Articles

The first grand tour of 2023 is about to start and our servers are buzzing from visitors trying to find out who has the most chances of winning or which dark horse to pick for their fantasy team. Our GC ranking is a decent starting point, but the specifics of a grand tour require something better. The GC ranking is designed to tell who is the best GC rider over the past 12 months, not who is the favorite for a specific race.

Before we formulate the 2023 favorites we need to take a look in the rear view mirror and find out how a typical grand tour is won. Riders can win time on their competitors in various ways, (1) in time trials, (2) climbing stages, (3) bonifications or (4) other ways (e.g. echelons, TTTs, etc). The amount of time won depends on the number of TT kilometres, number of climbing stages and  bonification size. The following chart depicts the time won by the final winner on the top-5 riders in Grand Tours of the last 10 years.

• 2:50

  170

  Mountains
• 0:09

  9

  Other
• 1:14

  74

  TTs
• 0:24

  24

  Bonis

As mountain stage we have taken stages with a minimal profileScore of 150 or a profileScore for the final of 75, excluding TTs. All time trials are included in the TT bucket, also TTs with a high profile score. All other stages are grouped together into the 'Other' bucket. The bonifications are considered per event. The time in the chart is the average time won on riders in the top-5 per event. In other words, the winner of a grand tour had an average time advantage of 4:37 on riders in the top-5. Out of this time, 61% was taken in mountain stages, 27% in time trials, 9% from bonifications and 3% in more or less flat stages. 

From this analysis we know that we have to focus on climbing and time trial capabilities when crafting a good formula. Instead of looking at the final GC points, we look at individual races and stages and use the specialty point scales instead of the PCS point scale. This way, also lower ranked time trial results are awarded points and there is less difference between a climbing stage and a climbing one day race. We take into account all results from the past 18 months, where we multiply results in the past 3 months with 2, the past 6 months with 1.5, the past 12 monts with 1 and before that with 0.5. From our analysis we know that more time can be won from climbing, so we value that with a higher factor than time trial. When we put this all all together we get the following chart, where the scores are normalized with the #1 set at 100%.