Competition Notes: Can You Predict a Fencing Tournament outcome?

Before the Cadet SJCC in Las Vegas, I was seeded 63rd out of 171 fencers.

That got me thinking about a question: How much does seeding actually tell us about what will happen in a tournament?

What Is Seeding?

In fencing, seeding is basically a ranking system used to organize a competition. Higher-seeded fencers have usually earned more points through previous results, while lower-seeded fencers are still working their way up the rankings. At first glance, predicting tournament results seems easy. Just pick the higher seed. Most of the time, you'd probably be right. But if seeding were perfect, there would never be upsets. And anyone who fences knows that upsets happen all the time.

Looking Beyond Seeding

Before this tournament, my dad and I talked about what factors might help predict performance. Seeding is one obvious factor. But it's not the only information that's available.

On FencingTracker, you can also find things like:

• National ranking

• Recent tournament results

• Win-loss records

• Pool performance

• Strength of schedule

• Head-to-head history in some cases

The question is whether any of those measurements add useful information beyond seeding. Or is seeding already capturing most of what matters?

What Seeding Doesn't Measure

One thing that stood out to me is that seeding only measures results. It doesn't measure the athlete. For example, seeding doesn't tell you:

• Height

• Reach

• Age within an age category

• Training volume

• Recent injuries

• Sleep and recovery

• Travel fatigue

• Mental confidence

• Competitive experience

Some of these factors clearly affect performance. But nobody includes them in tournament predictions. Maybe that's because the data isn't available. Or maybe those factors matter less than we think. I honestly don't know yet.

Why Do Underdogs Win?

This is the part that interests me most. If a lower-seeded fencer beats a higher-seeded fencer, what happened? Was it random? Did the lower-seeded fencer improve faster than their ranking suggests? Was it a favorable style matchup? Did one athlete have a better day physically or mentally? Or are there hidden patterns we haven't discovered yet?

The Start of a Research Project

The Las Vegas tournament made me realize that I don't just want to know who won.

I want to know why. Over the next few years, I want to explore whether tournament outcomes can be predicted using publicly available fencing data. Can we build a model that predicts results better than seeding alone? Which metrics actually matter? Which ones don't? And are there factors nobody is measuring today that could explain why some athletes consistently outperform expectations?

I don't have the answers yet.

But every tournament creates another data point.

And that's what makes the question interesting.