Sandra Vi going for record across America

Apologies for identifying you with Derek Murphy. I will correct that. I must have heard somebody else say it erroneously.

That makes sense now about Sandra's watch sampling in an 8-10s range. I still think that smoothing makes more sense than sampling though as the watch is doing so internally (it is measuring it's cadence value over those 10s, not randomly choosing one of the 1s measurements which will be more prone to random errors than sampling over a longer time point I would have thought). The best plan would be to smooth but than randomly sample in the windows. I don't think it will change my point though.

Obviously not all data follow Bedford's Law, and these don't seem to. But they do seem to follow something. Of course it can be corrected, but it's an iterative process and not one that could have been done ahead of time. Checking for repeating patterns and harmonics was one of my next plans.

I agree that all evidence matters, and it was interesting to try to spoof the data. But you and others were very adamant that Mimi's data were fake and garbage. I think that you were wrong and it is good to see you say so. For my part that is all I care about, and I hope that we will see some video soon to fully exonerate her.

stupid runner wrote:

Hi all. I recently started looking into some of the allegations against Mimi regarding data forgery to see if I could add an additional point of view to some of the accusations. I ended up putting this into a blog so that I could show my working, show my reasoning, and importantly show my code in case anybody spots any errors. It is really just touching the surface of things, but based on this I believe that Mimi's Strava data are genuine.

Thanks for the very thorough analysis, this is great! Please ignore anyone who is hurling insults. I really appreciate the effort you put into this. Remind me not to cheat on my taxes and try to sneak it by you.

I agree with your conclusion that the data is genuine. I'd more-or-less reached the same conclusion, based on the seeming lack of technical skills on Mimi's team compared to the skills needed to construct a convincing fake. And also on the strange bimodal cadence data, which would actually be *more* effort to fake than more normal-looking cadence data. A couple of your findings removed any lingering doubts I had:

1. The bimodal cadence data is matched by bimodal pace data.

2. The data passed two clever tests for evidence of fakery (little blips in the sample intervals, and the 2nd digit distribution of cadence). The known fake data failed both tests.

I agree with those who suggested Benford's Law doesn't apply here, but I still think your analysis of the 2nd digit distribution is valid and useful.

FYI, there's something wrong with your graph of "Pace Comparison Between Data Sets" and the accompanying text. You said Sandra's normal pace was about 11.5 minutes/mile and Mimi's was split between 11 min/mile walking and 8.5 min/mile running. Those numbers can't be correct, because both women averaged around 14 min/mile overall. Maybe you meant minutes/km?

If the data is genuine, and there's any remaining question about possible cheating, it could only be from another person running some segments instead of Mimi. We can't really prove or disprove that. However, the strange bimodal cadence is there from Day 1 of the data, which suggests it was probably Mimi running all of the segments.

In short, I believe it's very likely Mimi ran the whole thing exactly as she claimed.

Actually, no. I have no idea when that was, sorry. I haven't really been following that closely I'm afraid. I haven't analysed all of the data, just between 2nd and 7th October. The Strava data when averaged over all runs over these days shows Mimi ran an average of about 16 min/mile while Sandra ran an average of about 13.5 min/mile. So that fits with what you are saying. Higher cadence does not necessarily mean higher speed, right?

So far today wrote:

Look, the reason her cadence was buggered was because she was running on a bum leg. The buggered cadence occurred right after the vacation day. Do you know what that day is. Do you know what else happened on that day? Since you have analyzed everything you will know the answers immediately. Then you will understand how tragic it is that Mimi ran on an injured leg. Received medical attention twice yet continued on through all the pain. How she was able to run at all on a leg where the cartilage had ceased to exist days prior.

The movescount data can tell a story. For example, look at the following movescount data of pace plotted along with cadence, and the annotations:

(Contrary to the suggestion by 'stupid runner', it is the high sampling rate of the suuno that provides the fidelity by which we can deduce these telling 'breadcrumbs'.)

From similar data scrutiny of later data, does the cadence data possibly reflect labored walking partially due to pain?

Edit: "...it is the high sampling rate of the *Suunto..."

(Sorry about that.) Also, I'm somewhat impressed with the Suunto movescount app.

I think all the analysis done will end up to the same place of us wanting to see the uncut videos of Mimi running from her film crew. I have seen that brief YouTube video of Mimi from years ago of her running on the treadmill for the world record and yes, I do see some ~190 spm cadence.

From Mimi's current cadence data, she spends a significant amount of time in the 185-195 spm range, which there is no video evidence for. Of the two videos we have of Mimi running for any notable length of time, the first on Tim Anderson's Facebook page shows a ~140 spm.

The only other notable video of Mimi running is from the Scrumptious montage with many different cuts. The best I can make out from these many cuts is Mimi running at 160-168 spm:

https://vimeo.com/238314279

There is a clear gap in supporting evidence and it is important to obtain full video from both before and after our attention. I really do wonder how Mimi could have been running so fast over very hilly terrain the first two weeks if she is switching from walking to running so often.

I think this all boils down to whether we can get all the video.

Note that my conclusion from the Benford's Law analysis was that the cadence data do NOT follow Benford's Law for exactly the reason that the cadence is highly bounded. But following that I looked at the insignificant digits (which is the second digit for 2 digit numbers, but the third for 3 digit numbers, so will not follow the trends of Benford's Law). That is where there seemed to be a difference. I don't know whether a uniform distribution is "the norm", but one could get a whole load of data from Strava to test that. Might be interesting to do.

For "stupid runner", another point in your sh Faustus that should be corrected is that Mimi and her team often didn't upload data for days at a time as can be seen in their blog posts and not the "15 minute" delay in your article:

http://www.mimirunsusa.com/news/page/2/

Typo in my post above, should say in your article.

SANDY VI: 5-MILE SPLITS FOR OCT 19th

Time of Day: Miles completed

02:28 AM: 0

03:48 AM: 5

04:58 AM: 10

MILESTONE: 05:55 AM: 14 miles. Reaches the Mississippi. Goodbye Iowa. Hello Illinois.

06:13 AM: 15

07:18 AM: 20

08:21 AM: 25

09:27 AM: 30

10:08-10:39 AM: 33 Lunch Break (31 minutes)

11:07 AM: 35

Scam_Watcheroo wrote:

For "stupid runner", another point in your sh Faustus that should be corrected is that Mimi and her team often didn't upload data for days at a time as can be seen in their blog posts and not the "15 minute" delay in your article:

http://www.mimirunsusa.com/news/page/2/

I think there was one occasion where she didn't upload for one day. So neither often nor days.

Nah wrote:

Scam_Watcheroo wrote:

For "stupid runner", another point in your sh Faustus that should be corrected is that Mimi and her team often didn't upload data for days at a time as can be seen in their blog posts and not the "15 minute" delay in your article:

http://www.mimirunsusa.com/news/page/2/

I think there was one occasion where she didn't upload for one day. So neither often nor days.

Liar. Jihad Gary?

"As soon as Mimi finishes this evening I’ll be able to come back into town and download everything to Strava after several days of not being able to upload."

SEVERAL DAYS

And this happened very often in the beginning of the scam.

stupid runner wrote:

Actually, no. I have no idea when that was, sorry. I haven't really been following that closely I'm afraid. I haven't analysed all of the data, just between 2nd and 7th October. The Strava data when averaged over all runs over these days shows Mimi ran an average of about 16 min/mile while Sandra ran an average of about 13.5 min/mile. So that fits with what you are saying. Higher cadence does not necessarily mean higher speed, right?

I expected you have not bothered to read the thread at all. If you have you have noticed that I have provided 5-miles splits data for more than a week. And you would know that Mimi's Stava excludes break times while Sandy's includes them. You need to use start of day to end of day times for both Mimi and then exclude her actual breaks (also included as part of the splits). So, one error you have is not having an apples-to-apples comparison for time either for the entire day, or with breaks excluded for both ladies.

Scam_Watcheroo wrote:

For "stupid runner", another point in your sh Faustus that should be corrected is that Mimi and her team often didn't upload data for days at a time as can be seen in their blog posts and not the "15 minute" delay in your article:

http://www.mimirunsusa.com/news/page/2/

Okay. Sorry I didn't mean to make out that it was only a 17 minute delay, but yes I see your point. I have edited it (note that I am hosting through GitHub, and having a few issues pushing my updates, but will do so as soon as possible).

What I think is important to see if the data shows it is if the data indicates an injury occurred and when it started. It seems to me that is one explanation for the wacky cadence data. The other is that a different runner took over. But something did change in Mimi's run performance.

If it was an injury that started and got progressively worse it could be helpful to know that. It could help coaches or a doctor know that something is not quite right with the athlete. And they could determine this from a distance and decide to check further on what is really going on.

Mimi has a severe knee injury, and now back and maybe hip issues. That is not the proper end result. It should never have gotten to the point it did for Mimi. I find it saddening that her fans continued encouraging her to continue onward.

But how can Mimi's data exclude break times? The time differences that I have calculated are based on time of day, not on the time registered by the watch. They are all 1 second apart There are as many data points as there are seconds in the amount of time that she has been running (give or take literally one or two skipped seconds). That means that it was never paused - any breaks are included in the data, no? Sandra's on the other hand has data points that are several minutes apart (again based on time of day, not time on the watch). I have literally scraped the data from the Strava website. I have done nothing to them. I genuinely don't know why that differs with what you are seeing. And also note this was never supposed to be a comparison between the two. I just included them both to show that Mimi's data looks real and thought that things like the differences in their strategies was interesting.

this is amazing, no? wrote:

(Contrary to the suggestion by 'stupid runner', it is the high sampling rate of the suuno that provides the fidelity by which we can deduce these telling 'breadcrumbs'.)

Sorry, what suggestion? I agree that the high sampling rate of the Suunto is great as it provides fidelity - when did I say different? Or have I misunderstood your point?

Looking over a longer time period to see how Mimi's running changed is certainly worth looking at. Certainly in the time-period I looked at she was fairly consistent with her cadence, but her split of walking to running changed day by day.

Mimi's run can always serve as an object example of how NOT to do something of this nature. A complete failure in every sense. The only way it could have gone worse is if she had gotten hit by a car.