CapoVelo.com - | How to Predict Individual and Team Performance
91054
post-template-default,single,single-post,postid-91054,single-format-standard,no_animation

How to Predict Individual and Team Performance

article by Spencer Martin and Steve Maxwell

Cycling is different from almost all other sports. While it is undeniably a team sport, individual riders get most of the attention. And, unlike most professional sports, there aren’t really many analytical techniques to rate individual performers or teams as a whole. One can rate the number of victories, podiums, or top ten finishes for individual riders or teams, but the wealth of individual and team statistics that characterize sports like football, baseball, or basketball are almost completely lacking in bike racing. The UCI and different commercial groups – like ProCyclingStats – have designed intricate point systems for ranking performance, on the basis of specific achievements. The Outer Line has utilized these points systems to analyze and evaluate team and individual performance in various detailed ways in the past. Nonetheless, all such metrics are inherently subjective and will inevitably fail to capture the entire dynamics of bike racing.

All of this renders objective big-picture retrospective analysis of pro bike races very difficult. It also makes predicting future performance more difficult. However, pro cycling analysts have to make do with the information that is available. And in an effort to provide such an objective anchor, the Beyond the Peloton newsletter has created a new rating system – dubbed the BTP NET rating. This system attempts to project how individual teams have improved (or weakened) their relative team strength from year to year, based upon incoming or outgoing riders, as well as their prospects for 2021.

To create this relative rating system, BTP utilized the ProCyclingStats database, which is generally regarded to be more detailed and precise than the parallel UCI points ranking system. The NET rating takes each individual team’s final year-end points total from the 2020 season and then adds or subtracts the number of 2020 points contributed by new incoming riders, or lost via outgoing riders. In other words, the ranking shows how the teams would rank if every rider currently on the roster were on the team in the year prior.

While somewhat simplified, the NET ranking provides two basic semi-quantitative metrics and insights:

  • clear illustration of which teams did better and which did worse during the off-season transfer process, and
  • rough illustration of which teams may be the most successful in 2021.

Table 1 provides a clear illustration of the first metric – which teams gained or lost ground for 2021. The net gain/loss column simply tallies the 2020 points of both the incoming and outgoing riders. This metric is typically tied very closely to the movement of a small number of top riders. Ineos showed the biggest net gain, by picking up just three riders with a significant number of PCS points – Richie Porte, Adam Yates, and Dani Martínez. Movistar moved up, largely on the acquisition of Miguel Ángel López, as did Team UAE with the surprise early-2021 signing of Marc Hirschi. Meanwhile, Israel Start-Up Nation also jumped up, as we have previously detailed, with most of the points coming from Mike Woods and Daryl Impey. (Chris Froome, though generating by far the most media publicity, brought in only 33 points – the result of a weak 2020, following his horrific accident in 2019.

On the other end of the spectrum, three teams stand out as having lost a considerable amount of talent. Team DSM (formerly Sunweb) losses are largely attributed to the loss of Michael Matthews, Wilco Kelderman, and as mentioned above, Marc Hirschi. No incoming rider, including Romain Bardet, brought even half the points that each of those riders represented. EF Education-Nippo’s loss of Mike Woods and Dani Martínez accounted for the bulk of their net loss; likewise, Richie Porte’s departure accounted for virtually 100 percent of Trek’s net loss.

These semi-quantitative data points underline and support several “gut feel” types of assumptions or predictions about the competitive landscape as we head into the 2021 racing season. For example, Movistar performed pretty poorly in 2020, but based upon the talent added during the off-season, they can be predicted to do better in 2021. On the flip side, for example, Team EF Education-Nippo had a pretty successful 2020 – winning stages in all three grand tours – but lost a number of key riders in the offseason, and hence may be predicted to struggle more in 2021. Even if high-performing riders such as Sergio Higuita and Hugh Carthy repeat their success from 2020, it may not be enough to lift the entire team to the same level of success, since those performances are already baked into the team’s BTP points totals.

There will, of course, always be surprises – unexpected breakthroughs or collapses – that history-based systems like this can never predict. Additionally, the natural decline of aging riders and the rise of youngsters go somewhat unaccounted for in this exercise. For example, who expected Marc Hirschi to achieve the level of stardom and points that he did in 2020? (Indeed, as we’ve just demonstrated, his movement at the beginning of this season, completely reoriented the competitive strategy and performance ranking of both his new and his old team.) And who could have predicted the virtual collapse of the 2018-2019 consensus “golden boy” Egan Bernal, who stumbled and then fell completely out of the shortened 2020 season?

The NET rating also gives at least a rough indication of which may be the best-performing teams in the upcoming season, as shown graphically in Table 2. Although there will undoubtedly be surprises here as well, BTP uses this adjusted NET ranking to group the 19 WorldTour teams into four groups for 2021, as broken down below:

Tier 1 Teams – “The Galacticos:”
1)  Ineos Grenadiers – 7742 points
2)  Deceuninck-Quick-Step – 7483
3)  UAE-Team Emirates – 7457
4)  Team Jumbo-Visma – 6170

Tier 2 Teams – “The Challengers:”
5)  Bora-Hansgrohe – 4802 points
6)  Astana-Premier Tech – 4480
7)  Groupama-FDJ – 4299
8) Movistar Team – 4053
9) Bahrain-Victorious – 4008

Tier 3 Teams – “The Dreamers”
10)  Team BikeExchange – 3944 points
11)  AG2R Citroën – 3822
12) Team DSM (Sunweb) – 3459
13)  Trek-Segafredo – 3483
14)  Israel Start-Up Nation – 3281

Tier 4 Teams – “The Basement Dwellers:”
15)  Cofidis – 3247 points
16)  Lotto Soudal – 2918
17)  EF Education-Nippo – 2811
18)  Intermarché Wanty Gobert – 2248
19)  Team Qhubeka-Assos – 2035

In terms of these PCS points, the top four teams are clearly head and shoulders above the rest of the peloton. The spread or distinction between the teams in the lower three groups is a little less clear-cut, and some qualitative assumptions necessarily have to be made to determine the point levels at which to segment these groups.

Currently, it seems that many teams select riders to either fill a specific roster need or based on their longer-term historical accomplishments, as opposed to attempting to track their recent past production as a guide to their likely near-term future performance. Adjusting this rider recruitment strategy could lead to significantly increased team success without any increase in financial commitment.

Indeed, the argument can be made that team managers who do have more a numbers-based approach can capitalize on the inefficiencies which are created by the general number-blindness prevalent in the sport. Even something as simple as just compiling the PCS data for your own roster, and following the simple logic that: (1) riders who have performed well in the recent past will tend to perform well in the near future; (2) younger riders tend to get better, and (3) older riders tend to decline, will take you a long way towards building a successful pro cycling team. For example, both Jumbo-Visma and Bora-Hansgrohe have built top-flight teams based on this logic, even though it is generally believed that they don’t have large budgets.

Cycling will always be a far cry from the sabermetric world of baseball, where quantitative and numbers-oriented teams – first popularized by Michael Lewis’s book Moneyball – are increasingly able to create marginal gains and minuscule relative advantages vis-à-vis their competitors, which can translate into victory. Cycling doesn’t lend itself to anywhere near that sort of quantitative analysis. But, this isn’t all negative. The nuances of cycling that make it frustratingly difficult to quantify are also the qualities that make it a truly magical sport.

 

 

The Outer Line 

Leave a reply
Share on