FanPost

The Time Trial and the Tour de France: An Inconstant Love Affair, in Graphs

Bradley Wiggins has won the Tour de France - but does his accomplishment really count? After all, as some people have said, the course was almost tailor-made for his skill-set. The parcours this year was almost freakishly laden with time trial kilometres - a Tour designed specifically to be 'given' or 'thrown', as people have put it, to Wiggins (or, perhaps, to Cadel Evans).

Except, of course, that that's not true at all. The amount of time trialing varies from year to year, and though this year's Tour was certainly more TT-heavy than the last couple of editions, you don't have to go all that far back to find a course with a similar number of TT kilometres. In fact, you only need to go back to 2009.

Well, that got me wondering. I'm no great historian of the Tour, and I didn't have much sense of how this had changed over time - where did this year sit in terms of 'normal' Tours? Was this an exceptional year, or a return to normality?

As it turns out, the truth is rather more complicated than either of those options. Maybe everybody knows about this already, but I didn't, so having done some 'research' (read: 'having skim-read some pages of Wikipedia'), I threw together some hopefully-understandable graphs; and who knows, maybe some of you aren't entirely up on your Tour history either, so here I am to share the fruit of my 'labours' (read: 'short period of time fiddling with a spreadsheet to avoid doing more important things').

Pretty pictures, and some explanations, after the break...

Tt13_medium

(because I'm interested in tendencies rather than absolutes, I didn't bother with a scale - and the three pieces of information are scaled to be pleasing to the eye)

Hmm. Well that's odd. To be honest, it's hard to really see much detail in this graph, because it's so dominated by the late 1920s. What the hell happened there?

Well, it turns out that the time trial was introduced into the Tour in 1927, with a 180km TT from Paris to Dieppe. That alone would have been notable, but the route-designers instantly followed it up with a TT from Dieppe to Le Havre. And then from Le Havre to Caen. And so on. And so forth. In fact, I think it's fair to suggest that the organisers may have got, shall we say, just a teensy-bit over-excited by their sexy new format. In 1927, there were 16 time-trials in total, covering a distance of about 2940km, or 55% of the entire race, including a 285km TT from Les Sables d'Olonne to Bordeaux. Ouch. The only consolation for the riders was these were not solo efforts - though they were not quite conventional team time-trials either. Instead, the riders were sent out in teams, but each rider was given their own time, and thus were theoretically free to ride as fast as they could, ignoring their team-mates. As it happens, the rider who made that mistake on behalf of the peleton was race-leader Ferdinand Le Drogo, who, enthused by local support in his home region, surged ahead of his team on the seventh stage, only to over-exert himself and drop twenty minutes to his rivals.

This early experiment with the time trials was not, however, ultimately about deciding the winner of the Tour de France. Yes, big names won some of these stages (Pélessier won the first, and Frantz won Stage 21), but the Tour that year was still decided in the mountains - Frantz seized control on the road to Luchon and never let go. About all the TTs did was whittle down some of the weaker riders - and since the move to TTs was accompanied by a dramatic reduction in average stage length, some riders may actually have been able to stay in longer because of them. No, TTs were an attempt to eliminate, once and for all, the great scourge of the Tour de France: sprinting. In recent years, you see, riders had increasingly started to travel as a 'peloton', barely attacking each other at all on the flatter stages, and hoping to gain some time or prestige in a high-speed charge near the end of the day. How dull. This, sirs, was meant to be an ultimate test of the rider as an individual, a display of man fighting against man, at the maximum limit of his powers, straining his sinews to bursting point every minute of every hour, constantly striving to land the killer blow. This was not some family excursion for sight-seeing purposes, thank you very much, where men might amble along amiably chatting to each other before energising themselves in the evening with a pleasant little friendly dash to the line! I think not! That, sirs, is tantamount to cheating in my book, and the organisers of the greatest sporting event on earth will be having none of it. A fig for your 'sprinting'!

So, since Desgrange could not eliminate flat stages altogether, he decided to break up the big rivals, letting the riders go in bunches organised by team (those not in teams had to find a way to co-operate with one another instead), the aim not so much to find the best cyclist (the TTs did not result in large time gaps in most cases), but simply to force people to ride. If nobody could see their rivals, he reasoned, they would not know how hard the other man was pushing, and so would be forced to ride as fast as possible just in case.

Unfortunately, Desgrange's new idea didn't work. Yes, it made riders race faster, but it clearly distorted the results, biasing the whole race in favour of riders from stronger teams and potentially allowing a weaker rider from a stronger team to defeat the strongest rider in the race. He tried again in 1928, this time introducing a substitution rule to help small teams replace their worn-out domestiques - but the results were even worse. Substitutions helped the big teams dominate even more, with Frantz's Alcyon team taking 12 out of 22 stages, and all three steps of the podium. He tried a third time in 1929, reducing the TT distance to a paltry 620km, but he still hadn't found a formula he liked, and the time trial was abandoned forever.

...or until 1934, at least. No, the time trial would never again be seen at the Tour de France in either the length or the plenitude of the '27-'29 glory years, but it wasn't dead just yet. In fact, a wholly new breed of TT was seen in 1934: the individual time trial. Oh, sure, it wasn't the first ITT ever. Or even the first popular and famous ITT. Actually, Desgrange stole the idea from the GP des Nations, a rival event starting by Paris-Soir, themselves rivals to the Tour's sponsers, L'Auto. But who cares about that - if it didn't happen in the Tour de France, it didn't happen. So the first Individual Time Trial Ever, Honestly, was a relatively benign little 90km tester, Stage 21B, from La Roche sur Yon to Nantes. The winner was Frantz's old team-mate, Antononin Magne, who had led from the second stage of the race, and who used the TT to hammer eight minutes into his closest rival, almost a quarter of his final winning margin. For once, this TT mattered - not because it decided the race, which it didn't, but because it could have done, if the race had been closer. In an era when top riders were increasingly able to put the race beyond contestation in the mountain stages (a problem that had been worrying Desgrange since the time of the first TTs), and in which team support was increasingly vital (particularly in a time of national teams - at that point, a French rider had won every Tour de France since the introduction of national teams (itself a last-ditch attempt to end the domination by Alcyon)), a way to neutralise both the climbers and the big teams was desparately needed in the interests of... well, of maintaining interest. The TT was that savior: a race of truth, where neither the advantages of blood and money nor the unjust and peculiar proclivities of climbing specialists could steal victory from the one man who merited it most: The Strongest.

And to see what happened next, we need a new graph, that edits out that wild brain-fever of the 1920s…

Tt22_medium

Since the big bang in 1934, there have been, by my reckoning, six eras of TTing. The first era shows an exuberant embrace of the format. The proportion of the Tour allocated to TTs instantly soared, with the new ITTs supplemented with 20s-style semi-individual TTs now rebalanced into shorter lengths in emulation of the ITTs. This formula continued until the Second World War; and yet the organisers clearly were not wholly convinced. In 1935, Romain Maes tore his opponants apart to win by over 17 minutes. In 1936, Sylvère Maes (no relation) won by nearly 27 minutes, helped hugely by Belgian dominance of the TTT (only the final TT, on the penultimate day, was individual - a format that has endured). Perhaps in order to reduce this dominance, throughout this era the organisers gradually reduced both TT length and TT number (similar measures were taken to neuter mountain stages). 1938 saw the nadir - no TTT, and only two ITTs, neither longer than 45km. It was a year that saw a striking duel between a TT specialist (the Belgian Vervaecke) and a climber (the Italian Bartali) - again, a common feature in years to come.

But times changed. The late thirties saw the arrival of a new director with new ideas, and there was some unpleasantness around that time in europe that disrupted a number of bicycle races and other sporting events - the Tour de France was not spared. Anyway, once the inconveniences had been gotten out of the way, the Tour's approach to the time trial entered a new phase. The TTs dwindled from a handful to only one or two each edition - yet this was not necessarily good news for those who struggled in them: the reduction in number was compensated for by a terrifying increase in length. Not, of course, back to the strange old days of the experimental twenties, but nonetheless the change must have been intimidating. In 1939, the penultimate day saw a TT of 59km. In the next edition, in 1947, the penultimate day saw a TT of a full 139km, 135% longer. It was nearly 50% longer than the longest TT since 1929. And although it's true that high-watermark that has only once been surpassed, it set the tone for half a decade of a very distinct pattern of TTing: few in number, horrific in length. Perhaps the worst year for the pure climbers came in 1949, when, not content with imposing a 92km TT in the first week, they went and stuck a 137km TT on the penultimate day. From Colmar to Nancy. For those of you, like me, who don't know Colmar and Nancy... well, if they'd gone in a straight line there'd have been a 1200m mountain in the way. I'm guessing they tried to go around that, but comparing the straight-line distance to the TT distance, I'd be surprised if they entirely succeeded. This was NOT a pancake.

By the way, if anyone's still bitter about Wiggins beating a team-mate who may have been a stronger climber, thanks to his prowess in the time trial: don't get too excited, this isn't the first time. In 1949, teammates Fausto Coppi and Gino Bartali was leagues ahead of any competition (much like Sky in 2012), but Coppi was clearly second-best in the mountains. Fortunately for him, he had 229km of TT to play with (Wiggins had 101), and he used it to claw back a massive eleven and a half minutes from his goatlike colleague. The final victory margin? Eleven minutes. A course with an abnormally high amount of TTing had allowed Coppi to turn a narrow defeat into a crushing victory.

You heard it here first. Bradley Wiggins: today's Fausto Coppi. [No accusations intended...]

In any case, the TT-heavy courses served up for Coppi would not last forever. 1951, the awesome power of Hugo Koblet seemed to make the other competitors superfluous – nobody could come close to matching him in the 182km of time trial, and although he only gained a handful of minutes over some specialists, he put 6 minutes into his eventual podium companions in the first TT, and 11 and 13 minutes in the second. Coupled with a solid performance in the mountains, Koblet’s TTing rendered him unbeatable, it seemed. [And if you think ‘awesome’ is too strong a word, consider this: on Stage 11, Koblet attacked on a flat stage, and held off both the sprinters and the heads of state to win the stage. Like Compiègne, I suppose. Except that Koblet attacked with a staggering 135km to go…]. In 1952, then, the organisers seem to have put a bullseye on the back of the Swiss trackie (in ’51, Koblet was the Swiss Pursuit champion, and took bronze at the track World Championships, again in the Pursuit). The TTs were dialed back from 182km to 123km – significantly shorter than the last three years, and the second-shortest TT distance at that time since 1935. What’s more, the climbers were given a massive boost through the introduction of mountaintop finishes.

It didn’t work. Koblet was injured and did not attend. In his absence, the clear favourite was the world’s second-best TTist, Fausto Coppi, who had been present in body but not always in spirit in ’51 (perhaps because he was mourning his brother; perhaps because he ate bad fish; perhaps because of an adverse doping reaction; whatever the reason, he was never at his best, and lost half an hour on a flat stage). Unfortunately for the organisers, the world’s second-best TTist was also the world’s best climber. He took the first TT, took back-to-back MTF stage wins on Alpe d’Huez and at Sestriere, at was 24 minutes up on his nearest rival after only 11 stages; free to cruise in the second half of the Tour, he only picked up two more stage wins (including an MTF on the Puy de Dôme), and sauntered around the final TT in relaxed fashion. The final tally: five stage wins including all three mountaintop finishes, the King of the Mountains jersey, the Team Classification, and a final winning margin of over 28 minutes. Perhaps more important than the winning margin (never again replicated) was the philosophical significance of that first finish on Alpe d’Huez. Tours could now be decided in the great gladiatorial arenas of the high peaks – the ‘race of truth’ had become superfluous. All it did was add to the already mighty winning margins of the greatest riders. In 1953, faced with four imperatives (excessive winning margins, complaints from the riders over the difficulty of the courses, particularly with the new MTFs, nascent concern over the role of PEDs, and the complete absence of any French winners), the organisers set out to castrate the Tour, making it both shorter and easier. The most obvious change was the overall length – slashed from an impressive 4807km in ’52 to a paltry 4479 in ’53, returning the Tour to the type of distance more typical of the early 1930s; and, less ostentatiously, a second victim was the TT. In 1951 there had been 182km; in 1952, there had been 123km; in 1953… only 70km.

This was the third era. Every year from 1953 to 1962, the main TT would be between 65 and 85km, generally toward the lower end of that spectrum. After ’53, would be one of at least two TTs – for 1954 was the year the TTT returned to the Tour. Or possibly made its debut at the Tour depending on definitions, because now, for the first time, times were awarded to teams, and not to individuals. These TTs, however, were kept minimised – between 10 and 20km, generally toward the lower end.

This format wasn’t fixed, and over time the organisers clearly came to want more TTing in their Tour. So, in ’57-’59 an additional short TT was added (9.8km, 21km, and 12km respectively), and ’58 and ’59 saw the short ITT turned into a ‘mountain TT’, with the initial TTT converted to yet another ITT (up to 46km and 45km). In ’60 and ’61, organisers fiddled again, cutting the mountain ITT, and shifting some length from the first to the second ITT. Finally, 1962 saw all the ingredients of the previous era rolled into one parcours: an initial short TT; a first ITT of 43km; a short mountain ITT (in this case to Superbagnères); and a final ‘long’ ITT of 68km.

Although this era had fewer and easier TTs than in the past, its latter half was nonetheless dominated by a time trial specialist: Jacques Anquetil. Anquetil won every Tour de France ITT in 1957, 1961, 1962, 1963, and 1964, with the exception of one mountain TT won by Bahamontes. It was no coincidence that those were the five years he won the Tour – in the years he failed to win the Tour, he failed to win any ITT. After Coppi’s victory in 1952, no Tour winner had won the polka dot jersey in the same year, with the exception of Bahamontes’ double victory in ’58. Perhaps the perfect demonstration of the age is Charly Gaul’s win in 1957. The Angel of the Mountains, he was called – he’d won the KOM twice already. So how did he win the Tour de France exactly? Well, forget the myths about his climbing, and forget the unforgettable destruction of the peloton on the slopes of Luitel. All that was just an attempt to keep up, to recoup the losses after days and days of defeat, both in the mountains and on the flats. No, despite the immortal heroics, and contrary to how the story is sometimes told, Gaul came out of the mountains that year down in third place; on the morning of the final ITT, he was more than a minute behind the leader, Vito Favero, and thirty seconds behind Géminiani. Twenty-four hours later, Gaul set off on the final stage to Paris, three minutes ten ahead of Favero; three forty ahead of Géminiani. There were three ITTs that year, and Charly Gaul, the Angel of the Mountain, the purest of pure climbers (at least as portrayed in our collective memory) won all three of them.

So, after a difficult post-war time for the French, in which Italian riders dominated difficult courses, organisers made the Tour a little easier. If their aim was to provide a French winner, they were more than successful. The French won the Tour in ’53, ’54, ’55, ’56, ’57, ’61, ’62, ’63 and ’64, and would go on to win again in ’66 and ’67. [An aside: the man who first took control of the new, ‘easier’ Tour, winning three consecutive titles, was Louison ‘Cry-Baby’ Bobet, so-named because when first confronted with the Alps in the Coppi years, he burst out crying and went home]. But were they satisfied? Not entirely. You see, it was the wrong Frenchman winning. Anquetil had never been the most popular of riders, and by the later years of his reign, organisers were getting plain fed-up with his annual steamrollering of the opposition. Illness in ’58, post-Giro tiredness in ’59 and a failure to anticipate the serious challenge of Bahamontes (previously a KOM contender only, who had rarely even made the top ten), and injury in ’60, were all that prevented Anquetil from winning eight Tours in succession. Worse, he had done it in a way that was both ‘passionless’ and borderline treasonous, repeatedly riding vindictively against his own compatriots (most infamously in 1959, when he rode flat-out to ensure that Bahamontes beat surprise Frenchman Henri Anglade). The public – and the organisers – did not want Anquetil and his metronomic TT victories. They wanted fireworks, they wanted riders like Bahamontes and Gaul. Most of all, of course, they wanted Poulidor. In the early years of the sixties, the organisers took action – they would end the domination of time trialists. And they chose to do it by… increasing the number of time trials.

A strange choice? Maybe. And yet it did make sense, in a way. Increasing the number of time trials allowed them to reduce the length of the time trials without drastically reducing the overall percentage of the race given over to them. This gave audiences just as much TTing – and brought in more spectators, since the TTing was now spread over more locations – but because each TT was shorter, there was less opportunity for the specialist to gain time on his rivals, since exhaustion was less of a factor.

There wasn’t a ‘big bang’ to start this new era, but the trends are clear. Before 1963, only one Tour (1938) since the introduction of the ITT had failed to have time trial more than 60km in length – and most had seen TTs of at least 70km. After 1963, until 1978, not a single Tour had a time trial longer than 60km, and many failed to have any time trial longer than 50km. At the nadir, 1973, the longest TT was under 29km. And at the same time, no Tour in this fourth era had fewer than three TTs, with some (1970, 1977) having up to six; by contrast, from 1947 to 1961, only two Tours had as many as three TTs, with some (’47, ’48, ’53) having only one. I’ll make the change clearer: in 1947, there was a single TT, 139km in length; in 1977, there were SIX TTs, but they added up to only 109km in length. Some years within this fourth era also experimented with very low total TT distances – 1974, for instance, had only 65km in total, until that time the shortest distance ever.

The era wasn’t completely level. It seems there was a bit of a sawtooth – steep decline in TTing, a rise back, a gradual decline, and then quite a steep rise back up. It doesn’t take a genius to work out why this was. The intial decline is the final years of Anquetil’s dominance, and a period when organisers weren’t yet sure whether they’d completely killed him off or whether he might return from the dead to haunt them yet again. Once he was safely out of the way, TTing was cautiously re-introduced… only for organisers to run straight into their worst nightmare. Anquetil was back in a new body, and this time he wasn’t even French: this time, he was called Eddy Merckx. Merckx, needless to say, smashed all his rivals in the time trials. The only exception was in 1974, when, shockingly, he managed only second place in the TT. By that time, TTing had been almost eliminated in a vain attempt to stop him. Once he was gone, however, it began to return.

A slight difficulty arises in my graphs at this point, at the transition into the fifth era. In 1978, you see, somebody somewhere in charge of course design went temporarily doolally, and decided that what the Tour really needed was a 153km TTT. In order to prevent this from destroying race balance, this was neutered by ignoring time gaps at the end – in effect, this was a race-outside-the-race, with the winning teams only gaining a time bonus. The same technique has been used off and on since (or the equivalent method of giving losing teams only fixed time penalties), which may be misleading if you look only at the raw numbers. For the purposes of these graphs, I’ve substracted the length of these bonus-TTs both from the TT distance and from the total distance of the Tour that year, but I’ve counted them as TT stages for the purposes of the raw ‘number of TT stages’ data. However, it’s possible I’ve missed some info for one or more of these years and counted kms that I shouldn’t have counted…

But speaking of graphs!

Tt32_medium

(The black line here shows the rolling three-year average TT percentage, and the orange arrow-line shows the linear trendline that fits that curve. The thinner line shows the actual TT percentage each year – red for increase, green for decrease).

It’s hard to understate the sudden change in the Tour from 1978 onward. The TT percentage soared from 2.7% in ’77 up to 9.2% in ’79. It was a categorical shift – every single year from 1960 to 1976 had a lower TT percentage than any single year from 1978 to 2004 (with one exception – the least TT-ish year of the later period was SLIGHTLY less TTish than the most TTish year of the earlier period). And that’s not even counting the long TTTs raced only for time bonuses.

Why the change? I don’t know. But it’s hard to overlook the issue of nationality. With Thévenet, suffering side-effects from his doping regime, considered past his best and likely not a contender to retain his crown, the hope of France rested on the young debutant shoulders of Bernard Hinault; but the favourite must surely have been the Dutchman, Joop Zoetemelk, three-time runner-up. How could Hinault beat this fine TTer, who was widely considered his superior in the mountains? Easy. Zoetemelk was good in a time trial, but Hinault was the best – quite literally, as by the time of the Tour he had become champion of the GP des Nations, widely considered the effective world championship for time trialing (former winners included Coppi, Koblet, Bobet, Anquetil, Poulidor, Gimondi, and Merckx; Hinault would end up winning the event five times, second only to Anquetil’s nine). So, the 1978 Tour was the second-shortest yet (the shortest yet, if the untimed TTT is excluded), and the TT percentage leapt. Hinault won. 1979 was even shorter and even more dominated by TTs, and Hinault won again.

Hinault’s reliance on the TT can be seen quite strikingly. In a period of eight years, Hinault finished six Tours, four of which had TT percentages over 5%. He won all four. In the remaining two Tours, with lower TT percentages, Hinault won only 50% - and that was the year he had no serious rivals.

In any case, more or less the same average TT percentage was retained from 1980 through to 2003 (I use the big thick black line to show the three-year average in order to help show the general trend in this era, rather than the yearly oscillations). And sure enough, the Tour kept being won by TT specialists – Hinault, Zoetemelk, Fignon, LeMond, Roche, Delgado, Indurain, Ullrich, and Armstrong, all of whom won at least one long TT at the Tour. The only exception to this TTist domination was Bjarne Riis, himself no slouch, but who could manage no better than fourth in the long TT the year of his victory.

Of course, looking at the graph, you can see I’ve missed something out. Yes, the average remained high – but for half a decade there was a strange aberration, and TT percentages dropped precipitously, almost back to pre-Hinault days. Why? I would guess simply frustration with Indurain’s constant victories. He was tolerated until 1995, but 1996 looks like a determined attempt to prevent him from surpassing the records of Hinault and Anquetil – and it worked. That was the year of Riis – the least TT-centric winner of the era. The organisers played around with the idea for a couple more years, but I suspect that the dollars lining up behind Lance Armstrong and his miraculous return from cancer cut short any thoughts of a return to the seventies. We often think of the French as hostile to Armstrong, but once he had crushed the TT in 1999, they threw the parcours in his favour as though the man were a born Frenchman – the TT specialist gifted TT-heavy courses up until 2003, a generosity that extended to the return of the TTT. TTist Armstrong on the strongest team around couldn’t have asked for more!

But, while the Tour may have been excited at first to see all the rich new fans from America, sooner or later they had to tire of the Texan’s monotonous crushing victories – and so, like Anquetil before him, Armstrong in his last few years had to deal with courses rather less pleasant for a time trialist – and now with the TTT neutered once again into a race for bonus time only. The great era that stretched from Hinault to Armstrong, via a brief detour in the mid-1990s, was on the way out.

So here we are. The sixth era. And what a strange and inverted world we are living in. TT length has continued to slowly decline. TT percentage has oscillated quite wildly, but largely seems to be continuing the ‘90s trend toward a return to pre-Hinault courses. Indeed, perhaps a better way to analyse it would be to say that the fifth era ended with Indurain, and that the sixth has been a trend toward shorter TTs ever since – with the exception of a handful of years designed to favour Armstrong.

And where does that leave Wiggins? Well, nowhere remarkable. A sharp uptick, yes, but not something that really stands out even in the post-Armstrong era. In the broader context of the last thirty years, Wiggins actually won a pretty TT-light edition. Then again, in the even broader context of post-war racing, this year doesn’t really stand out at all – there have been courses with a lot more TTing, and courses with a lot less as well.

Unexceptional. Or is it? Here’s one last graph, this time showing the relative changes in TT percentage, expressed as each year’s percentage as a percentage of the three-year average of percentages… ok, that didn’t make sense, did it? Basically, for each year, I’ve established how TT-ish it should have been, based on the average of the last three years, and then compared that to how TTish it actually was. So, above the line means there was an unexpected increase in TTing, as a proportion of the Tour, and below the line means there was a decrease. Here you go…

Tt42_medium

You can see a lot of what I’ve been talking about on this graph. There, for instance, is the prolonged dip after the Coppi/Koblet TT domination, and there’s the Anquetil rise. There’s the relatively gentle but prolonged ‘we’re fed up with Anquetil and his ilk, let’s try something different’ trough, the ‘Merckx is gone, we can use TTs again!’ col leading up to the hors categorie ‘Hinault is good at TTing! Hinault is good at TTing!’ patriotic summit. The ‘TTs are boring now, let’s have some more killer mountains instead’ EPO depression, the ‘oh, wait, our bank manager tells us that Armstrong’s fans are spending more than Pantani’s’ mound, the ‘OK, that’s enough of that now’ slump, the brief ‘Armstrong’s gone, we can use TTs again!’ rally, the ‘does anyone really want to see Cadel Evans and Levi Leipheimer fighting for a win when you could have Contador and Rasmussen?’ bust, and, at the end, the Wiggins surge.

And now it does look unusual. First off, the heigh of that final surge is quite big – it’s been a big rise in TTing. And, more importantly, look how steep it is. Since the war, I think that only the rises associated with the beginning of the Anquetil and Hinault eras was similarly abrupt. So is that the strange thing? Did the ASO throw it to Wiggins, not by giving him an unusually large number of TT kilometres, but just by moving it so much in his favour so fast?

Well, not so fast, because he’s an even laster graph:

Tt52_medium

Yeah, same thing, shorter time period. What I hope this one shows is that this year’s swing toward Wiggins does actually make a lot of sense in context: a context of increasingly dramatic changes. The oscillations are just getting bigger and steeper – that is, designers seem to be moving toward more varied courses, rather than following the same trends for years at a time. Rumours of next year’s course, apparently due to be extremely ‘climby’, reinforce that notion. Then again, this too is a return to the past – the thirties and forties made the modern era look staid and predictable.

I’ve passed the point where I should have found a dramatic ending, but what the hell, I’m keeping plowing on, because you know what, those WEREN’T the last graphs in this post! Oh no! But now I’m moving on to something new! Pie charts! Yes it is now late at night and I’m beginning to fall asleep and may be beginning to act a little oddly and over-enthusiastically! How can you tell!?

Anyway, pie charts. As I was drawing up this post, I was struck by how dominant time trial specialists had been until recently in the Tour. So, I decided to have a peek at that directly – not looking at the course per se, but looking at the winners. How many winners have been specialists? I didn’t bring in external knowledge to this, but simply looked at the TTs within the Tour itself, judging the eventual winner by his performance within the Tour (after all, there’s no point knowing that X was a great time trialist at some point, if he actually came twelfth in the time trial the year he won). I’ve used a five-level ranking: "Terrible" means that the winner finished miles behind in the longest ITT of the Tour – examples include Andy Schleck and Marco Pantani; "Weak" means that the winner came in within the top 25 riders in the longest ITT – like Roger Walkowiak or Carlos Sastre; "Respectable" means that the winner made it into the top eight on the stage – like Bjarne Riis or Lucien van Impe; "Seriously Strong" means that the winner reached the podium of the longest ITT stage (like Gastone Nencini or Bernard Thévenet), and/or won another TT within the same Tour (like Pedro Delgado or Jan Ullrich); "Brilliant" means the overall winner also won the longest ITT.

I could show you the proportions for the whole history of the Tour, but I’m not going to bother, because that’s not the interesting thing. Instead, here’s the way of things from 1947 to 2005:

Tt62_medium

So, winning the Tour is usually all about killing the ITT. From WWII until the retirement of Lance Armstrong, 66% of Tours have been won by the winner of the longest ITT. More than that – 88% have been won by somebody at least ‘Seriously Strong’, which is to say by somebody who was probably a serious threat to win the longest ITT. 92% have been won by somebody who made the top ten in the longest ITT. The Tour is a time trialist’s game.

But wait – aren’t those figures distorted by the handful of riders who won 5 or more Tours? All four of them were also dominant TTists. Well, I’m not sure that’s distortion – it’s not just a coincidence that it was four dominant TTists who were able to dominate the Tour as a whole. But even if we accept that and only look at years not won by Anquetil, Merckx, Hinault or Armstrong, we get this pattern:

Tt72_medium

And why is that interesting? Well, for that we need the next graph, showing events since 2005:

Tt82_medium

Wow! OK, it’s not statistically fair to draw too many conclusions from such a small sample size, and in particular to contrast two eras of very different durations. But I’m going to do it anyway. In the current era, only 43% of Tours have been won by serious TT threats, and fully 28% have been won by people we’d normally think of as bad TTists (a fourfold increase on the 7% that such riders had won up until 2005). Is this a statistical artefact created by the current era’s lack of a defining rider? Is it a reflection a desire on the part of the ASO, either to encourage climbers or else just to create a more varied set of winners? Or is it, perhaps, an encouraging sign that drug use is on the decline, given the traditional association of TTing and PEDs, or alternatively just a shift in effectiveness from TTing drugs to climbing drugs? I don’t know. But it’s certainly a striking change in the pattern.

Finally (finally finally) it’s worth looking at that same change from the other side. Who won the longest ITT in each Tour, and how much of a GC threat were they? Similar ranking to before: "Specialist" means the winner was absolutely nowhere on GC – like Fabian Cancellara; "Wildcard" means the winner didn’t end up a threat on GC, but wasn’t someone who should have been allowed a big break (I’ve taken this as 9th to 20th place on GC) – like Rudi Altig; "Dark Horse" means that they were inside the top eight on GC, like Roger Rivière, Abraham Olano, or Raúl Alcalá; "Contender" means that they were on the podium in Paris, like Lucien van Impe or Levi Leipheimer; and "Champion" means, well, what it sounds like.

Here’s the Olden Days:

Tt92_medium

And here’s the New World Order:

Tt102_medium

It’s obvious what the story is there: the dramatic rise of people with no hope at GC managing to win TTs. We are in the strange new world of the specialist – of, let’s be honest, Fabian Cancellara, who is, to drive the point home, the only man in Tour history to win the longest ITT in two different Tours without also having won the maillot jaune.

Just look at that sentence again. Every year now we’re talking blithely about Cancellara winning the long ITT at the Tour, or whether Tony Martin can beat him… but in the whole history of the Tour, while there have been quite a few men who managed to snatch an ITT when they were on perfect form, the only men who have EVER won the longest ITT in two different years have been men who also won the GC. In fact, multiple longest-ITT-winners form an extremely short list: Bobet, Anquetil, Hinault, Indurain, Ullrich (who, ironically, won the longest ITT twice, but only managed to win the Tour the year he lost the ITT), Armstrong, Cancellara. That’s the same as the number of men who have won multiple Tours since WWII, and the same as the number of men who have won the Giro and the Tour in the same year. I think that maybe we don’t give Cancellara quite the credit that he deserves on that score – and perhaps also we haven’t realise how much the Tour has changed to make this possible. Cancellara is also, it’s worth adding, by a massive margin the lowest-GC rider to ever win the longest ITT, with the exception of Oosterbosch’s DNF – just as Schleck, in the same year, was by some margin the worst TTist ever to win the tour.

So where does all that leave us? Well, on the one hand, yes, Wiggins won a Tour that had more TTing than most recent editions. On the other hand, it was a Tour that was very slim on TTing within the context of the most recent decades. On a different hand, it was a Tour that was pretty average compared to some earlier periods of the Tour’s history. Then again, perhaps the interesting thing is that while the bare facts of the parcours weren’t remarkable, the magnitude of the swing toward TTing was unusual. But not unique. And part of a modern pattern of increasing variability in course design. And in another way, having the same man win the longest ITT and take the overall victory isn’t an anomaly, but a return to the normal nature of the Tour, which has since the war been dominated by time trial specialists; and yet, in the last decade, there has been a trend toward the emergence of true pure time trial specialists who do nothing else, and in that sense Wiggins is an exception to the new norm. One fascinating question is the degree to which this trend is the result of changing behaviour among the riders (whether shady, in terms of different drug regimes, or innocent, in terms of increased specialisation), and to what extent it is the result of decisions by race organisers. While the former can’t be discounted, I think the latter should be taken far more seriously than most fans entertain. To give one example picked almost at random: imagine Tony Martin competing over the 1962 parcours. Martin is a brilliant TTist who can often cope with some pretty aggressive hills as well – in 2009, he finished with the heads of state on the Arcalis MTF, was twelfth (losing two minutes) on the short but incredibly explosive Verbier MTF (where Contador put in arguably the fastest climb in cycling history), only dropping out of the top ten on the fast-raced Saint Bernard passes (he then lost half an hour the next day, but I’m guessing he’d given up at that point). He’s a rider who always threatens not to be dropped, but always is eventually, particularly on some of the massive MTFs.

Well, in 1962, there were only four days in the mountains, and the only MTF was a short ITT up to Superbagnères. By modern standards, there were only two genuine HC mountains in the entire Tour (the Tourmalet and the Bonette), compared to five or six in 2012, and the other little hills would have struggled to make it into the top ten this year. [Although, sidenote: they should bring back the Luitel sometime, it’s one of the most historic mountains of the Tour, despite, or perhaps in part because of, its small number of appearances]. And, on the other side of the equation, there were a whopping 150km of time trials. Is Martin a new type of rider, or are we now just dealing with a new type of parcours? Do you know how many serious hills there were when Cry-Baby Bobet won in 1954? Four – Tourmalet, Aubisque, Izoard and Galibier. No summit finishes. Fabian Cancellara must wish he had been born half a century earlier – he could have won the Tour de France!

Well, there you go. I don't know for sure what the real message of this post is, but hopefully some of you will have seen something interesting along the way. Apologies if there are any errors, and for the rambling nature of the thing (should probably do some serious editing, but I've gestated this so long I can't stand to look at it any longer), and for the terrible image quality of the graphics - they're made in Excel, copied into Word, copied from there into GIMP, saved as .pngs, uploaded to Wordpress, and imported to this site... and it seems that they've lost crispness at every stage along their journey. Would welcome (not unreasonably technical or time-consuming) suggestions as to how to do that better in future.

[P.S. And now one last hurdle - the site has helpfully added in about fifty pages of gibberish interspersed through this post. It doesn't like people actually posting stuff, does it?]

Anyway. Have at!