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

Procycling - - CONTENTS - Writer: James Witts Im­age: Gru­ber Im­ages*

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.”


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.”


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.”


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

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.