Well, today I concluded something. I'd thought it for quite a while, but now I think - until I see more evidence pointing the other way - the case is more or less closed. Here's why.
Probably a fair few people here follow the Science of Sport guys. So I'm not the only one who's seen the graph they recently linked to of Alpe d'Huez ascent times. It's a great story-telling graph. [It was put together by somebody called 'ASimmons', and all credit should go to them, and none to me]. Here it is. It shows the five highest ascent times for the Alpe (the graph cuts off at 42 and a half minutes).
Now, what jumps out at you with this graph is the history of modern 'extreme' doping. Pre-1990, the times are all super-41'45, most are super-42', and in fact in none of the three pre-'90 years were there five finishing times sub-42'30 - 1 rider in '82 and '89, three riders in '87. We can't really say that this is riding before doping - Fignon and Delgado probably aren't high on anyone's list of 'cleanest of all time'. But we can say that this is what pre-modern doping (or cleanliness, you never know) might look like.
During the nineties, there's a big step up. At first, the leaders all together up their game and end up between 39'30 and 40'30. After three runs in the eighties where winning times were all within about 30 seconds of each other, these early '90s guys were ALL a minute and a half faster than the winners from the eighties. From then on, the nineties are raced at two speeds - Pantani and everybody else. Pantani immediately goes incredibly fast, more or less stays at that level, and then dies. Everyone else is more cautious, and they gradually ramp up the speed, and don't die.
Then there's 1999, and it's incredible how slow everyone is. The winner in '99 is more than 4 minutes slower than the winner in '97 - but even more incredibly, the winner is almost 2 minutes slower than the 5th-placed rider in '97! These guys are still significantly faster than the riders from the eighties, but slower than everyone since. And after '99, the race goes back to being held at two speeds. On the one hand, there's Armstrong. He quickly goes back to the insane speeds of the pre-Festina era - he never reaches Pantani levels, but he's in the region of Ullrich, Indurain, Zulle, Riis, Virenque. Everyone else, however, goes more slowly - crucially, even guys like Ullrich and Zulle are a lot slower than they were a few years earlier. But their speeds ramp up - they never get back to late-90s speeds, but they do reach mid-90s speeds.
So what we see so far seems to match what we're told about doping. Through the nineties it reached ludicrous levels; after Festina, people cleaned up their act; Armstrong returned to the old levels quickly, everyone else was more cautious but did gradually increase their doping levels.
And then in 2006 everyone stopped doping, and speeds plummeted again.
To me, it looks like this graph can be divided up into 'classes' of riders:
Class A: Pantani - 36'30 > 37'30
Class B: Festina-era patrons, and Armstrong - 37'30 > 38'30
Class C: Second tier (top riders from the early '90s, and from the later Armstrong years other than Armstrong, and from 2006) - 38'30 > 40'00
Class D: Hangers back (other early '90s riders, early Armstrong rivals) - 40'00 > 41'15
Class E: Boring riders (everyone else from 1990 to 2010, plus Contador and Sanchez in 2011) - 41'15 > 41'45
Class F: Shuffling zombies (pre-1990, plus 2011 minus Contador and Sanchez).
Now, it's only natural that times would improve over time, with improved training and equipment. So maybe we should see the ~41'30 times of recent years as the equivalent of the ~42'00 times of pre-1990. And of course there are always individual stories for each ride.
But after we've finished looking at the overall story, there's a time here that really stands out. It didn't stand out so much at the time, but it's a horrible sore thumb in hindsight.Yes, I am of course referring to Mr Carlos "Unimpeachable" Sastre, he of the heart and virtue whiter, so the common story goes, than the driven snow.
Cover up, if you can, every result from 1990 to 2006. What do you have left? 15 rides. Take out Sastre, and every one of those rides took between ~41'15 and ~42'15, which also covers all but one of the rides from the post-Festina Tour in 1999. I'll reiterate - take out Sastre, and of the remaining 13 rides since Beat Breu in 1982 - actually, let's say the remaining 21 rides, it's just the other 6 are too slow for this graph - none of the 21 riders since Beat Breu have been able to beat Breu's time by more than a minute. Fignon and Herrera couldn't beat it by more than 30 seconds, and nor could Valverde*, Frank Schleck*, Rolland, Velits, or Evans. Tour-winner Evans was only slightly faster than Breu. Contador* and Sanchez managed to be almost a minute faster than Breu - just like the post-Festina 1999 versions of Zulle, Tonkov, Escartin, and Armstrong.
Sastre was the Valverde/Fignon half-minute faster than Breu. And then the Contador/Zulle extra half-minute too. And then another half-minute faster than them, up to the region of Vinokourov, Moreau, and Sevilla. And half a minute more, up to the speed of Beloki, and 2001-era Ullrich. And then another three quarters of a minute faster still, up to the level of Indurain.
I'll repeat that: everyone else on the graph pre-1990 or post-2006 is within about a minute of each other. Sastre is another minute and a half faster than the fastest of them. The only winning margins that compare are Armstrong in 2001 and Pantani in 1994. To put it another way: the difference between Sastre and all other post-2006 riders is bigger than the difference between pre-EPO riders and early-EPO-era riders. Compare that ride in 2008 to previous rides, and it doesn't stand out at all (almost a minute slower than Kloeden two years earlier). But compare it to post-2006 rides and it requires divine intervention.
In fact, the most damning thing of all about that dot is how well it fits next to the ones that precede it. It's not just the gap between it and the other post-EPO rides. It's the way the dots cuddle together so pleasantly into two little clusters - one of post-2001 top dopers (with Armstrong as an outlier), and one of post-2006 top 'clean' (or at least 'cleaner') riders. And Sastre is quite clearly in the first cluster, not the second.
Sastre was a doper.
(And I strongly suspect Chiapucci was too).
Of course, this year we'll get some more dots for the graph. But I'll say ahead of time: if anyone goes faster than 41'00, they're doping. If they go slower than 42'00, I'll be happy. Faster than 41'30, I'll be suspicious.