Simulation of the Vasaloppet Cross-Country Ski race

Microscopic simulation of the infamous traffic jam after the start of the Swedish Vasaloppet cross-country ski race. Every skier is simulated individually. By such simulations, new options for the operations of this mass event can be simulated before implementing them. For example, would a staggered wave start instead of a simultaneous start help? If so, how to optimize the sizes and starting times of the starting blocks? See www.movsim.org.

Notice that the lane changes in congested situations are not yet very realistic (stuck skiers at the boundaries). I participated in the Vasaloppet 2012 race, and, now, I know why: There are no tracks in the tapered zones before the bottlenecks! This will be made more realistic in the next months

• Category

  Sports

[Tor Lindqvist]

This is a really interesting simulation for us ski neerds. One thing about the start that is heavily debated is whether you should go to the left or the right of the house in the first uphill. What does your simulation suggest?

[klettmax]

@TheTorLindqvist: I actually participated in the 2012 race! Besides the challenge in itself, this serves as a first-hand calibration of the model, this means, changing the model and its parameters such that it quantitatively describes the current situation. I've got some ideas for that ... To your question: If you are very good (Wave 3 or better), you will encounter little jams. For the rest, keeping right consistently yields better results 

[klettmax]

NB: If I feel that this advice gets a big audience, I will keep to the LEFT at the next Vasaloppet. Unlike weather forecasts, traffic forecasts may well be self-destructing prophecies!

[danielbelucha]

Very interesting analysis! I wonder, how sensible the suggested optimization of the starting waves (staggered start) is, when the "parameters" of the individual skiers change, or the assumed snow conditions change.

It might be difficult to change the starting scheme at the fly, as weather and snow conditions change.

Great work!

[klettmax]

While the fine details depend on the parameters (i.e., on weather and ski conditions), and on the model quality, the big picture is pretty insensitive to that. Since the Elite and the first 4 blocks encounter little problems and the delay is 45-50 minutes for the last block 10, I propose a simultaneous start for the first 4 waves, and a staggered wave start with 5 min between for the remaining blocks.

[klettmax]

The simulation shows that, regardless of the starting scheme, the last skiers will more or less need the same time relative to the start of the *first* waves to pass the bottleneck causing the jam (at the end of the uphill slope). This means, delaying the start of the last waves by X minutes will essentially safe X minutes standing in the jam! Of course, this is only true if X is less than the delay for the present situation!

[swswsad]

Awesome analysis!