Friday Feature: Lies, damned lies and statistics

Being of an analytical nature – my best friend at school was an Excel spreadsheet – I’ve been pondering about how best to quantifiably rank sprinters. The result of my various investigations? There is no one definitive ranking. Every method I tried led me to a ranking order which fell apart under even the slightest scrutiny.

It reminded me of the famous expression commonly attributed to former British Prime Minister Benjamin Disraeli:

There are three kinds of lies: lies, damned lies and statistics.

He was a smart fellow, old Ben.

Frustrating? Yes. But it was also an exercise worth going through. Let me share the fruits of my labour with you, although readers who suffer from arithmophobia – the fear of numbers – might want to look away now.

The ‘win’ method

The most obvious starting point is to look simply at the number of wins earned by individual sprinters. Off the top of my head, here was my selection and their win record in 2011 (this total includes only individual stages and excludes jerseys or team time trials):

Sprinters' league table, ranked by 2011 wins (selected riders only)

On this basis, first-year pro Marcel Kittel should be regarded as the best sprinter in cycling, ahead of even Mark Cavendish and Andre Greipel. Clearly this isn’t quite right.

Based solely on quantity of wins, Marcel Kittel would have been top of the sprinters' rankings in 2011 (image courtesy of 1t4i)

Although blessed with fantastic acceleration, the young German won just one stage at the very highest level, at the Vuelta. (Admittedly it didn’t help that his Skil-Shimano team weren’t invited to either the Giro or the Tour.) His four wins at the Tour de Pologne came against a relatively weak field, and most of his other victories came in minor races.

On the other hand Cavendish won seven times combined at the Giro and the Tour, claiming the points classification at the latter. There was also the small matter of his World Championship road race victory. There is truth in the numbers – Kittel is blisteringly quick – but better than Cav? No, not yet.

Similarly, I doubt any but the most blinkered of tifosi would suggest that Farnese Vini’s Andrea Guardini is on the same level as Greipel, despite sharing the same number of wins (11). Greipel beat Cavendish in a straight fight to win his first Tour stage last year, and has defeated all the top sprinters. By comparison Guardini achieved five of his wins at the Tour de Langkawi (which is not renowned for attracting big-name sprinters), with his only other notable win coming in the early-season Tour of Qatar.

Arguably the one rider who could claim to be closest to Cavendish was Peter Sagan, who recorded 13 wins. More of a power sprinter than a pure speedster, in the mould of a Thor Hushovd or an Edvald Boasson Hagen, Sagan won more stages (three) at the Vuelta than anyone else, and picked up most of his other wins in good quality races such as the Tour de Suisse. There’s certainly a strong argument for ranking him above Kittel, at the very least.

The WorldTour points method

Clearly the ‘win’ method is flawed because the true ranking of a sprinter should not be judged solely on the quantity of wins as the quality of the race and the opposition. This is where the WorldTour system – which awards points relative to how big the race is – should provide a fairer reflection. It does, but even then it is still flawed.

A quick examination of the 2011 WorldTour points table reveals that the rankings are rightly slanted towards those who are able to dominate the biggest multi-stage and one-day races: Classics master Philippe Gilbert topped the rankings, followed by Tour de France winner Cadel Evans and the ever-consistent climber Joaquim Rodriguez.

Not until we reach 11th position do we come across the first rider who could be even remotely considered a sprinter. But the leading fast-man isn’t Kittel, Cavendish, Sagan or Greipel – it’s Sky’s all-rounder Edvald Boasson Hagen. That’s the same Boasson Hagen who appears at the bottom of my previous table when ranked in terms of individual wins with seven – less than half the total of either Kittel or Cavendish.

WorldTour sprinter rankings (selected riders only)

Boasson Hagen's overall ability catapulted him to 11th in the rankings (image courtesy of Wikipedia)

Why is this? It’s all to do with the way WorldTour points are allocated. Firstly, not every race on the 2011 calendar counted towards the rankings – indeed, only 27 did, split between stage races (14) and one-day Classics (13). This explains why both Kittel and Guardini feature so low down, as the majority of their wins occurred in non-ranking events – indeed, the latter did not register a single WorldTour point among his 11 wins.

Secondly, as mentioned above, points are weighted heavily in favour of overall race winners rather than those whose stock-in-trade is stage victories. So, for instance, Mark Cavendish’s five wins at the Tour de France netted him the same 100 points gained by Boasson Hagen for winning the relatively modest Eneco Tour.

Indeed, Boasson Hagen only gained points at three events all year, but he made them count: 112 at the Eneco Tour (including a first, second and third), 68 at the Tour de France (where he won two stages and had four other top-five placings), and a win at the Vattenfall Classics to net a further 80.

Farrar didn't have many big results, but they all counted (image courtesy of Wikipedia)

Similarly Tyler Farrar was another who made his performances count. The perennial nearly man in the sprints – wherever you see the likes of Cavendish or Greipel raising their arms aloft in victory, Farrar is often to be seen coming in further back in the top five – he earned two of his five wins and all but four of his 108 points at three WorldTour events. A win and two seconds at Tirreno-Adriatico netted 14 points, while the bottom step of the podium at Gent-Wevelgem delivered a further 50. And at the Tour de France he finished in each of the top four places once: another 40 points. He will never be as prolific a winner as even Kittel or Guardini, but he wins (or at least places well) in the races that matter.

In many cases it’s possible for one big victory (or the lack thereof) to have a disproportionate effect on a rider’s ranking, which is why Andre Greipel languishes behind Matt Goss, Peter Sagan and even Tom Boonen. Goss earned nearly half his total by winning Milan-San Remo (100 points), Sagan gained more points by winning the Tour de Pologne (100) than he did from his three Vuelta stage wins (48), while one of Boonen’s paltry total of two 2011 wins came at Gent-Wevelgem (80 points).

Rojas earned a few points here, a few points there (image courtesy of Movistar)

Equally it’s possible to climb up the rankings by stealth without winning many races at all. Movistar’s Jose Joaquin Rojas had only three victories in 2011, and only one in a race of note: a stage at the Volta a Catalunya. However he also added two seconds, two thirds and a fourth in Catalunya to yield 19 points. By my count he also added 26 points at the Tour de France despite never finishing higher than third on any stage. Indeed, he picked up points consistently throughout the year without ever landing any big totals – enough to put him a more than respectable 54th in the overall rankings, 45 places above Kittel.

By comparison Cavendish, whose focus is on stage wins rather than overall victories, is disadvantaged by the fact that he can only accumulate his points in piecemeal fashion (and had a modest Classics season). So although his record of seven Grand Tour stages in 2011 was unparalleled, they ‘only’ earned him 132 points (5 x 20 at the Tour, 2 x 16 at the Giro). As for his World Championship win in Copenhagen, that netted him the grand total of zero points. [Go figure – Ed.]

Meanwhile Fabian Cancellara gained 230 points with two second and one third-place finish in the spring Classics. That’s not to downplay Cancellara’s remarkable consistency in any way, but it does highlight that a sprinter’s lot is not a happy one – at least in terms of being a rain-maker for WorldTour points.

Basically, when it comes to ranking sprinters, you can trust the WorldTour points system about as far as you can throw it.

So, what’s the score?

The WorldTour system works well for assessing the best riders overall. But although it’s far better than counting race wins alone, I do think it is weighted heavily against the sprinters and still doesn’t work when it comes to ranking them.

However you look at the stats, there's really no dispute who the number 1 sprinter in 2011 was (image courtesy of Wikipedia)

The heavily graded sliding scale used for allocating points for sprint stages rewards wins over consistency – a good thing. For instance, a Tour de France stage win is worth 20 points, followed by ten, six, four and two points down to fifth, which seems about right. But I can’t help but feel their importance is understated relative to minor stage races or even the one-day Classics. Is it really right that a Tour stage win is worth the same as eighth place at Milan-San Remo? Or that five wins in France is equivalent to fifth place overall, or winning the Tour of Beijing? I’m not sure the balance is quite right there.

And then there are the races that are missing. Kuurne-Brussels-Kuurne is not as important as the Tour of Flanders, but is it right that victory in the former is worthless in terms of WorldTour points? And why are the Eneco Tour or the Tour of Beijing included, but the Tours of California or Langkawi not?

Having said that, I am probably safe in saying that Mark Cavendish cares not one jot about the fact he was only 30th in last year’s WorldTour league table. I doubt many would dispute Cav’s claim as the outstanding sprinter of 2011 – all he has to do is point to his rainbow and green jerseys and his bulging palmares. When you see him beating all his major rivals time and again at the Tour de France, you can throw point counts of the window.

However, there is another more serious point here. WorldTour points are an important form of currency further down the food chain. In a sport where the stipulated minimum annual salary for veteran riders is a Scrooge-like €30,ooo (£25,000) – and a significant proportion earn less than €100,000 – a few points here and there could make a massive difference to a rider’s earning potential. It’s tough at the bottom, and doubly so if you’re a sprinter, it would seem.

Like the man Disraeli said: lies, damned lies and statistics.

3 thoughts on “Friday Feature: Lies, damned lies and statistics

  1. Pingback: Early season Italian races review « VeloVoices

  2. Joe says:

    Good read, I love stats! Would be interesting to see them ranked using the UCI “sporting value” points system, which does include points for the lower ranked events and world championships.

  3. Thanks Joe. As you can see, I’m very much a stats guy too. I’ve not investigated the sporting value system before – maybe that’s a topic for another post …

Leave a Reply