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.
For the record, before I get in trouble for politicizing this, I am not calling you a Nazi supporter. I'm calling you extremely naive and gullible.
Thank you so much for clearing that up... Perhaps think of better examples than Hitler next time.
I am neither naive nor gullible. It isn't gullible to believe in the good in people (or the possibility of good for that matter). It isn't naive to not automatically jump to a belief in fraud just because a runner makes a big statement without a track record to back it up.
I would MUCH rather be a positive person than a negative one. I would much rather believe in good than assume bad. I would rather attempt to shine a little light on hope rather than tear down like many have done on this thread, and as you have tried to do to me.
With most horrifically obnoxious LR posters, it's fairly easy to discern whether they're morons dumb enough to actually believe any of the crap they post, or just trolls with such deeply pathetic lives they simply have nothing better to do than look for ways to piss reasonable people off and disrupt any actual legitimate discourse.
Then there's a smaller population of guys like you, with whom it's just brutally hard to figure out.
It's hard to imagine you could possibly be that dumb --but you actually sound kind of sincere, so it really gives one pause.
You're either a fairly good actor (though really demented), or you've got the critical-thinking skills of the average cinderblock.
Either way, the only reason I had the misfortune of seeing your posts is that when I get logged out, the ignore function turns off till I log back in.
Gonna fix that right now, and send you right back into that vast unseen background haze of people not worth paying *any* attention to.
Regardless of the cheating or not, I’m sure it’s been said in the past 166 pages but he’s a hard guy to like/root for. Dismissive of anyone offering any advice in the comments, pretty rude to anyone who asks a challenging question, and comes off as arrogant/socially unaware in a lot of his videos imo. Although impressive to run that far, doesn’t give you license to act like that
Regardless of the cheating or not, I’m sure it’s been said in the past 166 pages but he’s a hard guy to like/root for. Dismissive of anyone offering any advice in the comments, pretty rude to anyone who asks a challenging question, and comes off as arrogant/socially unaware in a lot of his videos imo. Although impressive to run that far, doesn’t give you license to act like that
I choose to see the good in people. Maybe 69 is his "lucky number" and it motivates him to be successful. I build up vs tear down. Go #69!
Question for sneakers on the ground: Let me preface that I’m not a statistical guy by any means. I did some initial reading on Kalman filters. One of the things that I saw was it is appropriate to use when you have a normal distribution of data. The filter smooths out garbage in the data to get a better result. So the big assumption is that the data is normal right and what some are interpreting as van riding is just bad gps readings. In this case, how do we know it is? For example….if I went for a run and then hopped on a van for 800 meters and hopped out, wouldn’t your application of the Kalman filter on my run just say….yeah, that’s corrupt data, we will just smooth it out? What would you expect to see differently than what Paul’s data is showing if he really was hopping on the van? Genuinely curious.
If Kipchoge uploaded a 2:04 marathon to Strava I wouldn't bat an eye but if matt choi did I would scrutinize all the data. Since PJ does not have the level of training, success, fitness as PK of course he is going to get more scrutiny. His claim of even close to a world record is more unbelievable than PK, a highly trained successful ultra guy, running an actual world record.
In addition, a few things can be true at once. It's true that any donations to charity are good, he is putting himself through something difficult, etc.
It's also true that some of the data is fishy and doesn't add up, and since he has a higher burden of proof than PK would it is not unreasonable for us to scrutinize a bit.
@Washed88, I appreciate this response. It was made without blasting him or without calling him names (like fraud).
Yes, absolutely things can and should be questioned. My greatest challenge with this string on PJ along with the previous string on WG was the visceral hate and venom that has been spewed about them. And, that is why I've jumped back into the fray again.
Perhaps any future scrutiny can be based on the stats and actual facts rather than name calling and immediate jump to guilt.
Question for sneakers on the ground: Let me preface that I’m not a statistical guy by any means. I did some initial reading on Kalman filters. One of the things that I saw was it is appropriate to use when you have a normal distribution of data. The filter smooths out garbage in the data to get a better result. So the big assumption is that the data is normal right and what some are interpreting as van riding is just bad gps readings. In this case, how do we know it is? For example….if I went for a run and then hopped on a van for 800 meters and hopped out, wouldn’t your application of the Kalman filter on my run just say….yeah, that’s corrupt data, we will just smooth it out? What would you expect to see differently than what Paul’s data is showing if he really was hopping on the van? Genuinely curious.
That's not what is meant by normal distribution for Kalman filters. Check the highlighted text in the Wikipedia link below. Kalman filters assume that the errors in the data are normally distributed, meaning it assumes that the errors follow a standard bell curve:
For statistics and control theory, Kalman filtering, also known as linear quadratic estimation (LQE), is an algorithm that uses a series of measurements observed over time, including statistical noise and other inaccuracies,...
Thank you so much for clearing that up... Perhaps think of better examples than Hitler next time.
I am neither naive nor gullible. It isn't gullible to believe in the good in people (or the possibility of good for that matter). It isn't naive to not automatically jump to a belief in fraud just because a runner makes a big statement without a track record to back it up.
I would MUCH rather be a positive person than a negative one. I would much rather believe in good than assume bad. I would rather attempt to shine a little light on hope rather than tear down like many have done on this thread, and as you have tried to do to me.
With most horrifically obnoxious LR posters, it's fairly easy to discern whether they're morons dumb enough to actually believe any of the crap they post, or just trolls with such deeply pathetic lives they simply have nothing better to do than look for ways to piss reasonable people off and disrupt any actual legitimate discourse.
Then there's a smaller population of guys like you, with whom it's just brutally hard to figure out.
It's hard to imagine you could possibly be that dumb --but you actually sound kind of sincere, so it really gives one pause.
You're either a fairly good actor (though really demented), or you've got the critical-thinking skills of the average cinderblock.
Either way, the only reason I had the misfortune of seeing your posts is that when I get logged out, the ignore function turns off till I log back in.
Gonna fix that right now, and send you right back into that vast unseen background haze of people not worth paying *any* attention to.
Do you enjoy calling people that you do NOT know names like dumb or demented? Does that give you pleasure?
I don't care if you see this or not...Your level of hate and cynicism is beyond sad.
My mentality has made me a very happy person. My intelligence has given myself and my family a wonderful life. My critical-thinking skills has made many companies better off.
Why all the HATE on here? It is serving zero purpose. Can anyone stick to facts and actual observations rather than the immediate rush of judgment? Is that possible any more on here?
It is so easy to be a keyboard warrior where you hide behind a screen name and make hateful statement after hateful statement (sp2, I will give you credit for actually having a registered name vs those that are posting completely anonymous).
If you are questioning PJ, fine. But do so with facts. If you are questioning and you care enough then go monitor him without him knowing you are coming. If you are questioning the type of person he is, then go run with him and see for yourself.
Question for sneakers on the ground: Let me preface that I’m not a statistical guy by any means. I did some initial reading on Kalman filters. One of the things that I saw was it is appropriate to use when you have a normal distribution of data. The filter smooths out garbage in the data to get a better result. So the big assumption is that the data is normal right and what some are interpreting as van riding is just bad gps readings. In this case, how do we know it is? For example….if I went for a run and then hopped on a van for 800 meters and hopped out, wouldn’t your application of the Kalman filter on my run just say….yeah, that’s corrupt data, we will just smooth it out? What would you expect to see differently than what Paul’s data is showing if he really was hopping on the van? Genuinely curious.
That's not what is meant by normal distribution for Kalman filters. Check the highlighted text in the Wikipedia link below. Kalman filters assume that the errors in the data are normally distributed, meaning it assumes that the errors follow a standard bell curve:
But sporadic cheating will likely not be normally distributed. It only happens at certain times, like going uphill in the evening for example. Cheating likely doesn't look like noise. I don't believe you can make the Gaussian assumption here, therefore a Kalmin filter isn't worth bothering with. Correct me if I'm wrong.
Not a statistics expert, but my boss has a PhD in stats, maybe I'll run this by her. She'd get a kick out of it probably.
Don't worry, I am not making this post to spew positivity...
I just looked at my last two runs on Strava. On my run this morning it showed 14 times with zero cadence and on yesterday's run, it showed 8. But during every one of those zero cadence marks there was still a pace shown. One of the zero marks today I was stopped at a stop light so the zero cadence makes sense but it still showed me running a 9:37 pace at that exact moment.
I am not a statistician nor am I dumb (like some have attempted to say). But, with all of the outcry of PJ having zero cadence spots while still having a heart rate and a running pace, can you explain to me the issue? I am not trying to cross the country nor break any records. But, I am a runner who goes out every day...and it appears every day that I am completely stopping (zero cadence) when I might merely be in a walk break.
Don't worry, I am not making this post to spew positivity...
I just looked at my last two runs on Strava. On my run this morning it showed 14 times with zero cadence and on yesterday's run, it showed 8. But during every one of those zero cadence marks there was still a pace shown. One of the zero marks today I was stopped at a stop light so the zero cadence makes sense but it still showed me running a 9:37 pace at that exact moment.
I am not a statistician nor am I dumb (like some have attempted to say). But, with all of the outcry of PJ having zero cadence spots while still having a heart rate and a running pace, can you explain to me the issue? I am not trying to cross the country nor break any records. But, I am a runner who goes out every day...and it appears every day that I am completely stopping (zero cadence) when I might merely be in a walk break.
That's statistical noise. That's exactly what a Kalmin filter is for. If I go for a run there will be many spots like that. For example, I approach a light and slow down and then press stop. Depending on GPS accuracy in that moment, it might say I'm moving at the same pace but with slower cadence because cadence is independent of your recorded coordinates. These are small blips, it can happen when you stop, when you make a turn, when you run beside a tall building, and sometimes even when you're running in a perfectly straight line at a rock steady pace.
If you applied a Kalmin filter it would smooth out those tiny little blips and look like normal running data.
PJ's data has extended periods like this. That's the problem.
Don't worry, I am not making this post to spew positivity...
I just looked at my last two runs on Strava. On my run this morning it showed 14 times with zero cadence and on yesterday's run, it showed 8. But during every one of those zero cadence marks there was still a pace shown. One of the zero marks today I was stopped at a stop light so the zero cadence makes sense but it still showed me running a 9:37 pace at that exact moment.
I am not a statistician nor am I dumb (like some have attempted to say). But, with all of the outcry of PJ having zero cadence spots while still having a heart rate and a running pace, can you explain to me the issue? I am not trying to cross the country nor break any records. But, I am a runner who goes out every day...and it appears every day that I am completely stopping (zero cadence) when I might merely be in a walk break.
That's statistical noise. That's exactly what a Kalmin filter is for. If I go for a run there will be many spots like that. For example, I approach a light and slow down and then press stop. Depending on GPS accuracy in that moment, it might say I'm moving at the same pace but with slower cadence because cadence is independent of your recorded coordinates. These are small blips, it can happen when you stop, when you make a turn, when you run beside a tall building, and sometimes even when you're running in a perfectly straight line at a rock steady pace.
If you applied a Kalmin filter it would smooth out those tiny little blips and look like normal running data.
PJ's data has extended periods like this. That's the problem.
Thanks for that explanation. If PJ is not stopping his watch when he is stopping for food or stretch by the van, how long would Strava show a pace even without a cadence? I looked at one segment of a couple of minutes on one of his runs and it shows a zero cadence, still a pace, and his HR slowing down. Would Strava still show that with a pace even for a couple of minutes?
Thanks sneak. So how would the data actually look if someone was taking short van rides? What specifically would look different?
You are the only person asking the proper questions. Since we know from many photos and videos that PJ is using trekking poles, cadence drop outs and low cadence values can be expected to be seen in a sizeable portion of the data so cadence cannot be reliably used in this case as a signal.
It means that we have to look for sizeable portions of the data where the speed is above what is humanly achievable for an all day run. This is why it is so important to apply a proper noise filter to the GPS long and lat data before calculating speed since it comes in 1 second intervals and you would be swamped by relatively huge GPS errors.
Everyone on here that has been accusing PJ of cheating has been focusing on the wrong days and the wrong parts of the data because they don't know how to properly process it and what to look for.
Has there been a day where I think PJ cheated? Yes
Thanks for that explanation. If PJ is not stopping his watch when he is stopping for food or stretch by the van, how long would Strava show a pace even without a cadence? I looked at one segment of a couple of minutes on one of his runs and it shows a zero cadence, still a pace, and his HR slowing down. Would Strava still show that with a pace even for a couple of minutes?
Based on my experiences with a 5 year old Garmin, it shouldn't show a pace for more than a few seconds. It would be a very bad GPS watch if it did. And he has two, nice, new ones with all the bells and whistles.
Think about this. You start running from a standstill the moment you start your watch. Look at the pace. It might take 5–10s to actually display your real pace. Works the same when you stop.
But sporadic cheating will likely not be normally distributed. It only happens at certain times, like going uphill in the evening for example. Cheating likely doesn't look like noise. I don't believe you can make the Gaussian assumption here, therefore a Kalmin filter isn't worth bothering with. Correct me if I'm wrong.
Not a statistics expert, but my boss has a PhD in stats, maybe I'll run this by her. She'd get a kick out of it probably.
A Kalman filter is not whether someone has cheated or not, it's for minimizing the GPS errors in the long and lat since the data is at 1 second intervals and GPS errors are relatively large at this scale at human walking/running speeds.
Has there been a day where I think PJ cheated? Yes
Reading the thread, this feels like a breakthrough. PJ's strongest supporter believes that he has cheated.
That means that bickering aside, everyone is on a similar page, broadly speaking. That's to say that if he did break the record, it shouldn't count.
There's a difference between what I think and what can be proven with the data. I'm not in the same baseball field as the rest of you harassing PJ without proper proof.
A Kalman filter is not whether someone has cheated or not, it's for minimizing the GPS errors in the long and lat since the data is at 1 second intervals and GPS errors are relatively large at this scale at human walking/running speeds.
Okay, yes, we agree there. I thought you were arguing that filtering noise would prove he wasn't cheating. It appears I misunderstood you.
His response to 'the haters' is making things worse for himself.
If he's innocent then I don't really understand why he's approaching it the way he is. He could just drop the files people have asked for and then he, and everyone else, can move on.
Instead, posting a video showing that you're now intentionally messing with the data to 'piss people off' is just going to raise more eyebrows surely?
