Part 2: Scratching the Surface in Snowboard Cross

Introduction

            In Part 1 of our deep dive into snowboard cross, we found that head-to-head heats are a meaningfully different – and more difficult – skill from individual time trials. Now, for part 2, we investigate how the course and conditions impact these differences. To do this, I analyzed FIS World Cup and World Championship results from events having both a qualifying (time trial) round and a heats round from the start of the 2017 season through March of 2026.

The Model

            To determine the impact of course conditions on the difference between time trial results and heat results, I built a model based on an athlete’s qualifying rank from that event’s time trial and the qualifying rank of the others in their heat, incorporating various elements that make up course conditions to predict placement (1-4) within the heat. The impact of each of these elements was then calculated using SHAP values.

Small Finals and Third-Place Matches

            In initial versions of my model, the round was included as a factor. Most of the round values had very little impact on the model’s predictions, with one exception: the small final.

            In most snowboard cross events, the athletes who do not reach the final from the semifinal compete in a round that is referred to as the “small final”. In events in which each heat includes 4 competitors, the small final is designed to sort out rankings 5 through 8. And in the (highly unlikely yet technically possible) event that multiple participants in the “big final” are disqualified, the winner of the small final gets onto the podium.

            The effects for small finals are in the black box in the above graphic. The purple dots, representing heats that are not small finals, are all above 0. The yellow dots, representing heats that are small finals, are all below 0. And there is a gap between the yellow dots and the purple dots. Together, this is an indication that athletes do not compete as hard in small finals – after they have already been eliminated from contention for victory – as they do in other rounds.

            This effect likely transcends snowboard cross. A marquee recent example of it came in the third-place match of soccer’s World Cup. This game is not as intense or relevant as the semifinals or finals. In fact, of the 22 starters for France and England in 2026’s third place game, just 8 also started in their team’s semifinal loss.

Course Characteristics

            In the final modeling, I tested five course characteristics: gradient, vertical drop, course length, elements per km, and features per km, as listed in the official FIS competition notes for each event. Due to the difference in competitive depth between the men’s and women’s fields that I described in Part 1, I built different models for the men’s and the women’s fields. Scheduling (whether qualifying was on the same day as the rounds or a different day), temperature, and the difference in temperature from qualifying to rounds were also tested, but I observed no meaningful impact.

This is what the feature importances look like for the women’s field, in order from highest to lowest. The top 2, qualifying rank and mean opponent qualifying rank, reaffirm insights already uncovered in Part 1, as does the impact of field size. Those three, age, and heats completed are the most reliable predictors. Heats completed in this case is more likely a factor of the field filtering down to the athletes performing the best on the day rather than fatigue.

The season that the event is in also shows an impact. Compared to previous versions of this model, it likely took out the impact of gradient and added more on top of it, suggesting that the changes and development of both the women’s field and the courses over time have also made major changes to the dynamics of the racing. But there is one course-related factor that makes some impact on the women’s field: features per km. Essentially, on the women’s side, less technical courses create a little bit more chaos in the heats.

This is the graph of feature importances for the men’s field. Similar to the women, qualifying rank, field size, heats completed, age, and various opponent qualifying rank factors make the most impact. But the men do not, in fact, have a single course-related feature that makes a meaningful impact of any kind on heat outcomes.

            Combining these two graphs, a course layout designed to maximize chaos in results would be straight. This suggests cornering and passing zones as the areas where the very best make their biggest impact. This finding of snowboard cross as a heavily precision-based sport also matches the finding that more experienced an athlete is, the better they tend to perform in the heats relative to their time trial performance.

Home Field Advantage

            Quantifying home field advantage in snowboard cross can be a difficult ask. Over the 10 seasons of competitions in my dataset, two thirds of the events took place in just 6 countries. So, I calculated heat rank excess – performance in a heat minus expected performance in that heat – and compared them for home and away races.

Every individual rider varies, and sample sizes are invariably low due to the short season. But the distribution of heat rank excess difference between home and away races is roughly normal and centered near 0. This means that home course advantage is not significantly different between heats and the time trial. In other words, any home field advantage that may exist happens in the time trial. We will take a more definitive look at this phenomenon later in the process, once we have separated rider ability from that day’s qualifying round.

Conclusions and Next Steps

            In Part 1 of this analysis, we uncovered the reality that heats are more difficult than time trials. In Part 2, we attempted to discover what makes that true. We found surprisingly little impact from the course’s structure, including home course advantage. Given the evidence that we have right now, we can conclude that most of the difference between time trials and heats in snowboard cross comes from the presence of the other riders on course and the inherent chaos that makes this sport so entertaining. I will return to this analysis when I have used these conclusions to establish measurements of rider ability that do not come only from that event’s heats

            In Part 3, we will look at seeding and whether national governing bodies could maximize medal-winning probability by using the time trial rounds to spread their athletes across the bracket or by putting as many athletes near the top of the timetables as possible – and whether that would change if a medal were awarded for the time trial, as it is at world championships in kayak cross.