are questioning. And those are not supersonic sound waves. Those are sound waves as recorded by a microphone, and put into electrical form by the equipment that was used to transmit it and record it, and there is nothing supersonic in those squiggles that we have up there on that board.

Mr. EDGAR. Would you answer the question I wanted to ask?

Mr. ASCHKENASY. Well, because you have a bullet that travels faster than sound, it will get to someplace faster than the sound reaching that same point. We are talking about two components, the bullet and the muzzle blast.

The bullet flies, let's just pick a number, at 2,220 feet per second, so that it travels at twice the speed of sound for this particular example, when you fire the gun. And it flies, let's say for 200 feet. It will get at the target 200 feet away in a certain period of time. Just like a boat pushes the water ahead of it creating the V-shape wake behind the boat, that is similar to what you see in a shock wave from a bullet. And that shock wave is what is recorded by the microphone that is right next to the target.

Sometime later, finally the sound catches up to it and gets to the target, and the muzzle blast is recorded. That interval of time is fixed, by the fact that you have a certain muzzle velocity and you have a certain distance, they occur in a fixed time relationship.

We have also the first, it's covered by the photograph--could somebody remove that photograph, please. If I may point something out there.

Mr. EDGAR. Yes.

Mr. ASCHKENASY. I can point out here also these first impulses before the muzzle blast, those are the shock waves, and if you look carefully--I am sure you cannot look that carefully at that distance--but if you look at these graphs, because these microphones are located at different positions on the street, the relationship between the shock wave and the muzzle blast changes, and it changes in a predictable manner because the manner in which you expect them to change is related to where the observer, or the microphone is picking up both the shock waves and the muzzle blast.

Now, you measure here about on the average of about 14 milliseconds, 14 thousandths of a second delay between the shock wave and the muzzle blast. We go now here to the police tape and the measurement that we found was around 24 milliseconds here. It is now reasonable to assume because of the measured time interval that the impulse may have attributes of a shock wave.

If you expand the experiment tape and take an even better look at it, you find there is a little shock wave echo right in between the shock wave and the muzzle blast, and if you expand the police tape properly, you find similar patterns, implying to us that this impulse has the qualities, attributes of a shock wave. (5 HSCA 609-610)

You might ask yourself this question: Why do you suppose the NRC panel declined to offer any explanation for the N-waves on the dictabelt recording? Hey? Why do you suppose that was? Why do you suppose they never made your ludicrous claim that N-waves are "scattered throughout" the dictabelt recording? Hey? Why? Why do you suppose they opted not to further embarrass themselves by claiming that the N-waves were just loud human speech or bursts of static? Hey? Why?

What sort of “sampling bias” to I propose. After “finding” the second shot, along a wide stretch of road, from microphone 2 ( 6 ) through 3 ( 5 ), they picked 2 ( 6 ), since it had the strongest correlation. Using it they could estimate where the first shot was, just using the Warren Report’s 11 mph estimate. It would only take one person on the BBN team to be aware of this fact to make use of it. . . .[SNIP]

With a limited search like that, any correlation they do find will match their expectation formed after “finding” the second shot. It wouldn’t require a 1-120 chance at all.

So, once again, I have explained it to you. My theory may be true or false. But I am sure you will again ask me to explain how the correlation between time and the microphone locations could possibly occur, as if I haven’t done so already.

You could point out that I am speculating? Which I am. But if you propose they somehow completed in 10 days a compare/calculate process for all 3,024 combinations of the 7 1963 impulses with the 432 1978 impulses, you would be speculating as well. We are forced to speculate, because Dr. Barger never stated how through a search, they were able to complete in the 10 days. Did they search for the last impulse along microphones 1 ( 1 ) through 1  ( 12 )? We don’t know. But I am pretty sure they didn’t because they should have found “false alarms”, false positives, as they did from 2 ( 5 ) through 3 ( 8 ).

Why do I feel my “speculation” is superior to your speculation? Because mine explains why they had terrible correlations for the location of the shooter (finding correlations for both the TSBD and the KNOLL on the same impulse), terrible correlations for the location of the target (which generally did not match up with the location of the limousine) but good correlation for a plausible location for the motorcycle over time. Your speculation does not explain this at all.

This is just so bogus. Oh my goodness. I hate to sound like a broken record, but you leave me no choice: This is just unbelievably erroneous and misleading. You are either totally confused or you are being very disingenuous. You seem to be hoping that if you type enough words and cite enough numbers, no matter how silly or idiotic they are, that some people will think you know what you're talking about. But in fact you have, once again, totally mangled what the BBN and WA scientists did regarding false matches/false alarms and how they weighted the matches. You have also simply ignored the WA sonar analysis, which destroyed the false-match argument by reducing the microphone spacing from 18 feet to 1 foot and by reducing the acceptance window from 6 milliseconds to 1 millisecond.

The key is not making the calculation. They key is knowing what values to substitute into the equation. And Dr. Barger and De. Weiss would have a better idea of this than the insect expert Dr. Thomas.

No. It is far more plausible that Dr. Barger would remain silent if it disagreed with Dr. Thomas, then for him to remain silent if he agreed.

This “discovery”, would vindicate his work form 1979. It would remove a cloud that has been hanging over his head since 1982. He would try to make as many people aware of this “discovery” as possible.

On the other hand, if he disagreed, he might remain silent so as to not embarrass his friend. Besides, Dr. Barger needs all the support he can get.

More of your dishonest drivel. You are quite the piece of work. Again, the reason for the errantly assigned values in the WA calculation has been explained. No one disputes this. I quoted Dr. Thomas's explanation, and you did not touch a single part of it. When you replace the errant values with the correct values and redo the calculation, the odds that chance caused the grassy knoll shot come out to 1 in 100,000. Since you deny this, show us where Dr. Thomas's calculation is wrong.

And, yes, I am certain that Dr. Barger was pleased to find out that the WA probability of chance for the grassy knoll shot was actually even far lower than WA said it was. Again, why would Dr. Barger not agree with Dr. Thomas's calculation when that calculation can be verified by anyone with sufficient math skills to do so? You have ducked this question twice now? And, again, when you are going to explain why Dr. Thomas's calculation is wrong?

When Dr. Thompson's book Last Second in Dallas is released on December 3, just remember that Dr. Thompson has been consulting with Dr. Barger on the acoustical evidence for the last several years.

Wow. You are in real impressive company. To be worthy of being a member of such a prestigious group you should brush up on your basic math skills, which are based on reasoning skills.

Wow. Really? Uh-huh. Given the horrendous blunders you have committed in your amateurish attacks on the acoustical evidence, you are in no position to be telling anyone else about their mistakes. My math error was miniscule and produced a tiny difference of 0.83% vs. 0.8% and 99.17% vs. 99.2%. That error pales in comparison to your numerous embarrassing gaffes.

[Joe Elliott]

I hope someone is paying you to spend so much time posting such dishonest drivel. Dr. Barger himself noted that there was a 99% probability that the locational matches were not the result of chance. The HSCA report also noted this. I've quoted their statements on this fact.

What Dr. Barger has failed to do is to state:
•   Unequivocally state that all combinations of the seven-1963 impulses were compared with the 432-1978 impulses.
We need this statement, at a minimum to conclude that there was an amazing, 1-in-120, correlation between the entire data set and the position of the motorcycle, appearing to make its way up Houston and Elm in a consistent manner.

Question 1:

After 40 years, has Dr. Barger ever made such a clear statement?


Yes, we know there is good correlation with the consistent progress of the motorcycle hypothesis with the data the BBN did check in 1978. But this is meaningless unless we know that all possible correlations were checked.

If BBN only checked for correlations:

For the first shot, between microphones 2 ( 5 ) and 2 ( 6 ),
for the second shot, between microphones 2 ( 6 ) and 3 ( 5 ),
for the third shot, between microphones 2 ( 11 ) and 2 ( 11 ),
for the fourth shot, between microphones 3 ( 4 ) and 3 ( 8 ),
for the fifth shot, between microphones 3 ( 5 ) and 3 ( 8 ),

then there is no amazing 1-in-120 coincidence. The correlations they do find would naturally match the scenario of a motorcycle moving forward.

For all we know, if a complete search was done, we might find:

•   correlations for the first, second and third shots in the range of microphones 3 ( 9 ) through 3 ( 12 ),

•   correlations for the fourth and fifth shots in the range of microphones 1 ( 1 ) through 2 ( 4 )

and we would find what we found with the “location of the shooter” and the “location of the target” correlations, random correlations for the “location of the motorcycle”.

So, again, ideally, I would like from Dr. Barger:
•   Unequivocally state that all combinations of the seven-1963 impulses were compared with the 432-1978 impulses.
•   State the time needed to check and calculate one combination of 1963 impulse with a 1978 impulse.
•   The number of men assigned to this task.
•   The number of man hours that were used in this task alone.
•   A statement that these were “blind” comparisons. That the checker did not know which of the seven impulses he was checking nor which microphone he was checking for. This would eliminate the expectations of a checker failing to find a correlation for the fourth shot at microphone 1 ( 7 ) because he thought any such correlation would be impossible.

But after 40 years, we have had no such statement from Dr. Barger.


One last time, regarding the match between the data being consistent with a motorcycle making progress down Houston and Elm:

•   if a complete, thorough, unbiased, comparison of all seven-1963 impulses were each compared to all 432-1978 impulses, then can we conclude that the odds were 1-in-120. Well, actually somewhat less than that because some of the found correlations were out of order.

•   But if this was not done than the true odds may be more like 1-in-1, depending on how incomplete the search was.

There is no dispute that WA overestimated the probability that chance caused the correlations produced by the sonar analysis of the grassy knoll shot. The NRC panel realized this error was caused by assigning errant numbers for some of the values used in the calculation. Dr. Thomas has documented this fact in several writings, including his book, and I quoted a segment on this issue from his book a few replies ago.

And Dr. Thomas has explained in great detail why the actual odds against the chance explanation are 100,000 to 1. You have not said one word that directly deals with Dr. Thomas's detailed explanation, which includes all the math that produce those odds. As usual, you have done nothing but snipe at the point with dishonest evasions. Perhaps it is time to ask you to put up or shut up. If you have any evidence that Dr. Thomas's calculation is wrong, let's see it. Your failure to do so will be a clear admission that you can't refute it.

Dr. Thomas has explained this, but you can find no expert who agrees with him. Only an expert who agreed to proof-read his work and who has become a good friend of his. But no statement from any expert who says Dr. Thomas is correct about this 1-in-100,000 estimate of the chance.

And I will again note that you keep ignoring the fact that WA themselves noted that their probability of chance calculation was "highly conservative" and that the actual probability was "considerably less."

As to be expected if they had actually made “highly optimistic” estimates due to being subconsciously influenced by “p-hacking”. Which just happened to cause them to arrive at a p-value of 95%. The ultimate goal of those under the influence of p-hacking.

HUH??? This silly lie again? I've already debunked this nonsense. As I have documented for you, the BBN scientists screened all of the impulse patterns on the tape for N-wave-like characteristics. They did not have the test-firing data yet, but, being acoustical scientists who had dealt with gunfire cases before, they understood the general characteristics of N-waves and how they would appear on a oscillogram/spectrogram. This was one of the five screening tests, and only the six impulse patterns between 136.20 and 146.30 passed this test.

You and Dr. Barger have something in common. You both fail to see that there are not six impulses but seven. They are all listed on BBN’s Exhibit F-367. I will list them again. They are:

https://mcadams.posc.mu.edu/russ/infojfk/jfk2/f367.htm

136.20
137.70
139.27
140.32
145.15
145.61
146.30

They thought all seven were possible candidates as gunshots. They did comparisons of each with some of the 432-1978 recordings. They rejected 136.20 and 146.30 as gunshots because they could not find any correlation stronger than 0.5.

Nowhere do they state, from their “N-wave” characteristics or for any other property, that they were gunshots. It was only through the similarity of 4 of the waveforms with some of the 1978 test impulses that they concluded that these 4 were gunshots.

Again, you seem to be the only one who thinks, that from the 1963 impulses alone, one can tell that they could only be formed by gunshots.


Question:

Where is the statement form the BBN that N-waves can only be formed by gunfire?

Question:

If these N-waves can only be formed by gunfire, why did they reject the impulses at 136.20 and 146.30, as shown on BBN Exhibit F-367?



From reading their statements I glean that the BBN settled on these seven impulses because:
•   The number of impulses match the very upper limit to the number of shots that may have been fired.
•   The seven impulses occurred over a span of 10.1 seconds, and they made it clear they were looking for a span of at least 5 seconds.

These is nothing form their statements, that from these seven impulses alone, that they could tell that these were formed by gunshots. You seem to be the only one who makes this claim.

Eventually, they only concluded that they were gunshots, because of their similarity to either a test shot from the TSBD or a test shot from the Grassy Knoll. The only problem is, and it is a whopper, is that 3 of the shots had similarities with both a test shot from the TSBD and a test shot from the Grassy knoll. Indicating that they were finding at least some false correlations, what they called “false alarms”. Which casts doubt on all their correlations. I’m not lying. This is what their own data, in BBN Exhibit F-367 shows. Dr. Barger seems to have turned a blind eye to what his own data was trying to tell him.

[Michael T. Griffith]

Below is large segment from Dr. Donald Thomas's discussion on the WA sonar analysis in his book Hear No Evil: Politics, Science, and the Forensic Evidence in the Kennedy Assassination. In the segment below, Dr. Thomas explains every step in the WA sonar analysis and the findings that came from the analysis, the key finding being that the dictabelt tape contains the impulse pattern of a shot fired from the grassy knoll during the assassination.

Dr. Thomas is a USDA entomological research scientist and an expert in statistical analysis. His first, and ground-breaking, article on the acoustical evidence was published in the peer-reviewed criminal science journal Science & Justice. It's worth noting that Dr. James Barger, the lead BBN acoustical scientist for the HSCA, proof-read the article before it was published. Also, Dr. Thomas consulted extensively with Dr. Barger while writing the four chapters on the acoustical evidence in his book. Here is the segment from Dr. Thomas's discussion on the WA sonar analysis:

When Mark Weiss and Arnold Aschkenasy were asked by the Assassinations Committee to review and refine BBN's analysis of the Dallas Police recordings, it was specifically for the purpose of moving Barger's 50-50 estimate off the fence. BBN's analysis was adequate to meet its limited objectives, so far as they went. Barger and his team had spread a net to catch a motorcycle and they had succeeded. The order in the data meant that the sounds of the assassination gunfire were almost certainly on the police recording. But when it came to nailing down the number of shots and their origins, the conclusions were more tenuous. Call them what you will, false alarms or auto-correlations, the detection method was too porous to assure that quirks and spurious matches wouldn't leak through. The primary weakness in the analysis was in the width of the coincidence windows. The solution to the problem was to narrow those windows to a slit.

The only way to narrow the coincidence windows was to get a better fix on the motorcycle position. It was unlikely that microphone No. 4 was exactly in the same spot as the subject motorcycle, if one was there, and very likely to be off by six feet or more. One solution would be to go back to Dealey Plaza and saturate the area around microphone No. 4's position with microphones at 2-foot spacings and record more test shots. For a street 36 feet wide, and to cover nine feet in each direction, 180 microphones would be required to cover the relevant area. This procedure would narrow the coincidence windows to ±1 millisecond, but it raises other problems.

Measuring echo delay times to within an accuracy of one millisecond requires consideration of the fact that the subject microphone was in motion. During the one-third of a second that the gunshot resounds, the motorcycle would have moved several feet. Thus, in order to duplicate the conditions during the assassination, one would have to array 180 mobile microphones and have them moving in the same direction and at the same speed as the suspect motorcycle. The logistics of such an exacting experiment would be impractical. But, essentially, that is what Weiss and Aschkenasy did - but they did it analytically.

The basic principles of echo location are well known; well enough that submarines don't have windows. A submarine navigates acoustically by bouncing sounds off of its environment. The on-board computer calculates the distances and angles to surrounding objects as it moves through the water based on the echo delay time. The algorithms in the computer's sonar software adjust for the vessel's speed and compensates for the temperature of the water. Sound travels slower in polar waters than it does through the warm waters off say, the coast of Cuba. Weiss and Aschkenasy were the ones who wrote those software programs and they now applied the same principles to the patterns on the DPD recording.

If the impulses on the tape were the echoes of gunfire reflecting and refracting off of the structures in Dealey Plaza, then the first step in the solution lay in matching the echoes to those structures. Weiss and Aschkenasy obtained a surveyor's map showing the buildings, monuments, and other structures in Dealey Plaza accurate to within one foot. They placed a pin at the position where BBN had placed a shooter on the grassy knoll and a pin on the microphone position (No.4 of Array 3) on Elm Street which had recorded the test pattern which correlated to the impulse pattern on the police tape. With a string attached to both pins they proceeded to measure the precise distances along the most direct paths from the shooter position to each structure in Dealey Plaza, and from there directly to the microphone.

They included every structure that presented a significant sound reflecting surface or sound refracting comer in the northeastern quadrant of Dealey Plaza. Eventually they were able to correlate twenty-two such structures in all, accounting for 26 major echoes or re-echoes.

In spite of such perfection, the match so obtained, though impressive, was only a start. The evidence tape was dense with impulses, most of which were piston firings from the motorcycle's motor. There were so many and they were so closely spaced, at least some could be expected to align with impulses on the test pattern, and there was a mathematical probability that many could by chance.

So, for the next step in the analysis, Weiss and Aschkenasy used amplitude as the criterion to eliminate the motor noise impulses, basically the same procedure used by the BBN team, except that they had now identified a specific impulse in the evidence pattern as the putative muzzle blast.70  Including only the large amplitude impulses, those loud enough to exceed a sound threshold determined by the average background noise, a total of 13 such loud impulses on the test pattern remained and 15 such impulses on the DPD recording. When the echo delay times were compared, eleven of the impulses were coincident to within ±1
millisecond, giving a binary correlation coefficient of 0.79.71  A probability calculation of the odds of attaining this good a match from a chance assemblage of static or noise impulses was calculated to be about 5%.72  Thus, Weiss and Aschkenasy had succeeded in moving the mathematical probability for the grassy knoll gunshot from an anemic 50% to a robust 95%. (Thomas, Hear No Evil, 2010, pp. 593-595)

[Joe Elliott]

Below is large segment from Dr. Donald Thomas's discussion on the WA sonar analysis in his book Hear No Evil: Politics, Science, and the Forensic Evidence in the Kennedy Assassination. In the segment below, Dr. Thomas explains every step in the WA sonar analysis and the findings that came from the analysis, the key finding being that the dictabelt tape contains the impulse pattern of a shot fired from the grassy knoll during the assassination.

Dr. Thomas is a USDA entomological research scientist and an expert in statistical analysis. His first, and ground-breaking, article on the acoustical evidence was published in the peer-reviewed criminal science journal Science & Justice. It's worth noting that Dr. James Barger, the lead BBN acoustical scientist for the HSCA, proof-read the article before it was published. Also, Dr. Thomas consulted extensively with Dr. Barger while writing the four chapters on the acoustical evidence in his book. Here is the segment from Dr. Thomas's discussion on the WA sonar analysis:

Yet not one acoustic expert has expressed any support for Dr. Thomas’s most important claim. That there is only a 1-in-100,000 chance that correlation found by Weiss and Aschkenasy correlates with the impulse at 145.15 by random chance. Not even Dr. Barger.