Kimmage on Wiggins, Sky

[thehog]

After all he did covert 8.9m to 10m. - bending truth is an attribute.

 

[BigBoat]

The higher your power output (performance) in relation to your genetic ceiling the more likely you are to be doped in my very simple opinion.

With 6-6.4 watts per kilo at threshold one cant help be suspicious of Froome and Wiggins. Its very very hard to not wonder about Bradley's obvious improvements over the last 6 years. Especially as it appears he's produced more sustainable watts than his hero Floyd Landis.

 

[Kender]

thanks for posting this. I don't disagree with you. My point was a slightly different one, and perhaps I did not state it clearly. As your list illustrates, there is a strong correlation between winning the Tour and doping. But this does not mean that doping is strongly correlated with high rider rank because looking at Tour winners only gives us a small sample of the entire peloton. What happens when we look at the entire population of riders – does doping correlate with high rider rank? This could be true if, for example, being a rider who is in a position to win induces those riders to dope, while other riders are less incentivized to dope. I admit I have not done a statistical study, but instead simply examined the list of recent doping positives/decisions/arrests over the last year (on dopology) assuming this was more or less a random sample of doping positives. I also assume that the probability of getting caught was independent of rider rank. That said, the conclusion I reached is that doping correlates with low rider rank. Admittedly, one would have to bin riders according to finer bins of rank, etc. to do this properly. My intuition, however, is that doping is likely largely independent of rider rank since there are incentives to dope at all levels of rank. Low ranked riders may be more tempted to dope since the risk/benefit tradeoffs are greater, but that is speculation. Poor performing riders dope to obtain a contract, make teams, get selected to ride major races, etc. High performing riders dope to win major races, get larger contracts, etc. Given the structure of incentives, one would expect a fairly uniform distribution.

Interestingly, in a study (not cyclists) Petróczi found that doping behaviors appear not to correlate with win orientation, competitiveness, or goal orientation (Attitudes and doping: a structural equation analysis of the relationship between athletes' attitudes, sport orientation and doping behavior), which is consistent with the uniform distribution view.

you're all over the place. no wonder noone can see what point you;re trying to make

 

[thehog]

Taking drugs is a poor mans thing. Rich don't no do no drugs.

 

[Ferminal]

Those unable to afford expensive doping regimes are more likely to test positive or be caught by law enforcement, agreed.

 

[Cerberus]

Interestingly, in a study (not cyclists) Petróczi found that doping behaviors appear not to correlate with win orientation, competitiveness, or goal orientation (Attitudes and doping: a structural equation analysis of the relationship between athletes' attitudes, sport orientation and doping behavior), which is consistent with the uniform distribution view.

Nope, it's consistent with the view that doping use is independent of pre-doping result. If however one makes the radical assumption that performance enhancing drugs actually enhance performance that naturally lead to the conclusion that if PED use is uncorrelated with pre-PED result, PED use will be correlated with post-PED results. In fact logic dictates that every performance boosting factor (talent, training hard, low risk of injury, racing smart, PEDs and even dumb luck) is found more commonly in the top tier, because the people who have the most or best combination of those factors will by definition win.

 

[Neworld]

Won Giro - far tougher route - with a Canadian for crying out loud.

You must be joking with the Canadian slurr, or suffer from Xenophobia.

Ryder has been a winning Races since he was 11 and if you look at his history you'll see he's not a one year wonder.

 

[D-Queued]

You must be joking with the Canadian slurr, or suffer from Xenophobia.

Ryder has been a winning Races since he was 11 and if you look at his history you'll see he's not a one year wonder.

You sure you want to go there? It is off topic. (nod, nod, wink, wink.)

Dave.

 

[Ripper]

Yes, we get it. However, why point out Sky? Consider Garmin, for example. Most of these points are true of them as well. For example,

They hired Allen Lim after Phonak/Landis affair.
They hire suspicious riders - admitted dopers and ones involved in Armstrong investigation.
Equal lack of transparency.
Won Giro - far tougher route - with a Canadian for crying out loud.

Garmin is supposed to be the new model. The accusations can be made far worse against most other teams. There is an unproportional hatred of Sky, which is seated in an irrational, visceral, and emotional reaction to the team. The doping stuff is largely a post hoc rationalization of that reaction...

Actually, Garmin has been called out for many of the reasons you noted. Not everyone has drank that Kool Aid.

This thread is not about Garmin, which is why is is not really the focus.

 

[Cloxxki]

You must be joking with the Canadian slurr, or suffer from Xenophobia.

Ryder has been a winning Races since he was 11 and if you look at his history you'll see he's not a one year wonder.

He was faster than Meirhaeghe at Lugano MTB worlds. Jsut too tall to stay on the bike on the downhills. Meirhaeghe we know about. Not sure what to make of that. He was a modest pro when he made the switch, and gradually got better. If he's a freak of nature, and the rest gets cleaner, yeah it could work out.
Sorry for the off-topic.

 

[Caruut]

Here is a thread all about Garmin. You can talk about them there, if you would like. http://forum.cyclingnews.com/showthread.php?t=17361. This is a thread about Sky.

 

[mastersracer]

You must be joking with the Canadian slurr, or suffer from Xenophobia.

Ryder has been a winning Races since he was 11 and if you look at his history you'll see he's not a one year wonder.

it was a joke - I grew up in Canada (racing against the fenwick flyer).

 

[mastersracer]

Nope, it's consistent with the view that doping use is independent of pre-doping result. If however one makes the radical assumption that performance enhancing drugs actually enhance performance that naturally lead to the conclusion that if PED use is uncorrelated with pre-PED result, PED use will be correlated with post-PED results. In fact logic dictates that every performance boosting factor (talent, training hard, low risk of injury, racing smart, PEDs and even dumb luck) is found more commonly in the top tier, because the people who have the most or best combination of those factors will by definition win.

This is demonstrably false. The distribution I indicated is an empirical one and disconfirms this view. If there is a distribution of talent, and if riders along that distribution have access to PEDs, then the effect is simply to shift the entire distribution, not a riders place along that distribution.

 

[zlev11]

if Garmin had Vande Velde, Stetina, and some other former mediocre rider dishing out a ridiculous pace on the Stelvio, shedding off all but 5 other riders from the lead group, then there'd be something to be suspicious about.

 

[Cerberus]

This is demonstrably false. The distribution I indicated is an empirical one and disconfirms this view. If there is a distribution of talent, and if riders along that distribution have access to PEDs, then the effect is simply to shift the entire distribution, not a riders place along that distribution.

If it's demonstrably false would you mind terribly demonstrating it? Because I really don't see how there can't be a difference in pre- and post-PED performance unless you insist that PED in fact don't do anything.

 

[Ferminal]

So everyone dopes or no one dopes? Everyone who does dope, dopes with the same intensity and the exact same timing? Everyone responds exactly the same to dope?

 

[Square-pedaller]

This is demonstrably false. The distribution I indicated is an empirical one and disconfirms this view. If there is a distribution of talent, and if riders along that distribution have access to PEDs, then the effect is simply to shift the entire distribution, not a riders place along that distribution.

No you're wrong.

Say there's a normal distribution of talent, and that without PEDs performance (and hence success) is perfectly correlated with talent. Now take half of the riders drawn uniformly uniormly from the talent distriibution and give them PEDs. If we assume that PEDs increase their performance by some fixed amount, that will displace half the performace distribution by that amount to the right. Now if you look at the numbers of clean and doping riders at any point along the performance distribution, the ratio of clean to doping will be high if you choose a low performance, and gradually decrease as the performance level you choose increases - which is what Cerberus claimed.

Only if all riders dope, and there is no variation in doping level or response level, will the riders doping or responding most not get concentrated among the relatively high performing riders.
Only if everyone dopes, and there is no variation in the amount of doping or response, do the best dopers or responders not become concentrated in the higher performing groups.

Something similar would be true if all riders doped, but the amount of doping varied and the mean and variation in doping level was constant across the talent distribution, or if all riders doped the same amount, but there was variation in the response, and the mean and variance in response didn't vary with talent. In both cases, although all riders doped, the highest performing riders would contain the riders using the highest average level of doping, or have the best average responses: in either case the 'winners' will contain those riders who profit relatively more from doping - either by doping more or being a better responder. If the gain from doping a lot/responding more is large enough it is perfectly rational to assume that winning is a sign of having doped or being a good responder.

 

[mastersracer]

If it's demonstrably false would you mind terribly demonstrating it? Because I really don't see how there can't be a difference in pre- and post-PED performance unless you insist that PED in fact don't do anything.

There's some distribution of talent across the peloton sample (to simplify suppose it is watts/kg at threshold) measured pre-PED. Riders now take some PED. Re-measure talent. The effect is that the entire distribution simply shifts right (mean watts/kg increases). No rider has moved in terms of their relative position. Given individual variability to PEDS, this is an idealization, but if that variability is itself randomly distributed, it will not have much effect.

 

[Square-pedaller]

but if that variability is itself randomly distributed, it will not have much effect.

No you're wrong. If doping is equally spread across talent, if there is any variation in (i) doping at all, (ii) the amount of doping, or (iii) theresponse to doping, the most successful riders will contain a higher proportion of (i) dopers, (ii) heavy dopers, or (iii) good responders. It's the variation that does this. The extent of the variation will determine the strength of the effect.

 

[Cerberus]

There's some distribution of talent across the peloton sample (to simplify suppose it is watts/kg at threshold) measured pre-PED. Riders now take some PED. Re-measure talent. The effect is that the entire distribution simply shifts right (mean watts/kg increases). No rider has moved in terms of their relative position. Given individual variability to PEDS, this is an idealization, but if that variability is itself randomly distributed, it will not have much effect.

Ah, but that is as Feminal says assuming everyone dopes, responds the same way and does it the exact same amount, which is much less likely to be true than for training because doping techniques are secret. Those assumptions are all wildly unrealistic individually, let alone all together. My point was as Square-pedaller pointed out that if a potion dopes that potion improves relatively.

 

[mastersracer]

No you're wrong.

Say there's a normal distribution of talent, and that without PEDs performance (and hence success) is perfectly correlated with talent. Now take half of the riders drawn uniformly uniormly from the talent distriibution and give them PEDs. If we assume that PEDs increase their performance by some fixed amount, that will displace half the performace distribution by that amount to the right. Now if you look at the numbers of clean and doping riders at any point along the performance distribution, the ratio of clean to doping will be high if you choose a low performance, and gradually decrease as the performance level you choose increases - which is what Cerberus claimed.

Only if all riders dope, and there is no variation in doping level or response level, will the riders doping or responding most not get concentrated among the relatively high performing riders.
Only if everyone dopes, and there is no variation in the amount of doping or response, do the best dopers or responders not become concentrated in the higher performing groups.

Something similar would be true if all riders doped, but the amount of doping varied and the mean and variation in doping level was constant across the talent distribution, or if all riders doped the same amount, but there was variation in the response, and the mean and variance in response didn't vary with talent. In both cases, although all riders doped, the highest performing riders would contain the riders using the highest average level of doping, or have the best average responses: in either case the 'winners' will contain those riders who profit relatively more from doping - either by doping more or being a better responder. If the gain from doping a lot/responding more is large enough it is perfectly rational to assume that winning is a sign of having doped or being a good responder.

I responded before you posted this and was illustrating with the simplest assumptions. However, your comments can be accommodated with some more realistic assumptions that I mentioned earlier in terms of the empirical distribution from actual doping positives data (e.g., the worst riders are most incentivized to dope).

 

[Square-pedaller]

I responded before you posted this and was illustrating with the simplest assumptions. However, your comments can be accommodated with some more realistic assumptions that I mentioned earlier in terms of the empirical distribution from actual doping positives data (e.g., the worst riders are most incentivized to dope).

Hi mastersracer, Nice to meet you.

You've moved the goalposts. I'm talking specifically about your claim that Cerberus was wrong. Cerberus is only wrong if, as Ferminal says, everybody (or nobody) dopes, they all dope the same amount (no variation) and they all respond the same amount (no variation).

You can't 'accommodate' this in any way. Cerberus was right, and you are wrong in claiming Cerberus was wrong.

 

[mastersracer]

No you're wrong. If doping is equally spread across talent, if there is any variation in (i) doping at all, (ii) the amount of doping, or (iii) theresponse to doping, the most successful riders will contain a higher proportion of (i) dopers, (ii) heavy dopers, or (iii) good responders. It's the variation that does this. The extent of the variation will determine the strength of the effect.

This is not the case if the variability is randomly distributed - will respond more fully.

 

[Square-pedaller]

This is not the case if the variability is randomly distributed.

No you're wrong. Imagine that the range in clean performance is 80-100 performance units, and that the range of doping levels gives a performance boost of 10-20 units depending on how much you dope.

Now, those who start out at 80 end up at 90-100. Those who start out at 90 wnd up at 100-110. Those who start out at 100 end up at 110-120.

Anybody with a performance of 120 has to have been doping the largest amount. Anybody with a peformance of 90 has to have been doping the minimum amount. In between the average level of doping increases from the minimum (at post-PED performance of 90) to maximum (at post-PED performance of 120).

 

[mastersracer]

No you're wrong. Imagine that the range in clean performance is 80-100 performance units, and that the range of doping levels gives a performance boost of 10-20 units depending on how much you dope.

Now, those who start out at 80 end up at 90-100. Those who start out at 90 wnd up at 100-110. Those who start out at 100 end up at 110-120.

Anybody with a performance of 120 has to have been doping the largest amount. Anybody with a peformance of 90 has to have been doping the minimum amount. In between the average level of doping increases from the minimum (at post-PED performance of 90) to maximum (at post-PED performance of 120).

short answer: you are assuming talent is normally distributed. I didn't make this assumption (because I don't think it's a good one for this sample).