A closer look at EPO-era performances

I was wrong? That was not one of the choices, and a conclusion can also be a fact.

Apparently for 100m, he compared two groups of 120 performances, and not top-10 runners.

Apparently for 1500m, your suggested slowdown post-ABP does not appear to be significant enough to be considered dissimilar in a deeper analysis of "equality of variances" and "equality of the shape of the distributions".

My guideline wasn't to look decade by decade. If you look at the 5000m (top-5 average) by 4-year Olympic cycles, you will see that non-Africans peaked in 2009-2012, with post-ABP performances. Peaked is a big word, as overall progress from 1981-2012 was rather small.

Another guideline is that you need to do more than juxtapose fast performances in an era with the doping reputation of that era.

You provide enough information to start a conversation, but not finish it.

rekrunner wrote:

My guideline wasn't to look decade by decade. If you look at the 5000m (top-5 average) by 4-year Olympic cycles, you will see that non-Africans peaked in 2009-2012, with post-ABP performances. Peaked is a big word, as overall progress from 1981-2012 was rather small.

The guideline comment was in reference to the "focus on male non-Africans".

2012? Interesting. Recall that the last blood-doping paper used 2013 as the border for pre- and post-ABP, because that's when it got serious and people were caught and limited their doping.

rekrunner wrote:

Another guideline is that you need to do more than juxtapose fast performances in an era with the doping reputation of that era.

That was to show correlation. The science is well established: have a look at the various papers discussed here on letsrun. Yes I know you tend to discard them because they rarely include elite runners.

I also recall another recent study comparing 2011 World Championship data with 2013 World Championship data, hypothesizing a reduction of blood doping estimation from 2011 to 2013, as a result of the deterrent effect of ABP prosecutions, and finding the data didn't support the hypothesis, with Russia being the rare exception.

Then, when you look at 5000m from 2009-2012, it was Solinsky, Tegenkamp, Ritz, and Rupp all running sub-13. Afterwards, Solinsky got injured. Tegenkamp retired. Ritz moved up to the marathon. Rupp's main event was the 10000m, and eventually he also moved up to the marathon. The next cycle between 2013-2016, with three new Americans, was as fast as the previous 4 cycles, from 1993-2008, +/- 1 second.

There are ways to show correlation of doping among the fastest performances, and yours is not one of them. Even for the sub-elite runners in the papers from well established science, we cannot draw the same conclusions you want about their fastest performances.

Read again:

"people ... limited their doping", not: fewer people doped.

casual obsever wrote:

Read again:

"people ... limited their doping", not: fewer people doped.

Read again my response.

Even limiting their doping would show a shift in the cumulative distribution curve. Note that for country "R" in that study on 2011 and 2013 world championship athletes (speculated by letsrun sleuths to be USA), the men's blood doping was average, and remained flat from 2011 to 2013.

You keep showing us the Sunday Times RET% scores graph, which shows a significant reduction in "unlimited" doping already occurred as far back as 2005-2008 -- also speculated to be due to improved EPO testing so fearful that it frightened El G into retirement:

- 2005-2008 was a cycle where non-Africans got faster in 5000m.

- The "unlimited" doping of 2001-2004 corresponds with the slowest cycle since the "unlimited" doping of 1993-1996.

- These two slow cycles were slightly faster than 1981-1984, especially compared to previous jumps in improvement. (*)

For perspective, these top-5 averages over the above mentioned 4-year Olympic cycles were all slower than Dave Moorcroft ran in 1982. What you called "hugely" was in fact steady, but slow and small improvements over three decades while unlimited EPO came and went.

In any case, "limited their doping" remains your own personal speculation. It is your own custom rationalization to modify a hypothesis to better accomodate new contradictory information, but it remains an unconfirmed hypothesis, which apparently doesn't fit other known data.

(*) All of this data on 4-year cycles of non-Africans comes from another thread "A longer deeper look at all-time top performances" which addressed some of the issues identified with the "closer look" of this thread, and shows trends with more resolution than your decade by decade figures:

What's with all the troll responses? You argue against stuff I haven't said, and then confirm what I have said, to then conclude that I am wrong.

You wrote:

"the data didn't support the hypothesis"

but I didn't make the hypothesis that fewer people doped.

The observation, not hypothesis, that people limited their doping is for example a combination of the 2011 and 2013 data you just cited, plus the numbers from 2001 - 2012, with the Times "Sunday Times RET% scores graph", and similar ones that Tucker collected. From that we learned that

a) the number of blood dopers did not change systematically (as you just repeated!)

b) the numbers of extreme blood values went down (as you just repeated!)

Somehow you manage to argue against both a) and b) anyway...

And finally, you troll me with

And finish with making it personal while lying: "your own custom rationalization". Not worth my time, but thank you for trying.

You said, and asked me to re-read "Recall that the last blood-doping paper used 2013 as the border for pre- and post-ABP, because that's when it got serious and people were caught and limited their doping."

With all due respect, I'm just trying to make sense of the timeframe here, as it relates to the non-African men's 5000m. And, since you said that doping and the best performances were correlated, I'm looking for any signs of correlation in the data we know.

The 2011 data and 2013 data doesn't supported "limited their doping" either, with Russia being an exception. For most countries, their was no significant change. The metric used doesn't break down into "number of blood dopers" or "number of extreme values", so I don't see how where "limited their doping" is coming from, with respect to 2012.

Since there was a decline in Russia, it does make some sense to look at Russian women pre-2012 versus post-2012. But generally, and specifically for the USA men (who drove the 2012 period and 2016 period 5000m times), the data was stable, and furthermore, the USA data was not particularly high, but rather both average.

When you say I argued for b), I argued against it for 2012 (except for Russia), and for it in 2005.

Then I looked for the correlation to performance over all these periods dating back to 1993, finding no correlation, or sometimes a negative correlation, with non-African men's 5000m performances.

Looking back on this post, it's almost mindbogglingly unscientific.

So because the top 5 from 5 continents combined did not improve significantly in the EPO era, or as much as East Africans, that means EPO doesn't work?

Jesus.

And imagine the time it took for him to compile these 'statistics'.

Ok Jesus, give us the 'real science'

Does doping with EPO facilitate oxygen delivery beyond natural means? Before you answer, bear in mind the implications of what I am asking.

Given that British performance actually declined after 1990, I guess Rekrunner should draw the scientific conclusion that EPO not only doesn't work on Brits, but actually makes them slower.

Or maybe a more rational conclusion would be that

1/ The most talented Brits did not take EPO. (that 'everybody dopes at 53% is patent nonsense)

and

2/ There was less British talent competing in the EPO era - possibly as a result of the fact that East and North Africans were doping their heads off and running a second a lap faster than even Coe, Cram, and Ovett.

To some extent, the above was probably true for most of the 'rest of the world' outside of Iberia, North Africa, and East Africa.

The rest of the world did not dope to the extent of Iberia and Africa.

The rest of the world's performances actually dropped because of the demotivating effects of being dominated by EPO cheats, as well as not reaching finals or being invited to top races in order to run fast times (because they were filled by African and Iberian EPO heads).

This is just common sense. Just like it's common sense to see that WRs that were set by 'GOATs that all happened to appear in the small time-frame of the EPO era, and lasted decades or even to this day, were fuelled by EPO, because EPO evidently works.

It didn't take too much time -- a few hours here and there over the course of a week -- I just imported data into Excel and manipulated it with filters and formulas. I never suggested this was something worthy of scientific publication. It's really just calculating averages -- something we learned in 3rd grade.

I'm wondering though, where in there did you see any conclusion from me that "that means EPO doesn't work?"

My conclusion in the quoted post was something that struck "me as most anomalous, when you look at athletes from "5 continents"", leading me to conclude with a question (eventually paraphrased): "why so few, and then by so little?". In between insults and criticisms, I did receive a few illuminating responses.

If I didn't see the hypothesized effect in my data analysis, one possible response is to suggest another way to compile and view the data.

I also performed a second analysis, attempting to respond to some of the criticisms and got a better picture of the trends over time:

Scientifically, I cannot draw any conclusion about EPO performance effects from a set of data that only includes a single best performance for each athlete, with no information about whether it was blood-doped.

Seems like the most rational conclusion is that no correlation (let alone causation) can be concluded that cannot be observed in the data.

As hard as we might want to suggest that EPO was the only significant factor in the "EPO-era", we also need to consider the potential effect of a dozen or so other factors at the same time.

This was also partly inspired by the "mindbogglingly unscientific" reason we think "EPO works" for East Africans -- the massive improvement in track times in the '90s.

Sounds great! Now do this with legalized epo (aka cheat-shoes/spikes).

Hahaha. So much nonsense. We know EPO works because of the science and the scientists' studies and the coaches' and athletes' statements and the coaches' and athletes' actual experiences. The massive improvement in track times in the '90s is just nice corroboration of the above.

And now, with drastically enhanced testing (AIU) in Africa, the African track performances don't stand out that much anymore. Surprise!

This thread was kind of done with that ^^^^

The thread questions were really answered with this from page 2. No need for 43 pages.