Climbing is fun! Since 2007 the International Federation of Sport Climbing (IFSC) organizes competitive sport climbing events. These events cover the disciplines of lead climbing, speed climbing, and bouldering (as well as a combined format).
Here, we focus on the available data of all IFSC bouldering competitions. In total, 50+ events with 2000+ climbers are analyzed. Who were and are the strongest competition climbers of their time? Can we find climbing styles by correlating who climbed similar boulders? Who is able to climb through qualification, semi-finals, and finals without falling off the wall? Which competitions were the hardest?
Thankfully, the IFSC provides a comprehensive and accessible archive of competition data on their IFSC info site. These data cover all IFSC bouldering events between 2007 and now, including qualifiation, semi-final, and final results for World Cups and World Championships. For events after 2007, in addition to all participants' rankings in a competition, their attempts to zone and attempts to top per boulder are indicated as well.
To automate data acquisition, we use a Python Selenium script which is provided in our GitHub repository. Data are compiled into a comprehensive Excel table which is used for subsequent analyses. Since climbers' names in the original data show inconsistent capitalization, we change them to lowercase for simplicity.
Our analyses are performed in a Python Jupyter notebook which is available in the same GitHub repository. For easier accessibility, the results are published here as GitHub Pages. Among other analyses, we cover athlete statistics, for individual years and overall aggregates, a visualization of similarities in topped boulders between climbers, as well as competition statistics.
Please feel free to use these data analyses, contribute to our repository, and have fun gaining new climbing insights!
To start with, we compile a table of competition statistics by counting the number and fraction of participations in different competition rounds for each of the 2000+ climbers. Here and in the following we refer to absolute counts (#) and percentages (%) relative to the total number of competition participations. Moreover, we refer to qualification rounds (Q), semi-finals (S), finals (F), podiums (P), third (3), second (2), and first places (1), respectively.
Please click on a column header in the table to sort accordingly or use the search bar to filter for specific climbers.
The initial table view is sorted by the number of podium places (#P) which shows Akiyo Noguchi at the top with an impressive 60. Alternatively, sorting by the percentage of first places (%1) shows that an incredible Janja Gargret won 70% of all IFSC bouldering competitions she entered.
The previous statistics were aggregated across the whole recorded period. Now, we additionally visualize the number of finals (#F) per climber and year. For better visibility we show only athletes who participated in more than 3 finals. Additionally, we sort the figure by the climbers' highpoints in time (average of years weighted by number of finals).
The figure clearly shows a gradual shift when some climbers become stronger, stop competing eventually, and are a followed by new generations. We can also identify climbers who made finals in few years as compared to others who dominated the scene over long periods. Here, again, Akiyo Noguchi stands out with consistent final participation from 2007 up to her retirement from IFSC competitions in 2021.
Next, we dive one step deeper. We count the number of participations in different competition rounds over the years. We focus on the same set of climbers with more than 3 finals as before.
Use the dropdown menu in the figure to choose between the aggregation of all climbers' and individual results. Here, we highlight the young and astonishing career of Natalia Grossman, who participated in a handful of IFSC competitions before and burst onto the scene in 2021, making finals in every competition she joined since.
Due to the detailed competition data, we can even compare how similar different climbers performed in competitions and even individual boulder problems. To that end, we calculate the similiarity between two athletes as the fraction of identical boulders that they topped (between 0-100%). This yields a graph or network, where each climber is a represented by a node and their similarity is reflected by the edge width. To identify clusters of similar climbers, we keep only the 10% of edges with highest similarity scores.
Use the dropdown in the network visualization below to choose between men and women and years. The node/climber color indicates the number of podiums and the mouse-over tooltip shows additional climber statistics. The edge width reflects the similarity score between the two climbers in question and the mouse-over tooltip shows the boulders that were topped (1), not topped (0), or not attempted (-). (Info: The initially presented network is for 2021-W although the dropdown menu shows the first entry.)
Consequently, there are many networks to study. Here, we focus only on 2021-W and find two clusters of climbers, i.e. groups of climbers who topped similar sets of boulders. The first, small cluster includes Camila Moroni and Elena Krasovskaia. The second, larger cluster includes 10 athletes, most similar among them are Janja Garnbret and Natalia Grossmann who were the only ones able to climb all the hardest boulders.
Annotating different boulders with their styles (e.g. more static or dynamic, slabby or overhanging, slopy or juggy) might provide more informative interpretations for the found similiary clusters. Alternatively, the RIC method evaluating Risk, Intensity, and Complexity of a boulder problem might offer an informative and more establishe classifiation scheme. For now und since it is more prone to subjective assessment, unfortunately, this is beyond our current scope.
Topping a boulder in the first attempt, without falling once, is called a flash. Here, we count the number of falls per competation and climber. Clearly, this number is only well-defined for climbers who could top all boulder problems in all rounds a a given competition. Otherwise, the number of actual falls depends on the number of attempts, is not comparable or meaningful, and is therefore omitted.
The heatmap color below indicates the number of falls per climber and competition. Use the mouse-over to get more readable infos about the respective color point.
To start with, we note that the available data provide the set of topped boulders only after 2007 (see horizontal lines). Before that, color points only indicate participation in finals. Next, we see the sparsity of color points, showing that only few athletes ever managed to top all boulders in a competition. Notably, again, Janja Garnet achieved to top all boulders in 9 of her competitions!
Finally, we turn away from individual climbers' statistics and towards a short analysis of different competitions. As a simple measure of competition difficulty, we calculate the average fraction of tops across all boulders and climbers. More difficult competitions come with a lower rate of tops.
The heatmaps display the competition difficulty aggregated per year (top panel), per hosting country (middle panel), and non-aggregated (bottom panel). During most competitions, the rate of tops is between 30-50% (approx. first and third quartile). However, there are deviations. For example, the first 2022 Worldcup in Salt Lake City had a top rate as high as 77%. In contrast, the 2019 World Championship finals in Hachioji were hard with only 23% tops.
In the non-aggreated heatmap, we additionally provide average top rates per competition round. They support the general obervation that qualifiers (42%) are typically easier than semi-finals (39%), while finals frequently feature most tops (50%).
Thanks for your time, we hope you had fun and enjoyed some interesting results and climbing insights!
We'll update the current data set and analyses after the future IFSC bouldering events.
Please feel free to re-use or contribute to our work as explained here.
See you on the wall and ¡Vamos! Gamba, Gamba! давай! Geht schon! Dai! Ajde! Kom igen! Allez, Allez, Allez!