TRAX excluded from Colorado courtroom
In August of 2022 there was a Shreck (Daubert) hearing concerning the TRAX application. The outcome was a court order prohibiting the introduction of any testimony or evidence based on the software application. The following are links to the court order.
(Colorado v Jones, 22CR196 Sept 2022)
(Unhighlighted version of court order)
According to this court order, the TRAX software is not reliable, its methodology not based in science, and its creator Sy Ray, who testified at the hearing, not credible.
Official transcript of Sy Ray’s testimony in Colorado v Jones
A few highlights from Sy Ray’s testimony:
Page 18, line 2-3; “I don’t know anything about this case or the prior testimony, …” Page 24, line 5-6; “I think I saw something in this case that there was this assumption that a cell phone or a cell site could reach 45 miles.”
NOTE: Sy Ray testified he didn’t know anything about this case; so, how could he think he saw something in this case?
(Page 25); Sy Ray was asked if it was important that the software he developed was declared reliable; he replied; “I really don’t have much of an interest in it anymore”. At the time, Sy Ray was a director with LexisNexis (page 5) who owns the TRAX application.
Sy Ray admitted they used reverse engineering. (Page 15 line 13).
Sy Ray was asked by the court; “Did you consult with any engineers — radio frequency engineers about your formula?” (Page 45, Line 7). Mr. Ray replied; “Not specifically on the formula.”
On page 47, lines 16-21; Sy Ray stated they have 3 million drive tests and have mapped every cell tower in the United States.
NOTE: If 500 teams did a driving test every work day (260 days) of the year it would take over 23 years to generate 3 million driving tests.
UPDATE: Sy Ray recently admitted in court he has no idea how many unique separate driving tests they have in their database. Under cross examination he finally admitted the 3 million number comes from how many cell sectors they believe they have mapped based on some unknown number of driving tests. This is a complete change from earlier testimony where he stated they had 3 million driving tests.
Sy Ray states; “… we’re finding in our drive test data, across all 50,000,000 cell sites …" (page 47, lines 18-20)
NOTE: The most recent estimates indicate there are approximately 500,000 cell towers in America.
UPDATE: Recently, under cross examination of Sy Ray the accuracy of his statement concerning 50 million unique cell sectors being documented in their database was brought into question. It appears they might have 50 million records; however, many of which appear to be the same cell sector from different time periods.
Sy Ray admitted the formula used in TRAX was developed by trial and error (pages 43 - 45). The constant, .97 used in the formula was data driven (Page 47, line 13-15 & page 49, lines 6-9). Sy Ray stated they tried constants from .87 to 1.6 changing the value until they got the results they were looking for based on driving tests (page 48, lines 1-4, 9-16).
Sy Ray stated under oath; “Right. So, you know, I think it's important. There's two ways that we can come up with this algorithm, right? There's one way that we could sit down and we could review all the texts and all the -- the known scientific principles of radio frequency, and we can try to figure out what is the best way to approach this. Or we can do more of a data-driven approach, and that's the way that we did it.” (page 47, lines 8-15)
NOTE: Does Sy Ray admit they did not review or use known scientific principles of radio frequency when developing their algorithm?
It appears the more times Sy Ray’s testimony is scrutinized with intense cross examination the more issues with TRAX are exposed.
Interested parties should read the entire transcript. Mr. Ray was provided multiple opportunities to explain how the TRAX algorithm/formula was supported by scientific principles. No published scientific articles supporting his statements or the application were provided, no published scientific 3rd party peer reviews of the application were provided, or any 3rd party peer reviewed error rates. However, there was a scientific article published in an IEEE publication documenting issues with the TRAX application coverage area maps.
Published scientific article referencing issues with the TRAX coverage area maps
Ultimately, based in part by Sy Ray’s testimony the court ruled to exclude the TRAX application as not based in science. (People v Jones)