ANALYSIS: AMSTEL & LIÈGE
What is the hardest race of the year? Some metrics indicate that Amstel Gold and Liège-Bastogne-Liège are the main contenders for this title. Procycling asks the coaches and crunches the numbers to find out
We crunch the data to learn why the hilly spring Classics are the hardest races of the year
The hilly Classics of spring, Amstel Gold Race, Flèche Wallonne and Liège-Bastogne-Liège, have been criticised in recent years for their predictable racing patterns. The relentless climbing and, recent editions of Amstel aside, uphill finishes tend to favour strong climbing teams, who hold things together until the inevitable uphill sprint. There were glimmers of hope for fans of exciting racing in 2018, however. Michael Valgren won Amstel from a late breakway and Bob Jungels won Liège with a long-range solo.
But however the races play out, nobody would ever deny their sheer toughness. The long distances at Amstel and Liège, along with the relentless climbing, mean they fully deserve their status as Classics. According to some coaches, these two races are the hardest days of the entire cycling season. At the annual Science and Cycling conference held at Nantes University in July, Dan Green, formerly head of performance analysis at BMC Racing and now with Dimension Data, made this claim, saying: “You might think it’s Flanders, Roubaix or even Milan-San Remo but I have the data to prove it.”
A LITTLE TLC
The data in question emanates from power meters, but wattage numbers themselves are meaningless without analysis and context. Programmes like Training Peaks are used for this, but Green has devised his own software, called TLC (Total Load of Cycling). “I’ve put thousands of rides through TLC and Amstel came out top at 650 units, with Liège next at 590. Gent-Wevelgem was next, followed by Roubaix, Flanders, Milan-San Remo and, finally, Lombardy.”
These units are similar to Training Peaks’ Training Stress Score (TSS). Both scores quantify the difficulty of a ride based on intensity and duration. For both TLC and TSS, around 100 units is a hard hour’s riding. There are, however, subtle but significant differences that elevated Amstel and Liège to those top two spots. Green says that TLC is more sensitive to the demands of professional racing, where small power surges can make big differences, especially in oneday races. On the other hand, he says, TSS uses a 30-second rolling average.
“This smoothes the data, and discounts a huge amount of data above 700 watts, because 30 seconds is often too long for a rider to hold. At the Classics, especially the Ardennes, riders hit 700-1,000 watts for five seconds, 10 seconds… but not for 30. Some of these guys might be hitting over 700 watts for over six minutes in a race, but it’s simply not accounted for,” Green says. “These short spikes in power relate to what we call the anaerobic, aerobic and endurance components of riding. In power terms, a typical Classics rider records the following: anaerobic power will be over 550 watts; aerobic is from 260w up to 550w; and endurance is below 260w. For a climber, these numbers might be more like: anaerobic power over 500w; aerobic from 220w to 500w; and endurance below 220w.”
Broadly, in Green’s model, endurance is when the rider can almost relax, using fat as the primary energy source. Aerobic is tougher, tapping into a greater percentage of the body’s carbohydrate store – glycogen. Glycogen stores are significantly lower than fat stores. Then there’s the anaerobic side, which is generating energy in the absence of oxygen. This is very hard.
“Classics, in particular, have a much higher anaerobic component than stage races, due to constant pace changes,” says Green. “In Amstel, out of those 650 units, 205 (32 per cent) were from anaerobic means. There were a further 355 (54 per cent) from aerobic and just 87 (14 per cent) from the easier endurance zone.”
Compare that to the BMC riders’ annual total. On average, each of the riders rode for 876 hours, over 27,500km, of which 326km were ascents, recording a total TLC figure of 50,000 units. Of those, only 14 per cent came from anaerobic sources.
Green’s conclusions are supported by former HTC-Highroad coach Sebastian Weber. Weber has created INSCYD, a piece of software that looks at the metabolic shift of cycling to measure intensity. Weber put the course profiles for Milan-San Remo, Paris-Roubaix, Amstel and Liège through the software to calculate metabolically which was the most demanding.
“For Milan and Roubaix, I used a 78-80kg sprinter and rouleur, whereas for Amstel and Liège I profiled lighter (70-72kg) climbers,” Weber explains. “This is realistic for those races and affects the absolute power output. But ultimately there were two clear conclusions: ParisRoubaix was harder when it came to the creatine energy system, while Liège produced higher levels of lactate.”
The creatine energy system is anaerobic up to around 10 seconds. Weber says that the Roubaix pavé require numerous subfour-second efforts for accelerating and correcting, whereas Liège demands slightly longer anaerobic and hard aerobic efforts, which use precious glycogen.
That’s the physiological background. But how does this fit into the course profile? Carlos Verona is a 26-year-old Spanish rider who was 84th in Amstel and 80th in Liège for Mitchelton-Scott this year. He
says that distance plays a part, both physically and technically. At 263km, Amstel is 31km shorter than Milan-San Remo, but Verona says that the technical aspect is significantly more draining.
“At Amstel, it’s all about position,” Verona says. “The lead riders hit the hills at such speed that if you’re not in the top 30, you have to break from your group in an effort to catch them, and that’s a massive effort. It’s made harder because there’s always a massive fight for position to start the climbs well positioned.
“Liège is different,” he continues. “Position is important but there’s an endurance aspect to it. With 100km to go, it’s full gas with huge shorter efforts.”
Verona uploads his training and race rides to Strava. For Amstel, Verona rode 264.11km over 6:44 hours, covering 3,299m of climbing for what Strava call a historic relative effort score of 464 - which is broadly based on heart rate. At Liège a week later, Verona covered 263.36km in just under 6:48 hours, ticking off 3,951m of ascent for a historic relative effort score of 453. On paper, Amstel just nicked it.
“That’s about right,” says Verona. “Amstel’s certainly mentally harder than Liège because you can’t relax. You’ve always got to be in position for the climbs, but also you need to be careful with all the road furniture.”
VALGREN DIGS DEEP
Verona’s a domestique at the Classics, but what about the load for a contender for the victory? Amstel was Michael Valgren’s biggest ever win, and he uploaded his power files to Strava, allowing Stephen Gallagher, who now runs Dig Deep Coaching, to analyse it. Gallagher’s review highlights the power fluctuations: Amstel riders have to dig deep into the anaerobic energy system.
“Valgren won Amstel in 6:40:07, riding at an average speed of 39.4kph,” says Gallagher. “In the first 15km, he averaged 271w – a big effort even for a pro.” With 100km to go, the pace increased further, starting with the Vrakelberg 4km later. “That’s where Valgren produced 520w for the 1:23 needed to cover that short and steep climb. Then on the Keutenberg, with its steep section of over 20 per cent, the riders in the main group attacked each other and chased hard to catch the break,” he continues. “Ahead, Lawson Craddock lost 32 seconds on the 1.1km-long climb alone, climbing it in 3:12 at 436w. In the meantime, Valgren put out 508w to climb it with Valverde and co in just 2:43.”
Valgren registered “big numbers” on the Cauberg (at 244km) as the thinned-out group of favourites went over it for the last time. “Here, Valgren generated 595w for the 1:26 it took to cover the steepest sector. This was over 100w more than the previous two ascents of the climb and it brought a select group, including Valgren, Peter Sagan, Alejandro Valverde and [Roman]
Kreuziger, up to the escapees. The second acceleration was a big punch and this 47-second effort was enough to create a gap between him and the others, with only Kreuziger following.”
Valgren positioned himself perfectly for the sprint, sitting behind Kreuziger with 200m to go before jumping past. “That’s when he generated an average power of 925w for nine seconds, maxing at 1,050w,” says Gallagher. “It was a huge last effort.”
In this case, Valgren’s efforts eclipsed that of the domestique. But often, explains Green, in races the opposite occurs. “In many cases it’s the workers who are doing the lion’s share, sheltering their leader from the wind and covering attacks,” he says. “I’m convinced [Micha¯] Kwiatkowski will churn out greater load figures on many stages of the Tour compared to [Chris] Froome and [Geraint] Thomas. He does so much work, burying himself until dropping around 4km from the finish.”
Green elaborates on this role-specific effort. In 2017, for example, Greg Van Avermaet’s load at Amstel (where he finished 12th) was six per cent below the team average, whereas when he won Roubaix, he was about 20 per cent higher. “The data we have for other riders in the team shows that they sat up 50km from the finish or stopped,” Green says. “Whereas, of course, Greg won.”
A high load isn’t necessarily a bad thing. In fact, it can be advantageous. This year, Michael Woods of EF-Drapac came 20th at Amstel and second at Liège. Woods says that Liège, in particular, plays to his strengths. “One of my fortes is anaerobic repeatability; I’m good at riding three to eight-minute efforts,” Woods says. “Liège is like that, where the longest climb is less than 10 minutes.”
Woods’s physiology suits these anaerobic surges because of his physical makeup, but also because of his past athletic life and current training regime. “I used to run 1,500m and 3,000m, which rounded out to around four- and eightminute, high-intensity efforts. My coach and I do a lot of work in those zones because we realise that’s my strength,” the Canadian says.
“In the WorldTour, once you find what you’re good at, you need to focus on that as it’s becoming increasingly specialised. In training for a race like Liège, we’ll do a long ride where, every hour, we’ll do specific efforts that simulate key climbs like La Redoute and Côte de Saint-Nicolas, by power.”
COBBLES OR HILLS?
How do cobbles change the equation? Despite Green’s conclusions, many think the cobbled races are harder than the hilly Ardennes. “By my calculations, in terms of training load – and I use TSS – the Tour of Flanders was the hardest,” says Quick-Step Floors trainer, Koen Pelgrim. San Remo came in fifth with a score of 331, Roubaix fourth 353, Amstel third 359, Liège second 379 and Flanders was a whopping 78 ahead with 457. “There are so few easy parts in Flanders, with riders always fighting for position, whether over climbs or cobbles.”
Flanders was a huge collective effort for the Belgian team, with Niki Terpstra winning, Philippe Gilbert third and Ždenek Štybar 10th.
This discrepancy could be down to, as Green would say, data variations coming out of TSS compared to TLC, but Green concedes no system yet factors in the brutality of the cobbles. “The extra energy to ride the cobbles means the physiological load of the power output is higher than what you normally see from that output,” he says. “They can ride 380w at 80 per cent of maximum oxygen uptake and heart rate on the flat, but 380w on cobbles is probably 84 per cent of maximum.”
Green says you’d need sensors and algorithms to examine the physiological load of riding on the cobbles. But his figures remain significant, because while the system doesn’t – cannot – account for the psychological and muscular impact of the cobbles, it does measure the increased heart rate and strain.
The riders, however, have their own take. “If you were to ask me to ride a course solo tomorrow, Liège would be the hardest,” says Dan Martin, the 2013 winner. “But as a race, maybe because I enjoy it, it flies by. Amstel is like a video game with the constant up, down, left and right. 2018 was the hardest Amstel I have done as the break was given a huge gap which meant the pace was very high for the final 200km. But it’s down to what type of rider you are. For me, Roubaix would surely take the highest physical toll, but that’d be the same for a cobbles rider taking on the Ardennes.”
Astana's Valgren celebrates after winning the sprint to the line in Amstel
Jungels attacks on the Côte de la Roche aux Faucons on his way to winning Liège
Liège's climbs play to Woods's strengths of riding hard for three to eight minutes