This thread was originally titled, "Incredible development in the $612,000 Transcon Goodge run, currently ongoing" but the new title is more descriptive. The description of the run is here.
Yeah, it's kind of like his cadence is rock solid when there are people around. And witnesses to say "look, he's really running. definitely not in the van". But then late in the day when they've left him alone, his cadence has... issues
Point 1) They will cause LOWER cadence, possibly as low as 50%, but not lower. His arms still swing. That's what is said on the Garmin forums, for those who didn't believe Sneakers.
Point 2) They will. If they show cadence data as above, I.e. 50-100% of his overall, then I'm happy.
Point 3) I don't think anyone is disputing he's running a fair way.
+1
Trekking poles will impact cadence. But they shouldn't remove it. They shouldn't sporadically remove cadence data points for minutes at a time before returning. Curiously, in his earlier days of running before we suspected he was cheating, his cadence is rock solid on big climbs where he was presumably using trekking poles.
Whatever you do, do not confirm your hypothesis with PJ's DNF from August 12, 2023 @ the Eastern States 100. No doubt was using poles on that notoriously hilly course.
Whatever you do, do not confirm your hypothesis with PJ's DNF from August 12, 2023 @ the Eastern States 100. No doubt was using poles on that notoriously hilly course.
Oof. Comparing ES100 with PJ's run yesterday, side by side, just a quick visual comparison, is pretty damning in and of itself. Much higher HR across the board. Low cadence seems to perfectly correspond to low HR and low pace. His cadence at WS100 never really drops to 0 except for tiny blips where his pace comes to a near standstill, which means he was likely taking a quick break. But none of this miles of low cadence nonsense.
You can view Paul's runs which have been flagged on Strava Desktop
Ok, now I see where you can do that. But it looks like anybody can flag any run on there, so I am not sure what to conclude from that alone.
I do wonder why they do not video the whole run from the chase vehicle. Unless maybe there are rules that the chase vehicle and the runner have to remain separated for some reason (e.g., unfair pacing) during the running portion of these types of record attempts.
Sneakers blindly defending WG way back in this thread, so I doubt he's affiliated with PJ as well.
But yes, the Garmin file contains more than you'll get from Strava.
Considering dusting off the ol' Python and attempting some clustering. Just learned a lot about HDBSCAN for work and it's definitely overkill, but it would be fun to take for a spin. Sneakers, what's your favorite clustering algorithm? Can you enlighten us since you're a data expert?
I like the dollar bill algorithm. You glue a dollar bill to the floor and watch as people cluster around it.
What are you looking to cluster with the data? It's going to be of limited use since we already know of the relationships in the data and just as important that there are existing precedent explanations for why you may not get the expected clusters (and no, clustering won't reveal anything about the heart rate data I've not already seen before). You will not find any conclusive evidence with clustering and if you thought I had a tough time even teaching you to download Strava data, how do you expect anyone but me to know what you are even talking about when you tell them to look at the clusters?
Thanks Sneakers. What would be fun and a great experiment in the name of science would be if one of us could post a series of runs on Strava. Some of the runs would contain “riding in a van” portions and some would simply be running with trekking poles in order to get the cadence dropouts. Then, it would be fun for you to take a look at the runs and tell us which runs were legitimate and which weren’t. I’d do it myself, but no matter what I do, I can’t replicate a run that doesn’t record cadence. If we can find someone who can record zero cadence runs, would you be up to the challenge?
Would be interesting to see if I could distinguish the data. Only if we get Cockerell or someone on his side to compete against me.
For PJ, what is needed is his accusers to come forward and go out to where he is and be eye witnesses. Only so much data analysis you can do, still need eye witnesses to ground it to reality.
Thanks Sneakers. What would be fun and a great experiment in the name of science would be if one of us could post a series of runs on Strava. Some of the runs would contain “riding in a van” portions and some would simply be running with trekking poles in order to get the cadence dropouts. Then, it would be fun for you to take a look at the runs and tell us which runs were legitimate and which weren’t. I’d do it myself, but no matter what I do, I can’t replicate a run that doesn’t record cadence. If we can find someone who can record zero cadence runs, would you be up to the challenge?
Would be interesting to see if I could distinguish the data. Only if we get Cockerell or someone on his side to compete against me.
For PJ, what is needed is his accusers to come forward and go out to where he is and be eye witnesses. Only so much data analysis you can do, still need eye witnesses to ground it to reality.
We need stealthy boots on the ground to surveil PJ without his knowledge. Then we can see whats really going on.
There has to be someone close enough to his route to do this ?
I like the dollar bill algorithm. You glue a dollar bill to the floor and watch as people cluster around it.
What are you looking to cluster with the data? It's going to be of limited use since we already know of the relationships in the data and just as important that there are existing precedent explanations for why you may not get the expected clusters (and no, clustering won't reveal anything about the heart rate data I've not already seen before). You will not find any conclusive evidence with clustering and if you thought I had a tough time even teaching you to download Strava data, how do you expect anyone but me to know what you are even talking about when you tell them to look at the clusters?
You could use these four dimensions to cluster: - Elevation - Cadence - Pace - HR
I would take that data and either take a rolling average of each of those dimensions based on a pre-defined quanta, basically simplifying the data set into, idk, 10s chunks, OR, apply a smoothing filter.
An algorithm like HDBSCAN will find the islands with the most prominence. My hypothesis would be that if I took known clean, multi-day runs from others, including PJ's earlier days or his WS100 run, you're going to see three clusters:
1: Running/walking flat/downhill 2: Walking/running slowly up hill 3: Standing still
1 and 2 might become 2 separate clusters, depends on how aggressive you set the parameters. Now compare that to suspect runs and I hypothesize that you'd see a fourth island form, and that would be his weird cadence dropping stretches.
Then, just for fun, do PCA to reduce the dimensions to two so we have a nice plot to look at.
1: Running/walking flat/downhill 2: Walking/running slowly up hill 3: Standing still
1 and 2 might become 2 separate clusters, depends on how aggressive you set the parameters. Now compare that to suspect runs and I hypothesize that you'd see a fourth island form, and that would be his weird cadence dropping stretches.
Then, just for fun, do PCA to reduce the dimensions to two so we have a nice plot to look at.
Any questions?
Yes, I have questions:
1- Do you have a lot of free time? Way more useful if you actually became boots on the ground instead of wasting time and effort on clustering that doesn't reveal anything I don't already know.
2- Of course there will be a fourth island with the dropping cadence. But this is something we already know and there is the precedent explanation of the trekking poles, it's not new information. How does that fourth island prove anything one way or the other? I have the data for another runner who did the run across America several years ago who also had dropped cadence data, and he was not as handsome as PJ and no one suspected him.
Basically, what you are clustering doesn't explain the physical reality of WHY the data is the way it is.
As much as I would enjoy to bring my work home and starting messing with data analysis, I would like to point out its not needed in this case.
We have his data on Strava from the last week where the cadence does not match:
* Previous runs.
* Previous runs that include walking & walking uphill.
* Previous runs that include using poles & poles on the uphill.
* Previous runs of long distances, including back to back long days.
* We even have runs in forested areas where the GPS signal was poor.
I don't need to start clustering, or adding filters, because I didn't mess with clustering or filters on any of his previous data. We are seeing a complete anomaly on this run irrespective of data analysis. This alone raises the suspicion to the level where everyone should be skeptical enough to want the additional Garmin data. The refusal to provide anything and immediate response of calling everyone who is skeptical a hater is not how an innocent person with exonerating evidence acts.
This post was edited 51 seconds after it was posted.
Note, this does not prove Paul is cheating, this simply is saying we have enough evidence (given my standard of evidence) to say that no one should be confident he is doing this legitimately.
When someone is claiming to do an incredible athletic event, similar to an FKT, I believe the responsibility is on the runner / athlete to prove they did it legitimately, not for others to prove they cheated. If it was the other way around I could claim to have broken the record and all of you would have to prove me a cheater.
56 miles in 13hrs so far today. They are going for 74, and have published their routes for the next 5 days in a change vs the pre-rockies mileage randomness .
Do they go through 11pm tonight to stick to their new schedule? Or will we see a recalibration after only 1 day attempting 75/day?
I ran with Paul a little over a week ago. I've given him and his crew a little advice here and there, and I feel like he trained to run 50-60 miles a day, which is far easier than 70+ when you factor the pace/recovery clock, but I have to say this whole run is not about a transcon record or some threshold that only Ashprihanal Aalto has achieved on a closed course. It never was. He's a decent enough runner to raise funds without preaching something so much. But it appears he only wants social media, and not listen to my advice. Paul and his crew had no idea it was me until I told them who I was at mile 4 in the RV. That was all I needed to know.
I ran with Paul a little over a week ago. I've given him and his crew a little advice here and there, and I feel like he trained to run 50-60 miles a day, which is far easier than 70+ when you factor the pace/recovery clock, but I have to say this whole run is not about a transcon record or some threshold that only Ashprihanal Aalto has achieved on a closed course. It never was. He's a decent enough runner to raise funds without preaching something so much. But it appears he only wants social media, and not listen to my advice. Paul and his crew had no idea it was me until I told them who I was at mile 4 in the RV. That was all I needed to know.
Welcome back to the thread, Pete. Some questions for you:
1- Did you meet the super secret sleuth that Cockerell sent to follow PJ on March 10th? The same day you ran with PJ? Or was Cockerell's secret sleuth you?
2- Do you agree with me in saying that PJ is by and large running this transcon and is not out to cheat? And PJ simply has bigger eyes than he can stomach?
I ran with Paul a little over a week ago. I've given him and his crew a little advice here and there, and I feel like he trained to run 50-60 miles a day, which is far easier than 70+ when you factor the pace/recovery clock, but I have to say this whole run is not about a transcon record or some threshold that only Ashprihanal Aalto has achieved on a closed course. It never was. He's a decent enough runner to raise funds without preaching something so much. But it appears he only wants social media, and not listen to my advice. Paul and his crew had no idea it was me until I told them who I was at mile 4 in the RV. That was all I needed to know.