Understanding rankings

How Dragon Calc uses Elo and MultiElo to estimate crew strength, update rankings, and help crews think strategically about race entry.

What is an Elo rating?

The BDA uses an Elo-based ranking system to estimate the strength of every crew.

Originally developed for chess, Elo works well for sport because it constantly updates based on who beats who and by how much.

The basic idea is simple:

Your rating is not a league position or a medal count. It is an estimate of your crew’s current competitive strength based on recent race results.

This rating sets an expectation of your performance at future events, and the system reacts when your result is better or worse than that expectation.

What is MultiElo?

Traditional Elo systems work best for one-on-one competitions. That obviously does not reflect a dragon boat race day.

MultiElo is an extension of Elo designed for races with multiple competitors. Instead of treating a race as winner versus loser, the system compares every crew against every other crew in the race.

For example, if 8 crews race:

The system then looks at:

  1. What the expected finishing order should have been
  2. What actually happened
  3. How surprising the result was
  4. The number of crews at the event

The more surprising the result, the bigger the rating change. The system expects higher-ranked crews to beat lower-ranked crews.

What data is added to the system?

After each race event, every entered crew is given a finishing position for each race. Those finishing positions are used to calculate a composite position.

For example:

Crews are then ranked by this composite number:

  1. Crew B — 1st, with a composite position of 1.33
  2. Crew A — 2nd, with a composite position of 2.66

How Elo ratings change

Winning crews take rating from crews they beat. Losing crews give up rating to crews that beat them, based on race position.

If a highly rated crew beats lower-rated crews, that was expected. The system effectively says: “you performed roughly as predicted.” The rating movement is therefore small.

But if a lower-ranked crew suddenly beats several strong crews, the system reacts strongly because the result was unexpected.

The ranking system also adjusts its expectations based on the number of crews at an event. The more crews in attendance, the more ratings can swing.

Beating strong crews is worth far more than beating weak crews, especially at larger events.

This means:

How do composite crews get points?

Not to be confused with the composite position. A composite crew is a squad made up of more than one team. For a team to count as a composite it needs to have 4 or more members of another squad.

Composite crews act as a stand-alone crew with an ELO rank made up of its parent crews and based on the number of paddlers from each crew.

Ranking points lost and won by that crew are then distributed back to their parent crews - again based on the number of paddlers from each crew.

Example 1

Composite Dragons - a new crew consisting of 16 paddlers from Team A (1016 ELO) and 4 paddlers from Team B (1059 ELO)

Attendance is also worked out the same way so Crew A gains 0.8 attendance and Crew B gains 0.2

Example 2

Combined Dragons - a new crew consisting of 10 paddlers from Team A (900 ELO) and 10 paddlers from Team B (1100 ELO)

Crew A gains 0.5 attendance and Crew B gains 0.5 attendance

How do I choose which race to attend?

If your crew has a low rank but is expected to perform strongly at a particular event, racing strong crews is beneficial.

Even mid-pack finishes can gain rating if expectations were low.

For example, if your crew is ranked low in the league and finishes above several crews ranked higher, your crew’s rating can rise significantly even without winning individual events.

This is one of the fastest ways to climb.

Should a strong crew attend a race if they expect to underperform?

This is one of the most interesting strategic questions in an Elo system. The honest answer is that it depends on what the crew values more: ranking protection or long-term competitive development.

What the rating system sees

The ranking system does not know:

It only sees:

If a top-ranked crew arrives weakened and finishes below expectation, the system will treat that as genuine underperformance, and that can lead to rating losses.

Option 1: protect the rating

A crew may choose not to attend if:

  • Key paddlers are unavailable
  • Travel massively weakens the squad
  • They expect a very poor result
  • The event offers little upside

This is rational, especially for highly ranked crews attending small, weak race fields.

For example, if a crew is ranked 3rd, missing several top paddlers, and considering a small regional race with mostly lower-ranked crews, the risk may outweigh the reward.

Option 2: race anyway

The alternative approach is to race whoever turns up.

This can be a healthier long-term approach because:

  • Strong crews usually still outperform expectations, even when weakened
  • Depth becomes valuable
  • Consistency matters
  • Experience matters

What can I use as a rule of thumb?

Probably worth racing if:

  • The field is strong
  • The race is well attended
  • The crew can still compete reasonably
  • Development matters and your crew needs experience

Consider skipping if:

  • Your lineup is dramatically weakened
  • Expectations are unrealistically low
  • The event is poorly attended
  • The potential rating downside is very large
  • Travel cost or fatigue outweighs value