Matthew DiSogra - Direct

MR. DISOGRA: Yes, I do.

MR. DISOGRA: Sure. Matthew DiSogra. Last name D-I-S-O-G-R-A.

MR. DISOGRA: I am a licensed professional engineer. I practice in the field of accident reconstruction and within that field I specialize in analysis, interpretation, and accessing all types of vehicle data.

MR. DISOGRA: I'm employed by Delta V Forensic Engineering, which is an engineering firm that specializes in accident reconstruction — all parts of accident reconstruction.

MR. DISOGRA: So accident reconstruction is a broad field. There are a lot of subspecialties within that. Accident reconstruction is everything from the initial investigation phases of a crash all the way through testimony, like here today. Some of the specialties within that could be things like biomechanics, could be human factors, and my specialty is vehicle data and event data recorders.

MR. DISOGRA: I am the director of engineering for our EDR services line.

MR. DISOGRA: Event data recorders. So EDR is a general term we can use to talk about all types of data from all types of vehicles.

MR. DISOGRA: It's — I wear a couple of different hats. First and foremost, I'm still a practicing engineer. I still conduct investigations. I still perform analyses, and I testify. On top of that, in the director role, I oversee a group — it's not just me doing this kind of work. I have three other engineers that work with me, all specializing in EDRs. So I sort of oversee that group and focus on kind of the day-to-day operations as well as the direction we're going — you know, with things like autonomous vehicles coming, the role of data is growing. So we're kind of looking to the future in a lot of those ways.

MR. DISOGRA: I began my career at Delta V doing traditional accident reconstruction in the broadest sense, and as I have progressed I have sort of narrowed myself down, and now specialize in that —

MR. DISOGRA: Ultimately, anytime we're doing EDR data analysis, it is in support of an accident reconstruction.

MR. DISOGRA: That's right. There are a lot of different ways data can be accessed, all the way from just plugging into a port on a working vehicle, all the way through — when a vehicle is catastrophically damaged and even the modules themselves are damaged — accessing the data at the chip level. So no matter what condition a vehicle or the remnants of a vehicle are in, generally we can still get the data from that.

MR. DISOGRA: I do.

MR. DISOGRA: Beyond Toyota being a common vehicle in general, we run into that data quite frequently. Just this year we published a paper specifically studying some of the behaviors of the additional data recorder that's in Toyota vehicles.

MR. DISOGRA: Lexus — I believe you're — so Lexus is a Toyota brand.

MR. DISOGRA: It's kind of a combination of a couple of words — information and entertainment.

MR. DISOGRA: So that's a generic term which more or less means the thing that's in the middle of the dashboard with the screen on it. A lot of manufacturers implement that a little bit differently, so they all look different, but it's a general term we use in the industry to describe that system.

MR. DISOGRA: GPS is a pretty fundamental part of vehicle data as a whole at this point, and we run into it in a lot of different places. We see GPS data from infotainment systems — so we're analyzing that data, most of that's GPS-based. I do a lot of heavy truck work, and almost every heavy truck has some kind of GPS system in it for the fleets to track them. So we're analyzing that as part of data. And then the last area would be when we conduct testing, any kind of research, we're generally going to put our own GPS systems on vehicles so we can study that data. So analyzing GPS is a regular part of my job.

MR. DISOGRA: Global positioning system.

MR. DISOGRA: That's correct.

MR. DISOGRA: What is instrumented testing as it relates to Toyotas and VCH testing and things of that nature? Can you explain that for the jurors?

MR. DISOGRA: Sure. What instrumented testing means is — let's say we want to study how accurate some data is that comes out of one of these vehicles. So we need an independent source of that information. So what we'll do is we'll equip a vehicle with one of these GPS data acquisition systems. Basically we can drive the vehicle around and we know exactly where it is, how fast it's going, all these parameters. We can do some maneuvers with the vehicle. Then we can download the data that the vehicle stored, and now we can compare the two and say, okay, well if we have really high confidence in our system, how accurate is the vehicle system, because the reality is there's always some small errors in these vehicle systems, and in a lot of cases what's important is quantifying that error.

MR. DISOGRA: That's correct.

MR. DISOGRA: Infotainment specifically, probably around 30.

MR. DISOGRA: Yes.

MR. DISOGRA: Clock alignment would mean that anytime we have a timestamp in any set of data, a singular clock value, there's a good likelihood that that's not 100% accurate. So clock alignment would mean finding an independent source of time to compare that to, preferably one that's more reliable, and then making the necessary adjustments so that time can be considered real time.

MR. DISOGRA: Thousands. Over a thousand. That's a regular part of data analysis. Any data set that has a timestamp — it's normal. It's part of the analysis. It's really the first step: let's get that clock aligned. So it's something I'm doing regularly.

MR. DISOGRA: It'll be my 13th year anniversary next month, actually.

MR. DISOGRA: Correct.

MR. DISOGRA: Certainly. So I began my academic career. I got an undergraduate degree in mechanical engineering from the University of North Carolina at Charlotte. I then went on and got a master's degree also in mechanical engineering from the same school. And then when I came out of school, I immediately started working for Delta V Forensic Engineering. In order to get a professional engineer license, I then needed to acquire four additional years of work directly under the supervision of another licensed engineer. So then I completed that phase of work, and then in order to maintain my license, I have to complete continuing education every year in the field that I practice.

MR. DISOGRA: Correct. Bachelor of Science.

MR. DISOGRA: 2011.

MR. DISOGRA: That is correct.

MR. DISOGRA: 2013.

MR. DISOGRA: So, in the state that I'm licensed in, North Carolina, they define that really as any activity you're participating in that's keeping you abreast with the field, keeping you up to date with what's going on. So taking formal classes counts for continuing education, but also some of the other work that I do, like publishing research in peer-reviewed journals, also counts as continuing education.

MR. DISOGRA: Yes.

MR. DISOGRA: Correct. Yeah. I attend — there's a couple of regular conferences in our industry specifically tailored to EDRs as a whole for passenger vehicles and heavy trucks. I'm a regular attendee of those. I'm also a regular speaker at most of those. And I'm usually presenting some of the new research that we have for that year.

MR. DISOGRA: Correct.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Correct. All that's accurate.

MR. DISOGRA: Correct. That's correct.

MR. DISOGRA: Yes.

MR. DISOGRA: IPC is an organization that offers certifications in basically soldering. I could say it in more technical ways, but it boils down to soldering. So as it relates to what I do, when we are recovering data from damaged EDRs at the chip level, that involves some really precise soldering work. And so IPC offers courses in how to do that at basically the highest level. And then they offer certifications for that.

MR. DISOGRA: I am an active member.

MR. DISOGRA: Yes.

MR. DISOGRA: I am. I'm currently an instructor of that course, which is offered through the Society of Automotive Engineers.

MR. DISOGRA: I am. I will be in Wisconsin next week teaching that very course.

MR. DISOGRA: I am. I am a voting member of the J2728 subcommittee, which is a subcommittee for event data recorders on trucks and buses.

MR. DISOGRA: I am.

MR. DISOGRA: It does, and it requires, as far as I know, when I joined, sort of proof that you are active within the field of accident reconstruction.

MR. DISOGRA: Correct. Correct.

MR. DISOGRA: That's correct.

MR. DISOGRA: Yes.

MR. DISOGRA: Specifically, this year we published a paper where we studied the effects of collision-induced power loss on the way data is recorded in Toyota systems, and that would include Lexus systems. Fundamentally they're the same components. It's just a different brand on top.

MR. DISOGRA: Yes, I was the lead author on that paper.

MR. DISOGRA: Sure.

MR. DISOGRA: There are differences of course between the data that comes out of passenger vehicles and the data that comes out of heavy trucks, but there are also a lot of fundamental similarities. Like for example, they all record vehicle speed. They all have accelerator pedals. They all have brake pedals. So a lot of the core data we're analyzing is very similar.

MR. DISOGRA: Yes, I did.

MR. DISOGRA: In that research, we introduced a new technique for getting data out of damaged modules. So, you know, they're either burned, they've been submerged in water, they're crushed physically from impact. Uh, we demonstrated a new method to be able to get data less invasively, but still at the chip level. And we also demonstrated that we could translate that data from the chips outside of the software that was provided. So, you could use the software and get the data that way. We demonstrated we could basically do it ourselves if we wanted to.

MR. DISOGRA: Yes.

MR. DISOGRA: Once.

MR. DISOGRA: Uh, in New York.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: All of my analysis is fundamentally based on all of the data and the numbers provided in the Aperture reports. So, uh, more or less the answer to your question is no.

MR. DISOGRA: Correct.

MR. DISOGRA: Correct.

MR. DISOGRA: Yes.

MR. DISOGRA: Uh, I analyzed the methods that Mr. Burgess proposed for aligning the clocks in that report.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: The event in order to be generated has to be triggered by something. So yes.

MR. DISOGRA: Correct.

MR. DISOGRA: For which event?

MR. DISOGRA: I do, your honor.

MR. DISOGRA: Correct.

MR. DISOGRA: Correct.

MR. DISOGRA: The trigger is a singular point in time.

MR. DISOGRA: Correct. It's a singular point.

MR. DISOGRA: Yes.

MR. DISOGRA: Sure. So when an event is triggered, what we end up with is some data from before the event was triggered and some data from after the event was triggered. The way it does that is it's kind of always recording and as soon as it sees a trigger it says aha and it goes and it grabs the previous five seconds. That number can vary. It can be more or less than five, but uh for the specific trigger we're talking about it grabs the previous 5 seconds of data and then it keeps recording for an additional 5 seconds and that completes its recording phase.

MR. DISOGRA: It wouldn't be accurate to say the trigger lasted that long.

MR. DISOGRA: The recording lasts that long, right?

MR. DISOGRA: Correct.

MR. DISOGRA: That's correct.

MR. DISOGRA: Correct.

MR. DISOGRA: Correct.

MR. DISOGRA: Customarily when we're doing any analysis we're referencing everything off of the trigger. Um, that's the normal way. Everything in the Aperture reports uh was referenced off of the end of the event. So, of the recording event.

MR. DISOGRA: Correct. And I didn't correct that. I just went with that convention for all of the analysis I did. But you noted it.

MR. DISOGRA: It would be larger by 5 seconds approximately.

MR. DISOGRA: Correct.

MR. DISOGRA: You did.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Um, regardless of what adjustment is made or which two points are used to try to correct the clock, the phone lock event always occurs after the end of the event — using their convention of referencing off the end of the event, not the trigger — but it always occurs after.

MR. DISOGRA: Correct. Correct.

MR. DISOGRA: I'm talking about the January report.

MR. DISOGRA: Yes, I was.

MR. DISOGRA: Uh, slightly after he issued it, but I don't know the exact date. It was issued — his report was dated May 8th and I was — I think I was provided it a couple days after that. I don't remember the exact date.

MR. DISOGRA: No, I don't know the exact date.

MR. DISOGRA: There was not.

MR. DISOGRA: Yes.

MR. DISOGRA: I did.

MR. DISOGRA: The purpose of the PowerPoint was really to um take a bunch of different data elements that were kind of placed throughout Mr. Burgess's report and Dr. Welcher's uh report in PowerPoint and compile them into a single source, a single table essentially to show as clearly as possible what the conclusions would be if we look at the data as a whole.

MR. DISOGRA: Yes.

MR. DISOGRA: No.

MR. DISOGRA: I received an amended PowerPoint by Mr. Burgess — not from him though, from you.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: There was.

MR. DISOGRA: In the first report, Mr. Burgess did some of the calculations to that are necessary to correct or adjust the times based on certain approaches, but never went through the final step of applying those calculations to the times and coming up with a final time. In the supplemental report, he then essentially took that analysis to the end. He approached it in a new way, a different way than he had previously laid out in the first report. But then he also took it all the way to the end by applying those calculations to the times.

MR. DISOGRA: In the original report, he laid out some methodologies for correcting the clock using phone calls that were shared between the vehicle and Mr. O'Keefe's phone. And in the supplemental report, he moved away from the phone call analysis and moved to attempting to align the clocks using a shared maneuver between the vehicle and the phone.

MR. DISOGRA: It was a three-point turn.

MR. DISOGRA: He did not have that analysis in the January 2025 report.

MR. DISOGRA: So I did the exact same process using the supplemental report as I did on the first report. And when I ran all those numbers and those calculations, it turned out there are now some scenarios in which the phone lock occurs after the end of the event. There are some scenarios where the phone lock occurs simultaneously — like literally at the same second — and there are some scenarios where it occurs before. So now there's a range.

MR. DISOGRA: Correct.

MR. DISOGRA: Yes.

MR. DISOGRA: I needed to include in my analysis the approach and the conclusions from the supplemental report by Mr. Burgess.

MR. DISOGRA: This is a printout of the PowerPoint presentation that I prepared.

MR. DISOGRA: Yes.

MR. DISOGRA: It appears to be.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: That is the title page of the PowerPoint that I prepared.

MR. DISOGRA: So this PowerPoint is largely identical to the first PowerPoint I prepared in that it seeks to take all of the different approaches between the initial Burgess report, the supplemental Burgess report, and Dr. Welcher's report — lay out all of those approaches in one coherent fashion so we can look at what conclusions we can draw from all of that data.

MR. DISOGRA: This is a slide from Dr. Welcher's presentation that I was provided. Specifically, this is his slide number 33.

MR. DISOGRA: This is where he introduces the first acknowledgement that there's a difference between the time that Aperture calculated for the end of event 1162-2 and the lock event on Mr. O'Keefe's cell phone.

MR. DISOGRA: It's up on the screen. I added the two green boxes and the text — the word "after." That's not Dr. Welcher's. And I added that because it's stated here that the event occurred within 26 seconds. But if we do the math — and he actually does the math — he really means "after." It means the phone lock event occurred 26 seconds after the end of the vehicle event as calculated.

MR. DISOGRA: 26 seconds.

MR. DISOGRA: 1162-2.

MR. DISOGRA: Right. I added the green box — the dashed green box — around the whole paragraph, the line, and the words "after." Those are my additions.

MR. DISOGRA: This is more or less a repeat of what's on the first slide but just shown with a little more focus. The first slide is a little busy.

MR. DISOGRA: I created everything on this slide apart from the top right, which came from Dr. Welcher's slide 33.

MR. DISOGRA: All of the data on this slide and all of my subsequent slides comes from the Aperture reports collectively — either Dr. Welcher's report or slides, or Mr. Burgess's first and second reports.

MR. DISOGRA: So in their analysis — in Aperture's analysis — they report the Lexus event 1162-2 ends at 12:31:43, and they also report there's a phone lock at 12:32:09. So if we just simply take the difference in those two times, we get 26 seconds.

MR. DISOGRA: 31 seconds.

MR. DISOGRA: If we wanted to conduct the analysis relative to the trigger, it would be better to use the actual trigger time. However, as long as we're consistent through the analysis, it's not entirely wrong to use the end. It would be wrong if I suddenly switched up and started talking about this relative to the trigger event — but they present it as the end of the event. So I conduct the rest of my analysis using that convention.

MR. DISOGRA: The end of the recording. Yeah. When I say "event," what I mean by that is the data event — the data recording event — but the trigger event itself is 12:31:38.

MR. DISOGRA: Yes, I did. For the sake of the rest of the analysis that I'm going to do, I thought it would be helpful to just refer to this difference in time as "John's interaction window," and that refers directly to the interaction with the phone — the human interaction of locking the phone — the 12:32:09 lock.

MR. DISOGRA: So here what I'm doing is laying out the bounds of what I mean by "John's interaction window," which should help — we have a lot of numbers to go through here. And so this should give us a very easy metric for determining whether we have an event before or after. What I mean by that is: when we calculate John's interaction window, if that number comes out positive, that means the phone interaction occurs after the end of the vehicle event. And if we calculate John's interaction window and that number becomes a negative value, then that means the phone interaction occurs before. So now we have just a convention we can use really quickly to say is it positive or is it negative? And we can maybe spare ourselves some of the math with a thousand numbers.

MR. DISOGRA: Positive means the phone lock occurs after the end of vehicle event 1162-2.

MR. DISOGRA: So, what I'm saying here, I'm kind of laying the foundation for why we're going to even do this. Why do we need to start doing math on these clocks? So, fundamentally, the phone — Mr. O'Keefe's phone — and the vehicle — Miss Read's vehicle — they're separate clocks, and it's very unlikely that they're going to be aligned perfectly in time. Right? If I compared my watch to your watch, it would probably be off. And that's the same case for any two independent clock sources. So, Aperture in their collective reports propose a number of different ways in which that can be corrected. So, I'm going to go through all of those ways that Aperture has proposed and show if we align the clocks based on how they laid out, does John's interaction window stay positive or does it ever become negative?

MR. DISOGRA: Where does it end up when we run all those analyses? Or another way to put it is: does the phone interaction always occur after the end of the vehicle event?

MR. DISOGRA: Yes, the phone interaction always occurs after the end of the vehicle event.

MR. DISOGRA: In the PowerPoint as a whole, all the information comes from the Aperture reports collectively.

MR. DISOGRA: Yes.

MR. DISOGRA: Yeah. So, this is the first adjustment we're going to talk about — and this one comes from Mr. Burgess's original report. And to summarize this adjustment, this is a key-on timing adjustment. And we're going to show that the key on — the actual turning of the vehicle on — occurs about 3 seconds before that time gets registered in the infotainment. So the time we see is basically stale by 3 seconds. We're all familiar with how computers turn on.

MR. DISOGRA: In this case, it's a button. I refer to key on because generally that event is considered the key on in vehicle parlance regardless of whether there's a physical key or it's a push button.

MR. DISOGRA: It means that when we're trying to use the timestamps that were recovered from the infotainment system that show when the vehicle was keyed on or turned on, that the vehicle was actually or likely turned on 3 seconds before that.

MR. DISOGRA: That came from the testing that Aperture conducted with the exemplar Lexus that they purchased.

MR. DISOGRA: No.

MR. DISOGRA: It could be different. In their testing, they only tested one exemplar vehicle. So I think there was a solid foundation that the exemplar that was used exhibited this 3-second delay. It was a very close match in terms of part numbers and software versions to Miss Read's Lexus. But without any kind of testing of — let's say — a second vehicle, a second exemplar, we wouldn't know. We wouldn't have any basis to say it must be three. We could test a second vehicle and it could be five. So we say, well, there's a delay, but it varies vehicle to vehicle. They could have tested a second vehicle and found it's also three. And then you start to build some confidence that maybe it is three across every vehicle.

MR. DISOGRA: But with only testing one and not testing the subject vehicle, it is difficult to say with absolute certainty it would be 3 seconds.

MR. DISOGRA: So, because we don't know whether or not this specifically would apply to Miss Read's vehicle, the best way to conduct that analysis would be to do the entire thing with the key switch adjustment at 3 seconds, because if we're going to use a value, three is the most reasonable one because that's the one shown in testing. But we should also do the entire analysis without it in case we feel that that's the most appropriate. So, in fact, that's what I do — and will show — is I ran the analysis both ways, with and without, to be as conservative as possible.

MR. DISOGRA: He did not.

MR. DISOGRA: This is an excerpt from Mr. Burgess's original report — his page 28 — and there's a screenshot also from his page 30 at the bottom, and this is where he states in that report that when they were doing the testing with the exemplar, they observed across four different key-on cycles this 3-second difference. So what I mean by that is they had a video camera in the car that had a timer on it. And so in the video you can see the button pressed. So in their testing, for example, they pressed the button at 21:48:05. And that's sort of nailed down. When they went back and downloaded the data from the infotainment, the key on was logged in the infotainment as 21:48:08. So there's the 3 seconds. That's where they're getting that from, and that's what he states — highlighted in red there.

MR. DISOGRA: It's a little hard to read, but it basically says the average across four runs was 3 seconds.

MR. DISOGRA: This is the only point in the report it's mentioned.

MR. DISOGRA: That is his report. Everything on the left-hand side of this slide comes from Mr. Burgess's report.

MR. DISOGRA: It was noted.

MR. DISOGRA: He did not.

MR. DISOGRA: He did not.

MR. DISOGRA: He did not.

MR. DISOGRA: So the positive 26-second window — if we only apply the 3-second key on — grows to positive 29 seconds.

MR. DISOGRA: That gap would grow to a positive 34 seconds between the trigger event and the phone lock.

MR. DISOGRA: The best practice would be to run the analysis with that adjustment included and without that adjustment included, because, like I stated, without testing of an additional vehicle, we really don't know if the 3 seconds is just a specific behavior of that exemplar or if it's representative of all Lexuses of this type.

MR. DISOGRA: So, now we're going to get into the next adjustments. That first adjustment we just talked about was really a vehicle-to-vehicle adjustment. We haven't talked about the phone yet. We're going to get into the phone now and talk about aligning that. But we had to get through the internal adjustment that has to get done first. Now, we're going to look on the subsequent slides here at all of the different ways the Aperture reports compared the vehicle to Mr. O'Keefe's phone.

MR. DISOGRA: The Aperture reports lay out two methodologies for adjusting. There is a methodology based on phone calls and there's a methodology based on the three-point turn.

MR. DISOGRA: From the Aperture reports — in Mr. Burgess's original report differentiating the January report.

MR. DISOGRA: In his initial January report, he never actually takes the calculation to the end by applying the differences in times to the calculated time for 1162-2, but he does do the math, which I'll show, for the differences on these phone calls.

MR. DISOGRA: Yes.

MR. DISOGRA: He describes in his initial PowerPoint slides that 154 aligns with the end of vehicle event 1162-1.

MR. DISOGRA: Now, this is the three-point turn event. So he uses that as an anchor point.

MR. DISOGRA: Correct. The Aperture reports consistently reference everything from the end of the events.

MR. DISOGRA: On the left is an excerpt of page 33 from Mr. Burgess's initial report. And what he shows here is he has identified seven call logs that exist both in the infotainment from Ms. Read's Lexus and in Mr. O'Keefe's phone. He puts them in a table form. Then he actually goes on and calculates the difference between those two times. So that's what I meant by he sort of does the math. He makes my job a little easier and the numbers are already there. He just never applies these to any of the times. So that's where I step in and say, if I take that third column that he lays out — those are in seconds — if I apply those seconds, that adjustment, to 12:31:43, where do we end up with John's interaction window in those times?

MR. DISOGRA: Sure. I did all of them. And so I highlight on this particular slide — I just highlighted the very first one he has in here. In that case he identified that Mr. O'Keefe's phone was 2 seconds slow relative to the infotainment system. So if we apply only that adjustment, then John's interaction window becomes positive 28 seconds.

MR. DISOGRA: He didn't apply any of the offsets. He calculated all of them, but he didn't apply any of them.

MR. DISOGRA: Yes, I did.

MR. DISOGRA: Regardless of which offset you apply, John's interaction window remains positive. The number changes, but it remains positive.

MR. DISOGRA: So in this slide, I just took the other end of the spectrum. There's a large range in values in this table. Mr. Burgess doesn't show the sign conventions on here — he doesn't show when he calculates the difference whether it's positive or negative. All the numbers in the table don't have a sign, don't have a plus or minus in front of them, but there's an implication. Some of them are plus, some of them are minus. So the first one that we just looked at showed we had two seconds slow. There's another value in here that's on the other end of that. It shows that Mr. O'Keefe's phone is 21 seconds fast relative to the infotainment. So that's the largest value in the table we could take. That would bring John's interaction window down.

MR. DISOGRA: And so if we apply just that adjustment, John's interaction window becomes a positive 5.

MR. DISOGRA: So I've shifted now from the phone calls and now we're going to talk about the three-point turn alignment. And the three-point turn alignment is first presented in Dr. Welcher's slide from his presentation here. And in this he doesn't say it explicitly, but he provides some math on the screen — which is a little hard to read — but he puts some math on the screen that shows he's aligning this 154 from Mr. O'Keefe's phone to the end of event 1162-1. This is the prior event. This is a three-point turn. And so he aligns — he says this is basically how you align it — and he doesn't do the math again. He lays out that — he says — a 19.2 second difference. I actually just adjusted that to 20, which is the most conservative way to interpret that.

MR. DISOGRA: I don't think we have the accuracy here to talk about a tenth of a second. Nevertheless, I just rounded that up to 20. And when we apply that 20 seconds to event 1162-2, then John's interaction window becomes positive 6. That's just applying this. We're not combining anything. We're just applying the difference shown in this slide.

MR. DISOGRA: Yes.

MR. DISOGRA: This is a table I made. We've talked about a lot of different approaches to how these clocks can be aligned and a lot of different variables. So I tried to summarize them in one table, put them all together, because some of them we can combine. We can combine the key-on with either a phone call or a three-point turn, but we can't combine, for example, a three-point turn and a phone call — that's not valid. So I put all the possible combinations — all of them with and without key switch, phone calls, and 154 — on one table, and I calculated John's interaction window for all valid combinations from the Aperture reports. And when we do that, the result is that regardless of what adjustment we make, all of the numbers are positive.

MR. DISOGRA: Meaning, regardless of what adjustment we make, the phone lock event always occurs after the end of event 1162-2.

MR. DISOGRA: No.

MR. DISOGRA: No.

MR. DISOGRA: I think if there's sufficient data — if there's testing or research or literature that in a specific case supports there is a best path or a best method — then we should use that. In this case, I don't think there is a best practice. There isn't any literature I'm aware of that would specify exactly the best path to align all of these. So in the absence of any external data to support one path, the most valid approach is to look at all of them as a whole and then draw the conclusion from the totality of that data rather than cherry-picking a single point.

MR. DISOGRA: Yes.

MR. DISOGRA: Regardless of what adjustment or combination of adjustments we make from the original reports, John's interaction window remains positive.

MR. DISOGRA: It means the lock event occurs after the end of the text stream event — 1162-2.

MR. DISOGRA: That's correct.

MR. DISOGRA: Mr. Burgess undertakes a different alignment in that report.

MR. DISOGRA: So what I have here on the left is a screenshot of page 42 from Mr. Burgess's supplemental report. And so now he's done two things new here. He's identified a new method for aligning the data based on the three-point turn. So now he's kind of following where Dr. Welcher started. But he's approaching it a little differently. And secondly, he is now calculating the actual time — he's doing the difference calculation and applying that to the time. So now he's taking the calculation all the way through, which is a little different from the prior reports. And so when he does that, he makes my job really easy because I didn't have to do really any math. Now he comes up with — there is a range now. So prior to this point, the original report by Mr. Burgess and Dr.

MR. DISOGRA: Welcher's reports all used discrete points in time. They sought to find an event that occurs at a singular second and another event at a singular second, and align based on that. This is a little bit of a departure from that approach. Mr. Burgess is now using a range of times, a range of uncertainty. He basically assesses that the three-point turn data from the phone and the three-point turn data in 1162-1 cannot be precisely aligned down to a single point — rather, they overlap in this range of seconds — and so he then says that that range must be applied. So now we have to apply a range, and what we end up with is nine possible values. So it kind of looks like one conclusion, but really there are nine conclusions on this slide.

MR. DISOGRA: And so if we're just applying the math he has in the supplemental report, John's interaction window now has a range of nine possible values. And that range spans from positive 5 to -3. So we have five positive values. We have one value that's zero — and the zero would mean it's simultaneous. And then we have three negative values.

MR. DISOGRA: He does not provide any data or analysis to support one point being preferential over the other eight.

MR. DISOGRA: There are more positive numbers in that range of nine than there are negative numbers.

MR. DISOGRA: He opines — I'm paraphrasing because I don't know the exact way he states it — that aligning the three-point turn on a singular point is not the preferred method.

MR. DISOGRA: I don't remember specifically.

MR. DISOGRA: I don't think it is more or less likely. I think it is a valid approach certainly, and so I think the results from the additional analysis should be included with the analysis of all of the results so far, and none of those results are more or less likely than any of the others.

MR. DISOGRA: It's my opinion they're all equally weighted.

MR. DISOGRA: I did not see any mention of that in his supplemental report. And certainly the 3 seconds isn't present in his math. I don't see it there.

MR. DISOGRA: This is very similar to the table we looked at previously where I look at all possible combinations. Now, however, I just have a lot more data points to show. So I've had to break the table into two tables. So we're going to look at this and then we'll look at slide 15 next, which is the second table. I just couldn't cram 30 lines onto one slide.

MR. DISOGRA: So think of these as a singular table. I just had to split them up.

MR. DISOGRA: So this slide — I'm not accounting for the key-on. And that's the main difference between my two tables. Table one, I'm not accounting for the 3-second key-on. In table two, I am.

MR. DISOGRA: Okay, go ahead. But again, I want to make sure we're thinking about this as a singular table because that's where we'll draw our conclusions from. So the first lines of the table, we've already seen those. Those were in my original table. I haven't changed those. That's the same math that's been there. All I did was I added the nine points now from Mr. Burgess's supplemental report using the range he shows. And so now we can see — if we include that, now we get some positive numbers for John's interaction window, we get that zero. And again, the zero would imply that the end of event 1162-2 is happening exactly at the same second as the phone lock. So it's kind of not on one side or the other. So I think it should be thought of as separate.

MR. DISOGRA: And then we have three negative values which would imply that phone lock occurs before the end of 1162-2.

MR. DISOGRA: Correct. That would imply simultaneity.

MR. DISOGRA: This is 14.

MR. DISOGRA: Correct.

MR. DISOGRA: This is the summary then of that table — kind of a verbal summary of the giant 30-line table. Now, if we look at all of the possibilities from all of the Aperture reports collectively, then we can draw some pretty straightforward conclusions. 25 of those 30 possibilities result in interaction with the O'Keefe phone after the end of event 1162-2. Looking at all the possibilities, two of the 30 result in interaction where it's simultaneous — that's the zeros, present in both tables. And then looking at all 30, three of the 30 result in interaction with the phone before the end of event 1162-2.

MR. DISOGRA: No.

MR. DISOGRA: Each of the approaches have their own individual strengths and each of them have their own individual weaknesses, and there is not really any published literature, any research, or any testing done that I'm aware of that would suggest one really stands head and shoulders above another one of the methods. So I think the most fair way to present it is to look at — again, if we combine everything together, then we can draw some conclusions like this in terms of how many of the total result in one outcome or another.

MR. DISOGRA: That's correct. I have not proposed any methods for aligning the clocks myself. All of the methods that comprise these 30 are from the Aperture reports themselves.

MR. DISOGRA: Yes. There would be three — three out of the 30 possibilities would result in that.

MR. DISOGRA: Yes.

MR. DISOGRA: That is 10%.

MR. DISOGRA: Yes.

MR. DISOGRA: 25 of them are positive. John's interaction window is positive for 25 of them and only three were negative.

MR. DISOGRA: Yes.

MR. DISOGRA: That would be the airbag control module. The ACM.

MR. DISOGRA: I was provided a copy of the download from the ACM.

MR. DISOGRA: There were no recorded events whatsoever on the ACM.

MR. DISOGRA: No.

MR. DISOGRA: It records a lot of operational —

MR. DISOGRA: No.

MR. DISOGRA: Yes.

MR. DISOGRA: Yes.

MR. DISOGRA: Right. Prior to 1162-2, there were 19 prior events — 19 total events.

MR. DISOGRA: Correct.

MR. DISOGRA: None of the events were triggered by a collision.

MR. DISOGRA: Neither of the events that occurred on that date were triggered by a collision.