Go Bananas!, Landis Case Info Now Posted Options

[CapeRoadie]

Okay, time to get the hard drive roostered up! Time to devour the documents! Thanks, Floyd, now I have something to read other than reams and reams of medical narratives today--just what I needed. I think I'll transfer everything to an audio file and listen to it on my training ride today. That way, I'll be sure to get in some serious base mileage.

Good luck, today, Champ!

[Pelotonium]

I've created a couple of topics to break up the 3 major points made by Jacobs to USADA in Floyd's defense. The 3 major points are:
A. That all 4 CIR metabolites should be considered, not just 1.
B. That the best CIR metabolite shows that Floyd was clean.
C. That there might have been equipment trouble at LNDD.

The topics are posted in this forum.

[dbrower]

Are the "problems with the site" with FL.com, trustbut, or archive.org?

-dB

[CapeRoadie]

Gotta love Arnie Baker. Way to go, doc! And there's more defense arguments we haven't seen yet, too. Thanks for the mirrored site, guys, the FL website has some serious bandwidth issues this a.m. I wonder why?

Go Floyd! You are the champion!

This post has been edited by CapeRoadie: Oct 12 2006, 02:03 PM

[wildeone]

Are the "problems with the site" with FL.com, trustbut, or archive.org?

-dB

from what i could tell (i tested it even though i've already downloaded the docs), it was FL.com -- totally over-loaded. a success, i'd say! well, as long as those not able to get in right now come back later. (besides, couldn't get the big doc from them; had to use your mirror site.) yours is up and running fine. archive is slow but strong. if patience isn't used, it could really rooster one up.

[Will]

Everyone has killed the site for a bit. Go to trustbut.

For the files, you can also try:

www.box.net

Login: PublicAccess
Pass: PublicAccess

[CapeRoadie]

I've created a couple of topics to break up the 3 major points made by Jacobs to USADA in Floyd's defense. The 3 major points are:
A. That all 4 CIR metabolites should be considered, not just 1.
B. That the best CIR metabolite shows that Floyd was clean.
C. That there might have been equipment trouble at LNDD.

The topics are posted in this forum.

Okay, read the statements, and they appear pretty strong.

My cross examination questions for Floyd:

1. Do you dispute the sample was yours?
2. Is there not still an exogenous source of testosterone in your sample?
3. Isn't the T/E ratio still high?

[duckstrap]

is the variation between the three results "significant" so as to throw all the results into dispute?

if i test something that is doped using a fair testing method, shouldn't i get basically the same results each time -- if i am doing the test correctly?

Um yes. First post here, (started lurking when I heard Floyd was here, and have very much enjoyed the discussion--thank you for letting me lose a couple nights' sleep!). I will admit to being relatively experienced with analysis of clinical data (work in clinical research for drug companies), but inexperienced with the specifics of testosterone metabolism (know a lot of statistics, some biology). That said, you would want to see consistent results for both the epitestosterone and the testosterone, and a conculusive result for the ratio.

So what makes a result conclusive? In the pharmceutical industry where you are looking for clear differences among treatments, you typically need to be 95% certain that the observed result could not occur by chance (reject the null hypothesis); ergo must have occured because of a difference caused by the treatment. Here, to establish that Floyd took testosterone we want to be very sure (95% certain seems about right) that the results we are observing did not occur by chance, i.e. that he likely needed to take some form of testosterone to achieve these levels. To restate another way, we want to reject the notion that his T:E ration is less than 4 with 95% certainty. Now, normally when we are trying to make this sort of argument, we take a reasonably large number of measurements and make some assumptions about the statistical distribution of the measurements, i.e. that they exhibit a gaussian (normal or bell-shaped) and do a particular statistical test on the data (our famous t-test). Here that doesn't really work, since there are only 3 measurements of epi levels and 3 of testosterone. Alternatively, since they are all, supposedly, independent measures of the same thing, I should be able to take any of the epi measurements and put it with any of the testosterone measurements and reach more or less the same conclusion (if the test were to be conclusive). Since there are 3 of each kind of measurement, I have a total of 9 possible combinations. If just 1 of the 9 combinations results in a T:E ratio less than 4, then I am not 95% certain that the T:E ratio is >4 (specifically I am no more than 8/9 or 89% sure). So let's see how that works out (I'm using your numbers for the moment, since I don't know where the 172/18 value comes from--although that would make this even juicier).

1) T=50 E=12 T/E=4.2
2) T=50 E=14 T/E=3.6
3) T=50 E=4 T/E=12.5
4) T=61 E=12 T/E=5.1
5) T=61 E=14 T/E=4.4
6) T=61 E=4 T/E=15.2
7) T=65 E=12 T/E=5.4
8) T=65 E=14 T/E=4.6
9) T=65 E=4 T/E=16.5

Note that case number 2 results in a value of 3.6, which is less than the magic 4.0, ergo I am less than 95% sure that Floyd's T/E is > 4. Note that if I throw in the additional measurement of 172/18 in there (again, these are supposedly independent measurements of the same quantity), I have 16 combinations, and because of the 18 value for epi, I will have a total of 4 combinations (or 25%) that have T/E ratios < 4. In statistical parlance, this would be a p-value of 0.25, and it is NOT significant by a long shot (I believe "roostered up" is the preferred term).

This just suggests that the T/E ratio is much less certain than it needs to be in order to be credible. The contamination issue gives us a fairly plausible reason why the results might be as they are; i.e. the contaminating bacteria create conditions where epi is preferentially removed relative to testosterone. I will curious to hear the more biologically inclined comment on this aspect.

Kevin

[floyd]

Okay, read the statements, and they appear pretty strong.

My cross examination questions for Floyd:

1. Do you dispute the sample was yours?
2. Is there not still an exogenous source of testosterone in your sample?
3. Isn't the T/E ratio still high?

Damn it, you have to read all of it and then you can cross examine me. Otherwise you will rooster up your questions.

[RADUSA]

Damn it, you have to read all of it and then you can cross examine me. Otherwise you will rooster up your questions.

Are you going to test us as to the contents before answering questions? Doh...

[cycling newbee]

Um yes. First post here, (started lurking when I heard Floyd was here, and have very much enjoyed the discussion--thank you for letting me lose a couple nights' sleep!). I will admit to being relatively experienced with analysis of clinical data (work in clinical research for drug companies), but inexperienced with the specifics of testosterone metabolism (know a lot of statistics, some biology). That said, you would want to see consistent results for both the epitestosterone and the testosterone, and a conculusive result for the ratio.

So what makes a result conclusive? In the pharmceutical industry where you are looking for clear differences among treatments, you typically need to be 95% certain that the observed result could not occur by chance (reject the null hypothesis); ergo must have occured because of a difference caused by the treatment. Here, to establish that Floyd took testosterone we want to be very sure (95% certain seems about right) that the results we are observing did not occur by chance, i.e. that he likely needed to take some form of testosterone to achieve these levels. To restate another way, we want to reject the notion that his T:E ration is less than 4 with 95% certainty. Now, normally when we are trying to make this sort of argument, we take a reasonably large number of measurements and make some assumptions about the statistical distribution of the measurements, i.e. that they exhibit a gaussian (normal or bell-shaped) and do a particular statistical test on the data (our famous t-test). Here that doesn't really work, since there are only 3 measurements of epi levels and 3 of testosterone. Alternatively, since they are all, supposedly, independent measures of the same thing, I should be able to take any of the epi measurements and put it with any of the testosterone measurements and reach more or less the same conclusion (if the test were to be conclusive). Since there are 3 of each kind of measurement, I have a total of 9 possible combinations. If just 1 of the 9 combinations results in a T:E ratio less than 4, then I am not 95% certain that the T:E ratio is >4 (specifically I am no more than 8/9 or 89% sure). So let's see how that works out (I'm using your numbers for the moment, since I don't know where the 172/18 value comes from--although that would make this even juicier).

1) T=50 E=12 T/E=4.2
2) T=50 E=14 T/E=3.6
3) T=50 E=4 T/E=12.5
4) T=61 E=12 T/E=5.1
5) T=61 E=14 T/E=4.4
6) T=61 E=4 T/E=15.2
7) T=65 E=12 T/E=5.4
8) T=65 E=14 T/E=4.6
9) T=65 E=4 T/E=16.5

Note that case number 2 results in a value of 3.6, which is less than the magic 4.0, ergo I am less than 95% sure that Floyd's T/E is > 4. Note that if I throw in the additional measurement of 172/18 in there (again, these are supposedly independent measurements of the same quantity), I have 16 combinations, and because of the 18 value for epi, I will have a total of 4 combinations (or 25%) that have T/E ratios < 4. In statistical parlance, this would be a p-value of 0.25, and it is NOT significant by a long shot (I believe "roostered up" is the preferred term).

This just suggests that the T/E ratio is much less certain than it needs to be in order to be credible. The contamination issue gives us a fairly plausible reason why the results might be as they are; i.e. the contaminating bacteria create conditions where epi is preferentially removed relative to testosterone. I will curious to hear the more biologically inclined comment on this aspect.

Kevin

Now there's something that makes sense. Kevin, excellent job explaining it in laymen's terms. I get lost in all the scientific discussion and appreciate how you were able to break it down for us. Great post. Thanks.

[floyd]

Are you going to test us as to the contents before answering questions? Doh...

No, come on, I just haven't had any coffee yet.

[smug]

get yourself some of f-rod's brew.

[shag]

I've created a couple of topics to break up the 3 major points made by Jacobs to USADA in Floyd's defense. The 3 major points are:
A. That all 4 CIR metabolites should be considered, not just 1.
B. That the best CIR metabolite shows that Floyd was clean.
C. That there might have been equipment trouble at LNDD.

The topics are posted in this forum.

The contention that the sample must be deemed contaminated is a "major point". It's a pretty bold claim and, if correct, should tank the whole deal . . . shouldn't it? No sample = no positive.

[dchase]

Let's DO THIS!

[cycling newbee]

No, come on, I just haven't had any coffee yet.

Boy, you really are a glutton for punishment coming here without your coffee first. Guess you really do have what it takes to endure the punishment necessary to win the Tour. I was in doubt (not), but NOW I'm convinced you won fair and square!

[CapeRoadie]

Nice, Kevin. A statistician ought to be employed by the defense in order to make the case for "junk science", which lawyers use to prop up their weak stance all the time. There are many "experts" out there who will testify to this or that. I think a revisit to the federal rules of evidence (e.g., Frye, Daubert, Kelly, etc.) is in order for the Landis team, although I don't really know what the rules of arbitration in this case are, nor am I a lawyer. You just reminded me that statistics as presented by a statistician can often be used in court to throw away junk science.

This post has been edited by CapeRoadie: Oct 12 2006, 02:56 PM

[amykate]

No, come on, I just haven't had any coffee yet.

Floyd,
My first cuppa coffee sure helped reading this! I Read the submission and saw the power point presentation. My eyes hurt. My brain hurts. My question is-- the power point mentioned 'anonimity' during the testing process. Can they prove they knew it was yours? That right there.....hmm. You look to have a pretty strong argument, Floyd. Maybe i won't have to eat my shoe after all....said i'd eat it if you weren't cleared. Not feeling too hungry for it now.....

Go get 'em Champ


(mod note - removed multiple postings)

This post has been edited by guidemd: Oct 12 2006, 03:22 PM

[RADUSA]

No, come on, I just haven't had any coffee yet.

good, in fairness, I started reading the document but couldn't keep my eyes open. The coffee didn't help, so am relying on others to help me understand.

[CapeRoadie]

No, come on, I just haven't had any coffee yet.

Good morning, Floyd! May I suggest Fast Freddie brand coffee. It makes you think faster, too.

Damn it, you have to read all of it and then you can cross examine me. Otherwise you will rooster up your questions.

I am simply the early-morning devil's advocate, no one get your feathers ruffled, please. I did read all of it. I guarantee these questions are coming none-the-less. I still like the use of a statistician to prove junk science in your team of experts.