All right. Thank you, ladies and gentlemen. Please be seated. The record should reflect we have been rejoined by all the members of our jury panel. And Mr. Neufeld, you may commence your cross-examination.
Good morning, ladies and gentlemen.
THE JURY: Good morning.
CROSS-EXAMINATION BY MR. NEUFELD
Okay. Dr. Weir, you said a moment ago at the conclusion of the direct examination that in assessing the estimates provided by the Department of Justice in California and Cellmark that you felt that they were, "Good estimates"; is that correct?
Umm, well, I think based on those data they are the appropriate estimates and for a single point estimates I believe they are the best, yes.
Do statisticians use the term "Best estimate" when applying a statistical estimate of the occurrence of something?
Not generally. We have more specific terms and that is--we probably need to explore that further. It is kind of vague. We might talk about--
Do the estimates that are provided by the Department of Justice and Cellmark in this case incorporate the corrections that you deem appropriate?
Well, that is a two-part question. As with the other labs, I produce an estimate by multiplying numbers together. That is the estimate. That is what I characterize as being a good estimate. Along with that estimate, as I explained yesterday, I would do two other additional steps, attach confidence limits and allow for appropriation structure, and those additions have not been done by the two labs.
Okay. So if you were giving the best estimate in this case it would--no. 1, it would account for a substructure; is that correct?
No. I think what I said is the best estimate is what they have done. Along with that estimate I would do these additional two steps.
When you do your additional two steps you come up with a new estimate; isn't that correct, Dr. Weir?
That's right. As I explained yesterday, in public opinion poles, we have two parts to the answer, the 47 percent and the plus and minus.
And so, Dr. Weir, if you made your--your additional step, which is the result of accounting for possible substructure, that would come up with an estimate which would be more favorable to the accused than the estimate provided by Cellmark and the Department of Justice; isn't that correct?
And sir, the other procedure, a step that you take that was not taken by Cellmark and the Department of Justice, is to include what you just said a moment ago called "Confidence intervals," correct?
If you provided a 99 percent upper confidence limit to the numbers provided by Cellmark and DOJ, you would again come up with a number which is more favorable to the accused than the estimates that have been provided by Cellmark and the Department of Justice; isn't that correct?
That's right. Their original estimates of course remain good estimates, but along with those good estimates we attach the second part of the answer. We say here is our good estimate and here is our confidence limit. Here is our good estimate and here is an appreciation of how it will change with population structuring.
It is not correct to characterize one as being right and one wrong. The answers are, as you appreciate it, are complex and they are multi-faceted. What we have seen on the board here is the simplest interpretation.
Sir, all I'm asking is would you agree that if you make the additional two steps that you have suggested, namely for substructure and for confidence intervals, which were not done by Cellmark and DOJ in their testimony, you would arrive at an estimate that is more helpful to the accused?
Well, I don't--I don't think that is quite saying what I said. We have--we have the estimate as presented on the boards here. Along with that estimate we give a range and showing how it would change. It is not that we are putting our faith on any one of these single numbers. We can't divorce the estimate from--from the range, they go together, so we don't--it depends on your point of view, I suppose, but the Defendant would be happy with the 47 percent because he would add on the three and the opponents would be happy with subtracting it. Of course they are both stretching the limit. The fact remains the data support the 47 percent along with that, so it must be a compound answer. The extreme limits will be more favorable to the Defendant of course. And the estimate remains the good estimate.
Sir, but by adding--by doing your two steps, you will get a frequency that is more common than the frequency articulated by Cellmark and DOJ; is that correct?
I will be presenting a range. I will be giving the original estimate together with the range. The outside end of that range, the very outside end of the range will be more favorable to the Defendant.
And it is that outside end of the range which was not presented by Cellmark and the Department of Justice, correct?
Well, I don't really have much choice. As a statistician I need to explain estimates and the ranges, and I think as we will--and I think I said that the amount of difference does not change the conclusions that we've heard. Matching profile based on several loci is a rare event. The fact that it is rare--and that also I should believe that we should be worrying about the fact that it is rare. That fact is not changed by any of these manipulations of the numbers.
Let me ask you this, Dr. Weir, just to change gears for a second: During your career, sir, have you ever made any mistakes as a professional?
Okay. Would it be fair to say that you may have made mistakes during your career in certain facts that were asserted?
Okay. And would you also agree, doctor, that perhaps you have even propounded theories that at some time in the future ultimately were refuted?
I don't believe so. I don't--I can't think of any published refutations of any of my papers.
How about unpublished refutations? In other words--I'm sorry, let me clarify. Have you ever propounded a theory not in the publication or not in writing which was subsequently refuted?
Goodness, I don't know how I could answer that question. If I didn't say it or write it, how could I--
I'm asking if you verbally expressed it to other people or other colleagues in the field?
Oh, I can't be held accountable for every coffee conversation I may have ever had.
KEY QUOTEOkay. Now, I believe you said on direct examination that you yourself do not operate a forensic laboratory; is that right?
Now, but you would agree, sir, that it is not necessary to have training in forensic science or to operate a forensic laboratory in order to provide an expert opinion on the appropriateness of certain statistical methods that are being used for interpreting forensic DNA data?
Okay. And would you agree, sir, that the relevant areas of expertise in this particular subject include theoretical population genetics and the field of statistics?
Now, there are--would you agree, sir, that there are often misunderstandings, if you will, about what statistics mean and what they don't mean?
Okay. Well, let's just try and clear up a few things for those of us who aren't statisticians, sir. Would you agree that the numbers that are reflected on the Prosecutor's exhibits in this case are not the probability of Mr. Simpson's guilt or innocence?
That's absolutely correct, and nothing I have said or will ever say has to do with the probability of guilt.
I appreciate that. And would you agree, sir, that these numbers that appear on these boards are not the probability that someone other than Mr. Simpson is the source of any particular stain?
No. I think we've been very clear what these numbers are. These are the probabilities or the frequencies of finding those stains if they were contributed by unknown people. And I will say it once more, because it is easy to get confused, these are frequencies of the stains, be they single or mixed, if they were contributed by unknown people.
So just so I'm clear on this, Dr. Weir, if someone said or suggested that these frequencies were the probability that someone other than Mr. Simpson was the source of a particular stain, that would be scientifically wrong, correct?
And the numbers up on this board and the other boards, Dr. Weir, they tell us nothing about the probability of a false or misleading match due to errors in the collection or handling of samples; isn't that correct?
These numbers on the board are the frequencies of the stains if they were contributed by unknown people. That is all they are.
Well, sir, therefore would you agree that these numbers tell us nothing about the probability of a false or misleading match due to errors in the collection or handling of samples?
And therefore they are not the probabilities or frequencies that I just described; is that correct?
That's right. They are the frequencies of the stains if they were contributed by unknown people.
And sir, would you agree that these statistics that have been propounded in this case by Prosecution witnesses tell us nothing about the probability of a false or misleading match due to evidence tampering?
These numbers on the board reflect only the frequencies of the stains if they were given by unknown contributors. That is all they--that is all they convey. They don't convey any of these other things.
Well, and by your answer, sir, are you therefore saying that they don't tell us anything about the probability of a false or misleading match due to evidence tampering?
And would you agree, Dr. Weir, that the numbers on these boards are irrelevant if the match is due to the combined effects of total degradation followed by cross-contamination with Mr. Simpson's blood, either from Rockingham swatches or from a minuscule amount of fresh blood from his reference sample?
Would you agree that the statistics on these boards, Dr. Weir, is irrelevant if the match is due to the combined effects of total degradation and cross-contamination with Mr. Simpson's blood either from--I'm sorry, with Mr. Simpson's blood from the Rockingham swatches?
I'm not sure I am here to talk about relevance. These numbers that I have verified are the frequencies of those stains if they were given by unknown contributors. As I think I said, my analysis starts with the determination of the laboratory that there is a match and these are the details of the matching profile. All of the numbers refer to the frequencies of the stains that are presented on the board. They have nothing--they make no statements about anything else and I'm not talking about relevance. That is beyond my expertise.
In fact, Dr. Weir, would you agree that the validity of all of your estimates and the estimates propounded by other Prosecution witnesses in this case assume that the laboratory work in this case was valid?
The validity of my results are self-contained. They follow from the profiles given. Once the profile is established and declared to match, everything else I have done is valid. Anything before that step I don't address.
Well, you say you don't address it, but you assume that all the work was done and all the results are valid; isn't that correct?
I think I have said that. I assume that the facts on the board are indeed the truth, they are the profiles and they do match.
Well, it is not just that they are the profiles. You are also assuming, sir, the validity of the underlying processing, collection, preservation and handling of the evidence by everybody who did the work up until the time that you analyze it; is that correct?
I--I said that I'm going to start with these profiles and the fact that they match and the details of them. Anything that goes prior to that is beyond my expertise.
Well, let me ask you this, sir: Referring to your report if you will--do you have a copy of it in front of you?
I call your attention to page 2. I'm sorry. On mine it is page 3. You state the following, and I quote: "These notes assume the validity of all DNA profiling by Cellmark Diagnostics, the California Department of Justice and the Federal Bureau of Investigation." Is that correct?
That's right. I think that is just what I've been saying. I start with a determination of the match and the nature of the profile. Anything prior to that is beyond what I do.
Well, you included, along with Cellmark and the Department of Justice, the Federal Bureau of Investigation; is that right?
Umm, are you aware of any DNA profiling that the Federal Bureau of Investigation did on evidence samples in this case?
No, I'm not, but it is obvious why I have that statement in there. My entire report is based on the FBI data. I must assume that those data are an accurate representation of the profiles of the people sampled. I have no choice but to assume those databases are valid.
Now, doctor, it is curious, however, that when you assume the validity of the DNA profiling of those different laboratories, you don't--or you fail to state whether you assume the validity of the DNA profiling done by the Los Angeles Police Department in this case?
And are you aware, sir, that the Los Angeles Police Department conducted DNA profiling in this case?
Well, when you became aware of it at 8:00 p.m. last night did you feel it was necessary to amend your report?
Umm, if I would have had time after I finished working at midnight I may indeed have done that and added in the words "LAPD."
Are you aware, sir, that it was the Los Angeles Police Department that was responsible for the initial collection, preservation, packaging and processing of the evidence in this case?
Would you agree, sir, that if the DNA profiles in this case were the result of laboratory error at the Los Angeles Police Department then all of your statistics, whether they are 1 in 50 billion or 1 in 5 million would be meaningless?
All of my statistics are completely meaningful because they contain a complete description of what they assume. They assume that these are the profiles, that they match, that they have this particular nature. With that starting point, everything else I've done is valid. It is certainly not meaningless. Everything I have done I believe is meaningful, given the starting point I've described.
Sir, so what I've asked you then, if in fact the DNA profiles were the result of error, then would your statistical estimates be meaningless?
This is the same question we have now asked for the eighth time. Proceed. I'm going to sustain the objection.
Now, on direct examination you discussed with this jury or described to this jury that when you use the product rule, when you multiply these different frequencies, it is essential that the various factors that you are multiplying together are independent of one together; is that correct?
Yes, they must. We are regarding this profile as a set of pieces of information. That is typically ten or twelve. Sometimes it is even a lot more separate pieces of information. And when I say "Separate," that means we are assuming that in fact they are independent. It is not--although we talk about the matching profile, it is actually a lot more than that, because the word "Profile" is a word that describes a whole lot of matches. They are matches at both the bands of all the probes. So it is a very--very--if you like, very compelling events that we have these--this complete match. It is compelling because we regard these as being separate items of the match, and to put a number on that we have to regard them as being separate and independent, so we can do the product.
And I believe you said, Dr. Weir, yesterday morning to this jury or yesterday afternoon, excuse me, when you were discussing the factors which suggested independence here, I believe that one of the things you said to the jury was that the marker systems that they are relying on for forensic typing are what is commonly referred to as junk DNA. Do you remember saying that, sir?
I don't think I said that. I said--I said, I imagine--I hope I said that this is DNA of no known function.
Okay. And you said that because the marker system that they are relying on for forensic work are DNA of no known function, that is an additional piece of evidence, if you will, to suggest that there is no association here; is that correct?
No, I didn't say that. I said because the RFLP types have no known function, we may expect them to be independent. However, we are not entitled to go on expectations. We must perform the tests that I described and did indeed conduct and it wouldn't really matter if there was some biological agent at work affecting these--these various pieces of DNA, these various types providing the data, providing their frequencies are consistent with independence, then we may proceed with the assumption of independence and do the product rule. So when I do my test, as a statistical exercise it makes no statement about the underlying biology. Some of these pieces of DNA may in fact be very important for our well-being. We don't know that. And why I say they may be important, because if they were important, it might be that people with a certain type of DNA don't live as long, don't have as many children, and that would affect the frequency, so there is all that kind of phenomenon that may be going on. We don't think it is, but I may indeed be. However, when we look at the data, when we look at the databases and the frequencies, we find that these frequencies are consistent with being independent, so as a statistician I'm doing a statistical test. I'm making no statements about the underlying biology.
Well, Dr. Weir, are you aware of the fact that for many of the PCR forensic markers that are relied on in this case, they do code for proteins and they do have some effect on our health and well being?
Some of the PCR markers are inside our genes and they could very well, as I've just described, have an effect. What is kind of interesting, I suppose, and it is certainly reassuring for this forum, is that their frequencies are consistent with independence. Their frequencies show DNA. In fact, we may multiply them together, regardless of what effect those DNA's actually have on us.
Now, this morning and yesterday, sir, you provided this jury with an approach to interpreting the mixtures; is that correct?
Yes, I did. We have given some numbers now which show us the frequencies of these various mixed stains if they were contributed by people unknown to us. That is what all these numbers mean and certain specified numbers of unknown people.
And when you provided this information to the jury, you simply provided what you found to be the most common frequency and then you gave the range of the rarest frequency; is that right?
But you also, at least in your report provided in tabular form a whole range of different frequencies depending upon the racial composition of two people or three people or even four people who may have been contributors; is that correct?
Yes. I think all these numbers do that. All these ranges, as we've said, talk about the frequencies of the unknown people and the various databases we have available to us.
And so, for instance, when you are looking at--when you work on the assumption that there were two contributors and you provided a range for these different items of evidence based on the assumption of two contributors, you don't assume that you know the identity of the two contributors, do you?
All these frequencies are for unknown contributors. They are all calculated throwing away any idea we might have who contributed, so these are all unknown people. These numbers refer to the frequency of which we would find that evidence if they were--evidence profiles were given by unknown people.
And so what you actually did, in arriving at these calculations and these estimates, Dr. Weir, is that you would list all the possible pairs, all the permutations, if you will, that if two people created a particular stain, a particular profile, what possible genotypes could each of them have; is that right?
Well, it is even more than that, of course. We have, say--the simplest case we have four of these bands, four alleles, so we know it is coming from two people, so I assume, well, I'm going to do a calculation as though they were given by two people unknown to me. One person might have these two bands and be African American. This other person might have those two bands and be Hispanic. Or one person might have the first and third bands and be Caucasian. And the second person will have the second and fourth bands and be African American. So you can see that there is a whole host of possibilities we have to consider. Remember that these contributors, under this calculation, are unknown to us. We have no idea who could possibly have contributed those mixed stains, and so we have no idea, we've got to consider all the possibilities that there are.
Setting aside for the moment that additional layer of complication, if you will, dealing with different races of the components, one of the things you did do, at least initially, is you would sum the frequencies of the different pairs of genotypes; isn't that right?
And it was your intent when did you this, Dr. Weir, was it not, to sum up all the possible pairs? You don't want to leave any out?
And would you agree, sir, that the more pairs that you include, the more common the ultimate frequency is?
So it is essential, to be fair to the Defendant in any case or to Mr. Simpson in this particular case, that you include all possible pairs; isn't that right?
And if you left out some pairs, then your number could be biased against Mr. Simpson; isn't that connect?
Let me ask you a question, DR. WEIR: When were you first asked to prepare statistical estimates of mixtures in this case?
Well, didn't you receive a call from one of the Prosecutors back in the middle of May when Robin Cotton was on the witness stand about a request to provide statistical estimates of the mixtures in this case?
In fact, you issued a report--preliminary report as early as May 11, 1995, didn't you?
Yes, yes, that's correct. The time Robin Cotton was testifying the issue of mixtures came up.
Excuse me, doctor. Would you allow Mr. Neufeld to finish asking the question before you start answering.
Let's take a look at one of the tables, if you will, in the report that you submitted to the Court in this case. I'm referring to page 33, Mr. Clarke, which is table 24-B. Could you put it up on the elmo, please. May I have it put up on the elmo, your Honor?
10:30? You know something, I think that is the same mistake I made last time, same issue.
Dr. Weir, you said that in preparation for this report that you reviewed at least the final reports of the DOJ and Cellmark laboratories; is that correct?
And you also said that you made a number of phone calls, at least to Mr. Sims, to assist you in interpreting the data; is that also correct?
Now, in this case, Dr. Weir, both Gary--I'm sorry, are you somewhat familiar with how DQ-Alpha typing is done?
That is why you had to rely on what Mr. Sims and Robin Cotton told you; is that right?
In this case, Dr. Weir, both Gary Sims and Robin Cotton testified that since there is no specific dot on the DQ-Alpha strip for the allele 1.2, it is concluded that it exists through inference and other dots or the absence of other dots on that strip. Have you heard about that?
Yes. I'm not completely clear of the details, but I understand that there are cases where the 1.2 may be present but we can't tell whether it is present or not.
Well, didn't you notice, sir, on the reports that you read from DOJ, that whenever there is a mixture and the mixture included both a 1.3 and a 4 allele, the 1.2 allele may be masked?
Well, the records I think aren't on the board anyway, so 1.2 possible and I think yesterday that I have ignored these qualifications. If the 1.2 is listed, I have assumed that it is present.
Well, Dr. Weir, you said, I believe, yesterday, that there were some instances where an allele may be--a dot may be very faint, may be very weak and in those situations you ignored the description of weak or faint?
Well, in the situation, sir, where the 1.2 allele is simply masked in a mixture by the presence of these other alleles, as has already been testified to, it is possible, is it not, that when you describe the various pairs that you have to list the pairs as if the 1.2 allele was there and then list the pairs as if the 1.2 allele wasn't there; isn't that correct?
I'm not sure about that. I've assumed that it was there because on the board it was listed as being there, as possibly being there, and as I said, I have ignored the words "Possible" or "Trace" or whatever, so I've just taken my calculations for those four alleles.
Well, let me ask you this: Were you told by Gary Sims or by Robin Cotton, that when we say possible 1.2 in a mixture situation, we mean that there may not be any 1.2 there at all? Didn't they tell you that?
So it is your position, sir, that in this particular table, for instance, that the reason you did not consider the possibility of no 1.2 allele, is because whenever it is listed as possible you included it in the calculations as if it was there?
That's what I've done, yes. I've taken every allele listed on the board as being there.
MR. NEUFELD: Well, Dr. Weir, I call your attention to table--(Discussion held off the record between Defense counsel.)
If I may, I call your attention to--I would like to put up as a next exhibit, Dr. Weir's table 25-A and 25-B.
Now, in this particular table, Dr. Weir, given the statement results from the DOJ laboratory--
--as you have for table 24-B, the table I just showed you, in this instance you chose to accept the possibility that the allele, namely, 1.2 which is the possible allele, could be masked and therefore not be there?
Well, I don't know what you mean. The statement holds of course all high calculations include the possibility of it being there.
And your calculations also include, in this particular table, the possibility of it not being there; isn't that right?
That's correct. I have--I have a whole range of possibilities either with it being present or not present.
So for table 25-A, which were certain mixed stains on the glove, G1 and G4, you chose to include both the frequencies, given the assumption that allele 1.2 is there and the assumption that allele 1.2 is not there; isn't that correct, sir?
Whereas in the table--by the way, as a result of taking that approach, sir, in table 25-A, how many different pairs did you have to sum up for the DQ-Alpha type?
And--one moment. Just so the jury can be clear on this as well, are the lists of pairs that you summed up reflected along this first column?
Right. Well, it is a pair--this would be the first person and this would be the genotype of the second person, correct, (Indicating)?
And so you have 18 different frequencies that you sum up on this particular table for those particular items; isn't that correct?
And you did that in a situation where there is a possible 1.2 allele which may or may not actually be there because of the masking phenomena; isn't that correct?
And consequently, the 1.2 allele in that particular mixture may actually be there or may actually not be there; isn't that correct?
But this time, sir, unlike the other items on the glove, you chose not to include the frequencies of those pairs assuming the 1.2 allele wasn't there; isn't that correct?
Well, let's talk about what that mistake means, sir. In this particular instance, Dr. Weir, your calculations are based only on summing frequencies for six pairs; is that right?
And would you agree, sir, that if you sum the frequencies for an additional dozen pairs, which would be three times as many as you started off with, you would arrive at a probability for that mixture which would be much more helpful to Mr. Simpson; isn't that correct?
It is kind of interesting that the frequencies in both these tables are comparable in magnitude.
Sir, would you agree and didn't you say a little while ago on cross-examination, that if you add additional pairs of frequencies you will arrive at a frequency that is more common than if you add fewer pairs of frequencies?
--on G--item G1 is a stain on the glove where it is not suggested that Mr. Simpson could be a contributor; isn't that correct?
So he is not considered a possible contributor there; is that right? He is not listed in the "Not excluded" category?
Right. And so on your calculations for those stains that are mixtures where Mr. Simpson is excluded, you've summed up additional pairs to--which would make the frequency somewhat more common; is that correct?
I've summed the table 25-B is correct. It looks to me as though table 24-B is incorrect.
When I do calculations, I do not consider any forensic implications, and if you are suggesting that I do, I will disabuse you of that right now.
Now, for--unlike the items of evidence where you did sum up the additional pairs, sir, table 24-A reflects items 303, 304 and 305 in the Bronco; is that correct?
And Dr. Weir, for items 303, 304 and 305--sorry--on the console, Mr. Simpson cannot be excluded; is that correct?
And the calculations you did for the frequency of the mixtures for those stains fail to include the additional pairings; isn't that correct?
And by failing to include the additional pairings in these samples, in these items, which do not exclude Mr. Simpson, the numbers that are arrived at by you and put on that board are biased against Mr. Simpson; isn't that correct?
And sir, would you agree that you made this same error, which is biased against Mr. Simpson, not only on items 303, 304 and 305, but also on item 31 in the Bronco?
Am I right, sir, that not only did you make this same mistake which has the effect of being biased against Mr. Simpson for items 303, 304 and 305, but also as to item 31?
Well, this item assumes that all the alleles listed on the board are present and I have made calculations accordingly.
Well, but of course, as you know from what you did on that other item, sir, on the glove, you know that you can't necessarily make that assumption that you assume that it is either there or then assume it is not there to consider the other pairs or frequencies; isn't that correct?
Were you told about how the system functions, though, because of this masking phenomena?
All my analyses are based on the reported profiles, whether or not they match, and their details. I have generally ignored, although apparently sometimes I didn't, but in this case there is no suggestion that the 2.1 allele is not there, so I assume it was there.
Sir, but isn't it--when you put on your own report, with respect to other items, both on table 25-A and again on table 26-A, if you care to look at 26-A as well--
I believe in those items the word "Possible" was listed on the board and on item 31 it was not listed as being possible, so I assumed the allele is there.
Well, you put that asterisk on the report. Before you did that, sir, did you ask either Gary Sims or Robin Cotton what is this thing about the 1.2 allele, that it is sometimes masked and sometimes isn't, before you decided to put in that asterisk and make those different computations?
I didn't decide to put in the asterisk arbitrarily. I put it in--at least my intention was and at least in my original intention back in May, to assume it was either present or absent in those cases where the word "Possible" was written. The word "Possible" is not written on item 31.
So you are saying that the reason you didn't do it on table 23-A is because the word "Possible" wasn't written in next to the 1.2 allele?
That is the only reason that would be consistent with what I did, and I assumed that is my reasoning at that time.
Well, now, when you say you assume that reasoning, do you know that to be true or are you guessing on that?
Well, I'm saying it is true. I can't think of every thought process I've had over the last several weeks, but that would--thinking about it now, and the way I do operate, I'm sure that is what I did.
Now, can you see, Dr. Weir, that for this particular table you were computing the frequencies of the mixture both for item 31 and I believe also for one of the glove stains; is that correct?
And is that what the sentence says at the very top there, sir, the top of the table?
And sir, on the report that you received from the Department of Justice with respect to item G10 which you have had front of you, I hope--sir, do you have that report?
No, I don't. I should say I based my calculations on these charts; not final reports.
Sir, did you tell us a little while ago that you actually reviewed the final reports of the Department of Justice and Cellmark?
Yes. Those statements are not inconsistent. I certainly reviewed them. I don't have them in front of me. All I'm saying is that after having reviewed the report, when I came to sit down and type in entries into the program to do these mixture calculations, I relied on these charts.
Isn't it a fact, sir, that with respect to the report from DOJ regarding stain G10 on the glove, which also is listed as not excluding Mr. Simpson, that report from DOJ told you specifically that the 1.2 allele was just possible; isn't that correct?
All right. So here again, Dr. Weir, you did the same calculation for G10 on the glove, a stain that allegedly cannot exclude Mr. Simpson? Even though you knew that the 1.2 allele was simply a possible allele, you did the same calculations, which as it turns out, are biased against Mr. Simpson; isn't that correct.
Isn't it correct, sir, that in your calculations for G10 you only summed the frequencies of six pairs of genotypes?
That's the way I have reported it. In fact, I didn't do any calculations on G10. I just lumped them in with the 31.
But your report, sir, says this is the correct data for item 31 and G10, does it not?
Yes. And in fact the correct way to do G10, if the 1.2 allele is simply possible, is to have the additional twelve pairs that you had on those items where you considered the 1.2 allele is possible; isn't that correct, sir?
I think you found my mistake, Mr. Neufeld.
As it turns out it looks that way, yes.
I became aware about 8:00 p.m. last night.
I can't be held accountable for every coffee conversation I may have ever had.
That would be wrong.