While I respect skepticism as in this case it's especially warranted, I do want to point out that you've gotten the investigatory body wrong - the NHTSA is not the NTSB. The NTSB preliminary report itself is really very straightforward and does not at all exculpate Tesla in this incident - a likely reason for that is highlighted in the Ars article:

For example, a Tesla defect known as “Sudden Unintended Acceleration” can occur when “components of the vehicle require additional power” and the draw on the battery causes “significant spikes in the system,” their lawsuit explained. If that happened to Butler’s car, the inverter may “incorrectly interpret that the accelerator pedal has been pressed” and rapidly advance to dangerous speeds.

It is fairly clear in the above NTSB release that they do not have proof positive that the accelerator was physically depressed, just that the logs indicate that is what happened. That ambiguity/uncertainty is likely the reason they have opened a special investigation into this incident, and given the vast horizons of potential biases in reporting about this tragedy I agree that best practice will be to look primarily at what the source has said, and treat speculation (both for or against Tesla) with a healthy degree of skepticism.

Edit: Further, the NTSB is likely pulling this data directly from the EDR itself - a device that must adhere to some extremely strict federal standards on retrieval and data integrity. Teslas do automatically report this information out to Tesla's servers in the event of a crash, but there has never been any indication I have been able to find that Tesla has even the capability to edit the information stored on the device itself - and the EDR is a system component frequently audited by numerous independent bodies for compliance, including that most reliable of compliance tester: open source enthusiasts.

Edit 2: clarity