Understanding Glicko-2 Ratings

Rating and Rating Deviation (RD)

Each pilot has a rating (e.g. 1500) and a rating deviation (RD), which measures uncertainty. A lower RD means we are more confident in the rating. Ratings are updated after each event based on results and the RD.

Expected Score and Pairwise Outcomes

Race placings are converted into pairwise outcomes. For each pair of pilots in a race, we determine who finished ahead. The system computes an expected score for each pilot against each opponent. If a pilot performs better than expected, their rating goes up; if worse, it goes down. The size of the change depends on the opponent ratings and on each pilot's RD.

One Update Per Event

Unlike per-race updates, we apply one rating update per event. All races and qualifying in that event are combined into a single set of outcomes, and each pilot's rating and RD are updated once. This keeps ratings stable and reduces volatility.

Class and Overall

Pilots have separate ratings for each class (Open, ProSpec, Micro) and an Overall rating that combines all classes. The Overall rating uses the same Glicko-2 process over all race results in the event, giving a combined skill measure across formats.

What Determines Rating Changes?

The size of rating increases or decreases depends on several factors. Understanding these helps explain why some results lead to bigger changes than others.

1. Opponent Strength

Beating higher-rated pilots increases your rating more than beating lower-rated ones. Conversely, losing to lower-rated pilots decreases your rating more than losing to higher-rated ones. The system calculates an expected score based on rating differences — if you outperform expectations, your rating goes up; if you underperform, it goes down.

2. Opponent's Rating Deviation (RD)

Opponents with higher RD (less established ratings) contribute more to rating changes than opponents with lower RD (well-established ratings). This means beating or losing to newer pilots or pilots with uncertain ratings has a bigger impact.

3. Your Own Rating Deviation (RD)

New pilots (high RD) experience larger rating swings because the system is still learning their true skill level. Established pilots (low RD) have smaller changes because the system is more confident in their rating. As you compete in more events, your RD typically decreases — but not always (see below).

4. Actual vs Expected Performance

The system compares your actual results against expected results based on ratings. If you consistently beat pilots you were expected to lose to, your rating increases significantly. If you lose to pilots you were expected to beat, your rating decreases more.

5. Number of Outcomes

Events with more races and qualifying rounds provide more data points. Each race creates pairwise comparisons between all pilots, so more races mean more comparisons and potentially larger rating changes. However, the system combines all outcomes from an event into a single update, so the change is proportional to your overall performance across the entire event.

Why can someone with more events have a higher RD?

RD is not simply "more events = lower". Two things can keep RD higher even with more events and races:

Volatility (consistency): The system estimates how consistent your results are. If you sometimes beat much stronger pilots and sometimes lose to much weaker ones, it treats your performance as volatile and increases a hidden volatility value. That volatility is added into RD every event, so your RD can stay high or even rise despite more races. Consistent results (finishing about where ratings predict) keep volatility low and let RD drop.

Pairwise outcomes, not just race count: What matters for reducing RD is how many opponent comparisons you get per event (each pair of pilots in a race counts). Five races in big heats (e.g. 8 pilots) can give more comparisons than 12 races in small heats (e.g. 3 pilots). So fewer races with larger fields can sometimes reduce RD more than more races with smaller fields.

Rating Ranges (Guide)

Everyone starts at 1500 with high uncertainty (RD). After events, ratings move up or down. These bands are a rough guide; actual distribution depends on your field and number of events.

The ± RD next to a rating shows uncertainty: a lower RD means we are more confident in that number.

Class Ratings vs Overall

Pilots have separate ratings for each class and an Overall rating that combines all classes. This allows pilots to have different skill levels in different racing formats.

• Open — Standard racing class
• ProSpec — Professional spec racing
• Micro — Micro/mini racing class
• Overall — Combined rating across all classes

National and Club Ratings

You have one national rating and a separate rating at each club you race at. This lets you see how you rank across all of Australia and how you rank within a single club's events.

• National — Your rating based on all committed events across the country. This is your overall standing in Australian FPV racing.
• Club (e.g. Sydney FPV, Melbourne FPV) — A separate rating based only on events assigned to that club. If you race at multiple clubs, you will have a different rating for each — reflecting your results at that club's events only.

Use the View dropdown at the top of the site (National vs a specific club) to switch between these. The leaderboard, dashboard, and your pilot profile all respect this view, so you can compare national rankings with your standing at a particular club.

Seasons vs All-Time

Ratings can be viewed for a specific season or all-time. Season ratings only consider events within that season, while all-time ratings consider all events regardless of season.

This allows you to see both current season performance and historical performance across all racing activity.

Competition Tiers & Multipliers

Events are categorised into tiers that reflect their competitive significance. Higher-tier events have a multiplier applied, meaning ratings change more at important competitions.

The multiplier used for an event is snapshotted at the time the event is committed, ensuring historical integrity — even if tier multipliers change later, past events always use the multiplier that was active when they were committed.

Tier	Multiplier
CLUB	1.00×
REGIONAL	1.25×
NATIONAL	1.50×
INTERNATIONAL	1.75×