Story of missing hiker in Joshua Tree NP

[Perry]

This sounds kind of like the streetlight effect. (Not to be confused with gaslighting in today's politics.)
https://en.wikipedia.org/wiki/Streetlight_effect
But I don't blame anybody. There were many other areas to search in the beginning. I also don't want to become the next Bill Ewasko, and there are still some easier areas left to search. The question of ping accuracy really complicated things. Without the 10.6-mile statement by Verizon, it's likely this area would have at least been searched by helicopter, if not also some ground searches. I would imagine that the ridges are easier than the ravines.

I noticed that Ewasko listed Johnny Lang Canyon on his itinerary. From Tom's typed version it appears that Ewasko thought Crown Prince Lookout and Lucky Boy Vista were also accessed from the Quail Springs Picnic Area. Another indication that there might have been something happening mentally before beginning his hike.

I came across a nice web site while looking up those places:
https://www.hikespeak.com/trails/crown- ... shua-tree/
https://www.hikespeak.com/trails/lucky- ... shua-tree/

Wasn't able to find a partner for tomorrow, so I'll just do something else.

[Ed]

The area between Johnny Lang Canyon and Lost Horse is a land of boulders and crevice hell. I did it once or twice and never again. Seriously, the canyons are solid boulders and one misstep and it's broken ankle time.

I hope people take this as a serious warning, OtherHand. I certainly do. Back in the day, I was an excellent boulder-hopper. But I never did Class 2 talus slopes alone, since a leg injury that immobilized you would almost certainly result in dying (and slowly) before you were found. No chance of being saved by electronics and a helicopter then. And even today, they should not be fully relied on.

[Perry]

Today I searched again with sh44. We'll get the GPS track uploaded probably sometime this weekend. I learned something interesting. Both of us have AT&T service, and it sure seemed like we were getting reception from the southernmost tower on this map:
https://caltopo.com/m/5R79

There was a place where we could barely see the tower and a ridge was blocking everything to the right of the tower and we were getting strong signals. It would have blocked the Serin tower which may have been too low anyways because we were not very high up. Shortly after we couldn't see the southern tower, I got a weak signal and couldn't send a text.

This would explain why early media reports of Nguyen/Orbeso said the ping came from the Maze parking area. The Ping 2 radius matches up approximately with the parking lot. Later on, an RF engineer might have analyzed the signal strength of each antenna section to get a rough direction, as Tom had mentioned. Apparently both the direction and distance were rough, to put it mildly.

So tying this back to Ewasko, I think there could be considerable error in the ping radius, well beyond 20%. Since the Orbeso pings were closer than their hiking route, that suggests this could also happen in the Ewasko case, even though it's Verizon and likely a different system. Sure, there's theoretical arguments that a ping radius is a maximum limit, but known empirical values suggest otherwise. Ewasko could have been as far as Ryan Mountain, in my opinion. I do think the ping had to be in an area with reception of Serin tower but not reception with any other Verizon tower.

[RichardK]

Is the tower at 34.0817 -116.3429 an ATT cellphone tower? I am seeing a tower for radio station KDGL and a satellite transmitter for radio station KODV in the FCC database. There is also a structure owned by American Tower which, I think, owns microwave relay towers. The FCC records are not easy to search or understand entirely.

It is within the realm of possibility that the ping distance for Bill's phone could have an error greater than the given 10%. However, the ping sets a hard outer limit for Bill's distance from the tower within whatever error it had. It is not physically possible for Bill's phone to have been further out. Reflections off rock walls or atmospheric layers may place Bill closer to the tower, but never further.

[Ed]

This is a subject I know next to nothing about. But I am accustomed to thinking of quoted errors as being statistical, typically two standard deviations, rather than a hard limit. So I am a little puzzled by RichardK's comments, one of which is consistent with my understanding, while the other is not.

[Perry]

Richard, it sure seemed like it when we were out there yesterday. That tower is very visible. We were hiking through this general area when I made the observations:
https://mapper.acme.com/?ll=34.05864,-1 ... 0Tree%20CA
I would be curious to hear from other hikers with AT&T service.

I'm guessing the ping algorithm assumes the tower is busy and averages the amount of time until it can pull data from a buffer. If the tower is less busy with other tasks, it might be able to pull the data sooner which would give a shorter calculated distance. Whatever the explanation, it looks like there's a huge amount of error in this AT&T tower:
Orbeso Ping 1: 2.58 miles
Orbeso Ping 2: 5.93 miles
Nguyen's Car: 5.77 miles
Final location: 7.12 miles
Approximate values, accurate to about 0.05 miles. It looks like Caltopo won't snap to markers in measure mode, or I just don't know how. Considering their likely hiking routes and the times of day and temperature, I think the best-case error is 5.8 - 2.6 = 3.2 miles. Possibly as high as 7 - 2.6 = 4.4 miles.

Extrapolating this to the Ewasko case, I think it's reasonable to say that he could have been as far as 14 or 15 miles from Serin tower. I'm just adding these errors (in miles) to the 10.6-mile figure.

It would be interesting to hear from someone with AT&T service who has hiked in the Maze Loop area and also the "trails" on the SW side of the road.

[Osmanthus]

Ed - my understanding of the reason for the hard outer limit is because the error applies to path length more than anything else i.e. the signal from phone to tower is either direct or indirect, but if it's indirect then it must travel further than the direct route and it would do so by reflection off surface between the phone and tower. If this is the (only) source of the uncertainty, then the provided range is in fact a hard outer limit and the error isn't a plus-minus in the usual statistical sense.

[Ed]

Osmanthus,

Thanks for the explanation.

Ed

[RichardK]

Bill's exact outer limit depends on how much error you believe is in the reported 10.6 mile distance. If you believe Verizon when they said 10% (meaning +/- 5%), then Bill cannot have been more than 11.1 miles from the tower. If you believe +/- 20% error for some reason, then Bill cannot have been more than 12.7 miles from the tower. Otherhand's field test suggests that +/- 5% is very reasonable.

I don't think anything about the Orbeso ping can be applied to Bill. Bill had a Verizon flip phone. Orbeso had an ATT smartphone. Orbeso's ping locations are so far off that I regard them as junk. Somebody or something got it wrong.

What Bill's ping proves is that he was alive after 3 days. Well, Ed Rosenthal survived for 6 days. The brevity of Bill's ping is a clue. I can account for a single ping in only 3 ways:

1. Bill turned his phone on with a nearly exhausted battery. It got out 1 ping before the battery died for good.
2. Bill turned his phone on and then immediately off for some reason.
3. Bill was ambulatory with his phone turned on in an area of no reception. He walked through a small finger of signal before walking back into the shadow.


Option 1 requires the coincidence that just as Bill got some reception his battery failed. Option 2 does not seem rational, but maybe Bill wasn't rational at that point. Option 3 requires Bill to have his phone turned on in an area of no reception. You would think he would conserve the battery.

[Perry]

People here have different opinions on the ping accuracy. This will just influence where we choose to search. Regardless, as long as people continue searching new areas and going to previously searched areas for specific reasons (newer scattering of evidence, looking under bushes and in boulders), I think there's a good chance that Ewasko will be found.

Looks like Ryan Mountain is 18 miles from Serin. That sounds probably too far away, in my opinion, but not impossible if the Verizon ping is total garbage. I think these systems were designed for urban areas where multiple pings can be averaged and for purposes where accuracy is not super important like choosing which tower to use and how much power to use.