Here is a copy of a letter from attorney Bill Risner to Attorney General Terry Goddard explaining why he should continue investigating Pima County 2006 RTA election. Risner sent the letter on July 14.
Dear Mr. Goddard:
I sent you a short letter on July 9th, 2008, together with Mr. Zbigniew Osmolski’s Affidavit.
I will be out of the County from July 15 through the end of the month. Accompanying this letter are various materials that may help you to better understand the nature of the allegations and more fully understand the past investigation by your office staff.
At the beginning of the database lawsuit, the Pima County Democratic Party, and I personally, had confidence in your Office’s integrity. Additionally, I was sensitive to political currents. That is why I informally told Jim Walsh what we were finding out in our lawsuit against the Pima County Board of Supervisors. It was a “heads up” conversation relating to him that we were acquiring evidence suggestive of criminal activity but not enough in my opinion at that point for your office to open an investigation and none was requested.
Later, attorneys for the Board of Supervisors forcefully suggested that I was obligated to make a criminal complaint if I believed crimes had occurred. At that point, I made an appointment with John Evans of your Office who agreed to open an investigation. The “suspects” were listed on your office form as the “Pima County Election Division.”
The Pima County Democratic Party offered technical expertise. Your office chose not to accept our technical expertise and we did not complain then nor do we complain now about that decision as your office can investigate in the manner that you choose.
I subsequently had a conversation with Mr. Evans in which I asked him what our role was in the investigation. He said it was a “one way street in which he could not give me information but he could receive information from us.” I then gave him the names of two witnesses including Robbie Evans, Jr., who for four years was the computer assistant to Bryan Crane. I explained that Mr. Evans, Jr. would testify that Mr. Crane regularly printed unofficial tallies or summary reports of actual votes before election day. Your Office investigators chose not to interview that witness, even though they knew his testimony would contradict Mr. Cranes’ prior testimony. Instead your investigators accepted Mr. Crane’s fourth different under oath story without comparison with the prior explanations nor did they question any contradictory witnesses.
During a subsequent conversation with Mr. Evans, I learned that your offices’ report from iBeta would be provided to the suspects, but a copy would not be provided to the Democratic Party, although Mr. Evans concluded the report would be a public record, he said he would require us to retain a copy from the County suspects. I have attached several of the letters that I subsequently sent to John Evans.
I am sure you are now aware that your office joined with the suspects in a joint study, permitted the suspects to direct the investigation and gave them a copy of the investigative report before conducting any interviews. Before commenting on the iBeta report, I would like to review the background of the decision to proceed in that manner. Mr. Evans had initially contacted Michael Shamos, a nationally known voting systems expert at Carnegie Mellon University. Mr. Evans and Mr. Shamos’ e-mails are attached. Mr. Shamos immediately recommended the ballots themselves be examined as he said: “Ultimately the proof of the pudding is in the ballots.” “My suggestion would be to re-tabulate from the original records. This should tell us very quickly whether the GEMS results were fudged. What is the difficulty with this approach?” Indeed!
Mr. Evans response was:
“As for the white wash, I would agree with you but the party to the civil law suit that discovered this problem is very much on board. They want the data base to be looked at and they have approved the scope of the project. The most vocal local naysayers have bought into this process.”
Mr. Evans was completely wrong. We had not “bought into this process.” He insisted on this process. Nevertheless, Michael Duniho, on behalf of the Democratic Party, strongly suggested that the ballots he examined. Mr. Duniho recalls a heated exchange with Mr. Evans.
Our deference to your office’s integrity at that point should not be characterized as being “on board” Mr. Evans’ flawed process.
Mr. Evans’ e-mail also contained this important reference to the “issue to be investigated.”
“Regarding your questions, the initial issue is about the absentee ballots that were run before the joint summary report. The next question is whether after the summary report there was a flip of the fields. So the accuracy of the absentee ballots is questioned and the accuracy of the subsequent ballots may be an issue.”
The evidence to resolve that key question was already available to the Attorney General. A.R.S. § 16-445 required Pima County to send “at least ten days before the date of the “RTA election” a copy of the ballot layout. In other words, the position of how the computer would read “yes” and “no” votes was on file. If the computer had later been instructed to read those votes reversed or “flipped” so that “no” votes would count as “yes” votes the computer data could easily have been compared with the data on file with the Secretary of State.
A.R.S. § 16-445D specifically provides that the data on file “shall be used by the Secretary of State or Attorney General to preclude fraud. . .”
In other words, the entire purpose of that data was for it to be examined in a fraud investigation by the Attorney General. Your office did conduct a fraud investigation where that evidence would have provided the answer, but it was neither used nor requested by your office.
Furthermore, your office actively attempted to obstruct the Democratic Party’s attempt to find that evidence, when the Democratic Party scheduled a deposition of the Secretary of State’s office. Your office filed a Motion for a Protective Order asking the trial court judge to prevent us from learning the whereabouts of that evidence. We ultimately prevailed over your office’s objection and learned it had been mailed back to Pima County where Brad Nelson personally handed the critical evidence to Bryan Crane, and it has not been seen since. The Arizona State Election Director, Joseph Kanefield, testified that the Secretary of State’s office was aware of the criminal investigation having been informed by your office.
As for the iBeta “investigation” jointly conducted by the suspects and your office, it is clear to us that the investigation was steered by the suspects down blind alleys. The statement of work written by iBeta contained no reference to either swapping ID codes or replacing a database with one modified on another computer. During the investigation, the suspects’ technical defense person, John Moffat, suggested the investigative contractor engineer look at the Preferences table in the database to see if the programming had changed, and also to back each batch of early ballots scanned out of the database to see if vote totals had been changed. But the simplest manipulation of the election database, swapping the codes that identified the Yes and No votes, would have been done in the Candidate tab and swapping the codes would not have changed any vote totals – it would have merely reassigned the votes. Needless to say, the investigative contractor engineer found no conclusive evidence of tampering – either because it did not know where to look or because he carefully avoided looking where tampering was likely to have occurred.
The iBeta report discusses five “tests.” Test 1 produces no useful information. Test 2 did turn up what appeared to be evidence of “tampering,” but the company accepted John Moffat’s explanation. Test 3 confirmed “five copies” of the test target file were identical. This was not a useful conclusion as the key issue was data that had been erased. Test 4 was a test “prepared” by John Moffat concerning the “Preference table.” I have previously noted the uselessness of that test. Test 5 was also “prepared” by John Moffat, and again, was a test not directed toward the allegations. That test was whether votes had been externally added which has never been an allegation.
John Moffat is paid $184,000 per year by the County for a 30 hour week. He works part-time, so he can continue to run a separate company he owns. He reports directly to Charles Huckelberry on an “oral” basis only. Since competent evidence, such as the Osmolski Affidavit, quotes Bryan Crane as saying that he was told to fix the election by his bosses, it is clear that County management has a potential motive to obstruct an investigation. An assessment of John Moffat’s role in your investigation, and in the civil case, indicates that his role has been to prevent an examination of past election practices. At a recent meeting of the Pima County Board of Supervisors, John Brakey reported that John Moffat said he would cooperate with the Democratic Party in the future if we would agree not to look into the past.
A press report this week quoted Bryan Crane as saying he had to look up on a map where the Boondocks Bar was located. However he got there, he was seen that evening by another available witness who knows Mr. Crane. Mr. Osmolski related his conversation with Mr. Crane to four separate people at the bar that evening.
The truthfulness of Mr. Crane’s confession can readily be determined by examining the ballots. As noted by Michael Shamos, the proof is in the ballots. The likelihood that the RTA election was fraudulent can also be inferred from the totality of the circumstantial evidence. The circumstantial evidence is strong. I have already mentioned that Mr. Crane received from Mr. Nelson the RTA pre-election tape sent to the Secretary of State’s office pursuant to A.R.S. §16-445. The box delivered by Mr. Nelson contained several tapes but only the May 16, 2006, RTA tape has disappeared. An inference can be drawn from the disappearance of computer data that has the specific ability to prove the crime by contradicting saved data.
The motive of the “bosses” could not be clearer. The proposal that a sales tax be approved for roads was defeated on some four prior occasions. The May, 2006, proposals were unanimously endorsed by all five supervisors. Supervisor Valadez was the RTA Chairman.
Months before the RTA election, the Board of Supervisors hired James Barry, a Special Assistant County Manager, to work under the direction of Chuck Hucklelberry, and develop a computer database of all previous County board elections by precinct to determine precinct by precinct voting patterns. Mr. Barry’s contract began the day after his retirement from the County. Mr. Barry was paid $75,000 for that work. At the same time, Mr. Barry received approximately $12,000 from the RTA Yes Committee for “consulting.”
The RTA was said to have passed by a surprisingly large margin. Yet the RTA Yes group was privately claiming in the weeks leading up to the election that their tracking polls showed the measure likely to lose.
A Microsoft access manual was seen and photographed in the vote tabulation room on election night. Use of MS access on an election computer was and is illegal.
The Chair of the Pima County Democratic Party requested days after the RTA election day for a party consultant to enter the tabulation room accompanied by Election Director Brad Nelson for the sole purpose of looking at the cables attached to the election computer. The request to enter the vacant room to see if another computer might have been connected to the election server was denied. This request occurred while all parties were present in a room next to the vacant room.
Chester Crowley, an election department employee, testified at trial that the election computer had in the past been connected to Bryan Cranes’ computer in his office and he believed Mr. Crane had printed unofficial tallies on his office printer directly from the election computer.
Mr. Crane’s assistant for some four years, Robbie Evans, Jr., testified that Mr. Crane regularly took home during elections a compact disc (CD) of election data. Isabel Araiza, perhaps the election division’s senior employee and the office manager prior to Brad Nelson being hired, testified that she had discussed with Brad Nelson the security problem of Bryan Crane taking election data home with him during live elections. Mr. Nelson did not object to the practice and did not instruct Mr. Crane to cease that practice. The GEMS system has a well-known security defect known as “the back door” whereby data can be changed using Microsoft Access without knowing or using a password. The GEMS audit log is not separate from the data itself. That means that election data can be changed and then the audit log itself can be amended to erase any history of the changes having been made.
The audit log for the RTA election shows evidence consistent with just that kind of manipulation and inconsistent with the normal operation of the GEMS software. Since Bryan Crane operated the GEMS software for ten years before the RTA election his normal style is known.
The May 10, 2006, audit logs demonstrate the normal operation of the ballot counting. On that day, election employees counted more than 13,000 early ballots over a four hour period. The vote total data from those ballots was backed-up and labeled as Day 1 back-up. If a CD of the election data had been made, it would not have shown on the audit log. Testimony has confirmed that the making of a backup CD was his normal practice.
The number of persons who could observe inside the counting room was severely restricted in the months just prior to the RTA election. Brad Nelson radically changed prior procedures so as to prohibit employees that previously had access to the counting room from doing so during the RTA election.
Bryan Crane was quite familiar with the ability of the GEMS system to export data and manipulate it off line. He had done so in 1996 at the instructions of Chuck Huckelberry. During one of Mr. Crane’s depositions, I asked him about 1996 at which point the County Attorney’s Office stopped the deposition and attempted to reach the trial judge to prevent any questions about that off line activity. The deposition continued only when I agreed to not ask any questions at that time about 1996.
The audit log of May 11, 2006, shows that thirty-three seconds after the election computer was opened that morning, Bryan Crane created a second “Day 1 backup” and erased the prior day’s data, replacing it with a new “Day 1 backup.” This action would be similar to your experienced secretary backing up a brief she was preparing for you before going home and then seconds after coming to work the next day again “backing-up” the brief when no additional charges had been made. Such an event is highly unlikely. Bryan Crane’s normal practices are known. The audit logs show that he backed up vote totals only after ballots were counted. Precisely what one would expect.
At his deposition, Mr. Crane had no explanation for the new Day 1 backup nor for the two separate unofficial tallies did he print thereafter. At trial, he was questioned by Deputy Pima County Attorney Chris Straub and explained that the writing over of the data had been a “slip of the finger on the mouse.” That explanation cannot be true, however. That is because the overwriting and destruction of the day one data required responding to two warning messages, one from GEMS and one from Windows. A box would have appeared on his screen that said a day one backup already existed and did he really want to wipe out that file and create another one with the same name. Such a sequence from a ten-year veteran of that system is unlikely in the extreme.
What the audit log evidence is consistent with is the re-insertion of a new data with reversed data. If Mr. Crane had taken home a CD of the election data he could have examined that data at home and reversed the votes of no to yes. That simple change would then cause GEMS to automatically make multiple changes. For instance, the computer would automatically change all four hundred or so precinct totals to match the new reversed count. Additionally, the computer would count all future No votes as Yes votes as its instructions would have been changed. As previously noted, this sort of election fraud is precisely why the ballot data was filed pursuant to A.R.S. §16-445 for the use of your office in a fraud investigation.
At the end of each election day, the data was normally backed up on the computer. One would expect that such a backup would be made since if it is wise to back up each day’s counting of early ballots. It would certainly be wise to back up data from each precinct that came in after the close of the polls. And that is the normal pattern. Before the election staff goes home on election night the audit logs show that a backup is made of that data.
Except for the RTA. For the RTA, such a backup was not made. This failure is a very significant departure from normal practice and suggestive that vote total manipulation was occurring off line. The data was not backed up “until three days later,” after the results had been published.
In any sophisticated computer crime, the variance from normal patterns offers clues as to what has occurred. Those clues are referred to as “badges of fraud” in the case law.
In the database lawsuit filed by the Pima County Democratic Party, the Board of Supervisors’ lawyers filed a pleading stating that they could not adequately defend the lawsuit because of the substantial risk that every employee who operated the Pima County Election computer would assert his or her Fifth Amendment privilege not be incriminate themselves. Such a written confession by the County’s lawyers is unprecedented to my knowledge in this country.
Neither that admission nor any other admission of violation of rules, criminal laws or good practices has resulted in any inquiries by County management or even a reprimand of any election department employee. John Moffat testified that he had been instructed by the County’s lawyers not to ask questions of Mr. Crane about violations of law relating to the printing of summary reports.
Joe Kanefield testified that he assumed the county had itself examined such allegations as would any organization or company. His assumption is the same as ours. Therefore, the total organizational failure to do so speaks volumes to the necessity of an outside review and clearly suggests that the management of the organization is complicit. In other words, it supports Mr. Crane’s statement to Mr. Osmolski that he fixed the RTA election on the instructions of his bosses.
Ten months prior to the RTA, the Pima County Election Division, at the request of Bryan Crane, purchased a “crop scanner,” a read-write device that is a computer hacking tool. That tool has no other purpose than to illegally alter the programming of precinct voting machines. Actually, it does have a legal use, but I am certain the election division was not using it to know when to irrigate their crops.
The Pima County Democratic Party’s election integrity Committee has an unusual number of individuals with extensive computer and election computer expertise. Dr. Tom Ryan, PhD. is a retired computer engineer who has been studying computer election issues for several years. The Pima County Democratic Party adopted a report he wrote in April 2003 concerning election computer problems. James March is a member of the Board of Directors of Black Box Voting, a National Organization of citizen election reform advocates. He was one of the first computer technicians to examine the Diebold GEMS software. He has been consulting with the Democratic Party on election security issues.
Michael Duniho (“Mickey”) has retired to Tucson from a career with the National Security Agency where he was one of fifty “master programmers.” He has spent innumerable hours learning election and ballot processing procedures. John Brakey, another computer, expert, is self-taught, but has an excellent grasp of the GEMS system and its potential use in fixing an election.
All those informed individuals are in agreement that sufficient questions exist to merit a hand count of the RTA ballots.
All of our freedoms in the Untied States are ultimately guaranteed at the ballot box. Anything less than an honest count of ballots is a crime that strikes at the heart of our Democratic system.
All of us who have been active on issues related to election security believe that the ballots for the RTA must be preserved and counted. Only you, as Arizona’s Attorney General, can take control of the ballots as potential evidence of a crime and count them.
Our community, your political party, and our core freedoms, will be protected only if you act to determine whether a major crime has occurred against the Democratic process. The issue is not the fallout of that crime but whether the crime has occurred.
Pima County management now asserts that they want the RTA ballots preserved, but they want a judge to tell the Pima County Treasurer what to do with the ballots. The ballots can be preserved and counted only if Arizona’s Attorney General does the job he is required to do.
The obligation to determine if a crime has occurred is not for the Democratic Party. The political party is not a prosecutorial agency. It has been involved in order to preserve its core role of election observation. The prosecutor’s role is yours. Whether or not a crime has occurred can be simply and definitively determined through an examination of the ballots. We ask for you to personally direct that the current investigation be conducted in such a manner as to arrive at an answer that the people of Pima County can accept.
Very truly yours,
William J. Risner