As a data science nerd, I've been closely following the professional cycling's promotion/relegation scheme since it began in 2020, tracking the UCI Team Rankings that will determine which 18 teams might get a WorldTour licence. I've picked apart each team's results, rosters and strategy while also taking a wider view on how the relatively new scheme has changed the sport.
For the purposes of this article, the Women's WorldTour - equally as interesting in terms of racing - will have to take a back seat. Between teams going away and fewer applications for WorldTour licences, there hasn't been the dramatic battle to avoid relegation that we've see on the men's side.
Origins of cycling's relegation system
It started with a fierce battle between the UCI and the most powerful race organisers, the Amaury Sport Organisation (ASO), who run the Tour de France, Paris-Roubaix, Paris-Nice and many other top events.
The UCI, in an attempt to create a cohesive calendar where the best teams and riders would regularly face each other, as well as more stability for teams, created the WorldTour (or ProTour as it was originally called). The concept was fairly simple: a set number of teams would be required to compete in a set number of top races so fans could enjoy the races with great competition.
From the inception, ASO objected to being forced to invite teams that they had no hand in choosing, citing a list of objections and demands over the years and complaining of a 'closed system'. Doping scandals like Operación Puerto in 2006 and the 2007 Tour de France doping cases involving stage 13 winner Alexandre Vinokourov and race leader Michael Rasmussen only bolstered ASO's case for stricter controls on who gets into the top tier.
To cut out years of history far outside the scope of this article, the ASO vs. UCI conflict eventually led to the ethical, organisational and financial criteria that we see for WorldTour licences today. But the last demand for sporting criteria took much longer.
The reason was simple: pro cycling had a pesky doping problem, and adding pressure for results would only push riders toward illegal performance enhancement rather than away from it. Once the Athlete Biological Passport was in place and showing an effect on the doping problem, it paved the way for the UCI to add sporting criteria to WorldTour licences in 2020.
Is it a coincidence that a French UCI president, David Lappartient, was the one to make it happen? Is it a coincidence that it came on the heels of Israel Cycling Academy (as it was then known), owned by billionaire Sylvan Adams, joining the WorldTour by purchasing Katusha's WorldTour licence in 2019? Possibly not.
How the points system improved from 2022
During the 2020-2022 promotion/relegation period, the points structure came under heavy criticism for the weighting of one-day races compared to stage wins in the Tour de France. Third place at the ProSeries Faun-Ardèche Classic was worth 125 points while a Tour de France stage win was worth 120 points.
Cofidis and Arkéa-Samsic took advantage of this and focused much of their seasons on lower-level races in the ProSeries or Class 1 (.1) events, because of the points imbalance.
Additionally, the system only factored in the results of the top 10 riders of each team counted toward their team rankings, and Grand Tour stages only went five-deep for points, Arkéa-Samsic stacking the top 20 of the Tro-Bro Léon made a lot of sense for the team's future, but perhaps did not make for thrilling racing. As a result, Arkéa were promoted with only 3,842 of their points coming from the WorldTour compared with Lotto's 6,518 WorldTour points.
The UCI re-worked the points structure to give fewer points for one-day races relative to the more heavily weighted Tour de France before the start of the 2023-2025 cycle. WorldTour races victories now range from 1,300 for the Tour de France, 1,100 for the Giro d'Italia and Vuelta a España, 800 for any of the Monuments, down to just 300, based on historical importance.
You can see the relative differences in the available points in the chart below.
Since 2023, there are more points for WorldTour races compared to other series events, and the points are 15-deep for stages of Grand Tours and 60-deep for the GC or one-day races, which gives riders something to fight for.
The team rankings now include the top 20 riders rather than the top 10 of each team, giving a benefit to teams with deeper rosters.
Is it more even? Yes - the Grand Tours have a much bigger piece of the overall pie, while one-day races, which accounted for almost half of the total available points in 2020-2022 are now about a third.
Has the system rewarded the most qualified teams? In the first cycle, Lotto and Israel outperformed Arkéa in WorldTour races but were relegated. Since the revamp, Uno-X Mobility are likely to be promoted once the Lotto-Intermarché merger is final, but they scored the fewest points of any of the top teams in WorldTour races, still scoring most of their points in the ProSeries.
XDS Astana's miraculous turnaround
The stress of the relegation battle has been palpable both in 2022 and this year. The tension ramped up in the last months of 2025 when the relegation came down to just a few hundred points over three years of racing, with Cofidis most notably ending up on the losing side of the battle.
One team that had been sweating relegation for two years and should have been worried in 2025 were XDS Astana, but their miraculous season didn't just have them sneak by to stay in the WorldTour.
With a focus on improving everything from recruiting, to calendars, and even race tactics, they exceeded all expectations. But the only way the team could have done that was with a bigger budget.
XDS Astana were the prime example of how a cash infusion can turn a team's fortunes. They were well on their way to relegation after the 2023 and 2024 seasons, and came into this year dead last of the WorldTour teams. However, they secured the Chinese bike maker XDS as a new sponsor, and with the infusion of cash, were able to bring in a raft of talent to help execute their plan to remain in the WorldTour. Unfortunately, pro cycling is secretive about budgets, so it's difficult to accurately analyse results based on it.
However, Astana's new sponsor helped the team bring on a data scientist who helped the managers and directors come up with a calendar and select just the right kinds of riders, along with a stronger team of coaches and staff to support the riders throughout the season and ensure they were in fighting form.
It also supported a massive increase in the number of races in which the team were able to score points - they went from scoring in 74 and 88 races in 2023 and 2024, respectively, to 124 races this year. Their average points taken in those races went up from 96 and 76 to 135 - a massive increase that fuelled their ascent.
The changes were astonishing: XDS Astana ended the year in fourth for the 2025 season behind powerhouses UAE Team Emirates XRG, Visma-Lease a Bike, and Lidl-Trek.
Their tally of points for the season - 16,712 - was more than enough to lift them out of the relegation zone, climbing from 21st to 15th for the three-year period. I've highlighted the team in blue in this animated bar chart race so you can see their progress in the three-year rankings over time. (Note, the chart only looks at final rankings for 2023 and 2024 to keep the animation reasonably short.)
In the next interactive chart, I calculated the number of unique races each team scored in over the last three seasons, and the average points per unique race. This is where Astana's diverse calendar really popped out.
Play around with the filter - you can click on a team's name to show only their data. You can zero in on a single year with the drop-down filter.
You can see how XDS Astana banged away at a bunch of races, not scoring a ton of points but it added up to more than enough.
Across all three years, it's astonishing how UAE Team Emirates excelled in both measures. They're the big teal blobs in the upper right, scoring a lot of points in a lot of races.
Visma-Lease a Bike focused their efforts more, scoring in 60-70 races each season but averaged more per race than UAE in most cases.
In the lower right quadrant, you can see that Cofidis (pink, hover over for the year and more info) scored in more races than most teams and still were relegated because the points they gained were so low per race, on average.
I also wanted to see what UCI category of races each team focused on in terms of collecting points. While it's obvious that the WorldTour races have the most points, and each team's percentage of points should favour those events, Uno-X Mobility's largest category was the ProSeries at 36% of their points over three years.
It's interesting that Visma-Lease a Bike and Bahrain Victorious almost exclusively earn points from the WorldTour. Arkéa-B&B Hotels and Cofidis, the two relegated teams, had the highest percentage of Class 1 races in their totals.
Next, I wanted to see how important the transfer market was for teams in the promotion/relegation battle. It was a tedious process, but I came away with the chart below, which shows what percentage of a team's points were scored by riders new to the team (incoming transfers) and the returning riders. I put the bars in Cyclingnews colors just to get your attention.
For Astana, it was surprising that, after their massive recruitment for 2025, it was largely the returning riders who scored this season. Christian Scaroni and Simone Velasco have been with them since 2022, and Lorenzo Fortunato, who came in 2024, were the top three for the team, and 60% of their points for this year came from returning riders. The opposite was the case in 2024, when the incoming riders buoyed the team, but the overall total was much lower across the board.
Still, across three years, Astana's incoming riders were responsible for the largest percentage of points of any of the other 21 teams that I examined at 34.78%.
In contrast, Cofidis struggled in the transfer market, and overall, only 18.68% of their points came from new riders. They lack the budget to bring on an existing powerhouse rider, and instead, tried to develop younger riders from within the team. After bringing on a host of riders for 2023 like Ion Izagirre, Bryan Coquard and Benjamin Thomas, they also lost solid riders over the three years like Victor Lafay, Max Walscheid, and Guillaume Martin. Of their investments for 2025 like Dylan Teuns, Emanuel Buchmann and Alex Aranburu, only the Spaniard paid off as expected.
The team blamed manager Cedric Vasseur for focusing too much on scoring points, which seems counterintuitive. But the opposite strategy - keeping victories as the main focus and letting the points follow - is what helped bring Lotto back into the WorldTour, at least until their dismal 2025 season. Because so much of cycling is psychological, it makes sense that teams succeed by celebrating positive outcomes and motivating riders to try for good results rather than just scoring whatever points they can.
Quality over quantity
Speaking of wins, I thought it would be interesting to see what each team's average points were per win, and I was surprised at the outcome of this, too.
It shows that teams don't necessarily have to win a lot to stay in the WorldTour; they can make an impact by being good at really valuable races.
Alpecin-Deceuninck are the shining example: they don't feature high on the wins per month column, when they do win, they make it count more than any other team.
This is mainly due to the fact that Mathieu van der Poel wins a lot of their races, and in major Classics. Paris-Roubaix and Milan-San Remo are worth 800 points to the winner. Additionally, Jasper Philipsen had a bunch of Grand Tour stage wins worth 180 points or 210 for the Tour. Their average win was worth just shy of 200 points, more than even UAE Team Emirates!
This is probably the only chart where UAE, Visma and Lidl-Trek aren't at the top.
Compare that to the heat map below showing the number of wins per month by each team across three years.
This heat map clearly demonstrates just how dominant UAE Team Emirates-XRG came to be over the three years. Grouped by month, the darker colors are fewer victories, and the lighter the color, the more wins. This data only looks at one-day and GC victories and ignores other points coming from leading or winning a minor classification in a stage race.
UAE Team Emirates XRG recorded the most wins in a month in June 2025, in part because five riders won national titles. But the team won three stages and the GC in the Dauphiné with Tadej Pogačar and the Tour de Suisse with João Almeida. Filippo Baroncini's win in the Baloise Belgium Tour and Ivo Oliveira's stage in Tour of Slovenia put them over the previous maximum of 13 held by the same team in July 2024 and July 2025 and by Visma-Lease a Bike in September 2023.
Is winning 'momentum' real?
When I saw that Paul Magnier kicked off a massive winning streak for Soudal-Quickstep starting with the GP de Fourmies on September 14 in 2025, I wondered if wins usually come in streaks after periods of drought.
It was hard to call what Soudal-Quickstep did this fall a winning streak, just because they win often enough to make a series of victories unusual.
I decided to go looking for teams that had an unusually long period without a win, and then went on to win as much or more than they normally do.
What constitutes a drought depends on how often a team usually wins. A streak of five wins in nine days was big for Bahrain Victorious in June of 2024 because they were only getting one or two per month before then.
So I asked Google's Gemini to crunch the data because it was beyond my capabilities. After much internal debate and experimentation, I asked it to define a drought as a winless period of one standard deviation longer than the team's average period of being winless (not including those spanning the off-season), and a streak as a period of an average or greater than average number of wins.
It found 23 instances over the three years of this WorldTour relegation cycle that met this criteria, but only a handful were statistically significant. The least random was XDS Astana's nine wins after a period of 51 straight days without one, from July's Tour of Qinghai to Christian Scaroni's Giro della Romagna victory. After that, they won four stages of Tour de Langkawi, two and the GC in Tour of Taihu Lake, and a stage of the Tour de Kyushu.
Visma-Lease a Bike cracked a 38-day drought from April to May in 2023 with a series of 18 victories in May and June, and Uno-X Mobility spent a torturous 78 days without a success at the start of 2024 until they turned it around with 14 wins in three months. Here's where the formula falls down, as those were too scattered to qualify as a streak in most people's minds.
I learned more than I really wanted to about statistics and what are called 'self-exciting processes' through this exercise. More experienced data scientists, feel free to drop suggestions in the comments on what parameters to use.
But I'm pretty sure most periods of drought are not broken by streaks of victories even though it often feels like it.
What makes a good WorldTour team?
What has been on my mind for a lot of the last six years is what makes a good team, other than having superstar riders? Most of the teams in the WorldTour are quite evenly matched below the top five.
It's likely a mix of factors: budget, management, recruitment and results but in the end, pro cycling is a sport that involves human beings. If a team is relegated, it's not the fault of any one rider or sports director, but a team-wide strategy failure, just as a promotion is a full-press team effort. They're good dogs!
I tried to visualise each team's qualities with a radar chart, and while it shows some interesting things like Ineos' high number of unique winners and Alpecin-Deceuninck's deep reliance on a couple riders, what is most stunning is UAE Team Emirates are amazing in every category. The two relegated teams? They barely register with the scale needed to accommodate UAE!
The possibilities of analysing the WorldTour data are endless so I'll leave it here. Drop a comment if you want to see anything else.
The full dataset used in this analysis is available here.
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.