The Olympic step
How much faster does each nation go at the Olympic Games than at the World Championships just before them? The same measure across the flying 200, team sprint and team pursuit, with the track taken out, so that peaking, held-back equipment and selection show up as a single number per nation.
Nation × event
Percentage change in the nation's best time, Olympics against the preceding Worlds, for each event. Negative (green) is faster at the Games. The mean column averages the selected events the nation contested; the field row is the average of all nations. Pick one Games above; the trajectory chart below shows all six. By default nations are ordered by step × evidence: the mean step pulled towards the field average in proportion to how few riders it rests on (weighted = (riders·step + 8·field)/(riders + 8), with a team pursuit counting four riders, a team sprint three — two for women before Paris — and a flying 200 one), so a nation with two riders in one event cannot top the table on a single good day, while GB's or Germany's figure, built on ten or more riders across several events, barely moves. The weighted column shows the value used. "Order by programme budget" ranks instead by the published high-performance award to the cycling programme for that cycle (£m, confidence grade; hover for what it covers), with nations lacking a published figure below, ordered by weighted step. "Min. events" (default 2) hides nations with fewer event entries than that in the selected Games — a one-event, two-rider nation can post a huge step that is one good or bad day.
| Nation | F200 M | F200 W | TS M | TS W | TP M | TP W | Mean | n |
|---|
Team events compare like with like: women's team pursuit 3 km (2012) and 4 km (2016 on), women's team sprint 2-rider 500 m (2012) and 3-rider 750 m (2024); women's TS at Rio and Tokyo was 2-rider 500 m while the Worlds had moved to 3 riders, so those two have no comparison.
Time jumps
The same comparison in seconds: each line runs from the nation's Worlds time to its Olympic time for the selected Games, one block per event. Green went faster. Follows the raw / venue-adjusted switch.
Across the Games
Mean step per Games for the nations chosen below, with the field average dashed. A nation that sits below the dashed line is gaining more at the Games than its rivals.
Every nation, every event
One dot per nation and event at the selected Games; the highlighted nation is the first slot above. Spread of the field is the context for any single nation's number.
Method and what it can and cannot show
For each nation and event, the nation's best time at the Games is compared with its best time at the preceding Worlds (Melbourne 2004 → Athens, Manchester 2008 → Beijing, Melbourne 2012 → London, London 2016 → Rio, Berlin 2020 → Izu 2021, Glasgow 2023 → Paris 2024; the Ballerup 2024 Worlds came after Paris). Times are qualifying or the fastest round ridden.
Venue adjustment: for each event a separate log-linear model with a fixed effect per rider (flying 200) or nation (team events), a linear season trend, and a fixed effect per velodrome, fitted on every ride in the database (Worlds, Olympics, World Cup, Nations Cup, Europeans). The venue effect is what remains after the rider or nation and the season are accounted for. It is identified by riders and nations who appear at several venues, so it is weakest for tracks used once.
Single-meeting venues: Athens, Beijing and Izu appear in the database only as an Olympic Games, so the venue model cannot separate "fast track" from "everyone peaked here": the venue effect for those three absorbs part of the field's Olympic step, and the adjusted numbers for those Games are relative to the field rather than absolute. Paris 2024 was ridden at Saint-Quentin-en-Yvelines, which also hosted the 2015 and 2022 Worlds, and Rio, London, Melbourne and Manchester all hosted other championships, so their effects are properly anchored. Switch to raw times to see the unadjusted change.
What the step contains: peaking for the Games, equipment and positions held back for the Games, a different rider line-up or a changed team order, and weather on the day. It cannot separate those. A consistently large step across several Games and several events is the pattern to look for; a single big number for a single event is mostly noise, and a nation with two riders in an event will swing more than one with twelve.
Budget against step
For the nations with a published programme budget: £m per cycle (all cycling disciplines unless marked) against the mean Olympic step at the selected Games. If money bought the step, the points would run downhill to the right.
Budgets are the public high-performance award to the national federation (UK Sport, BMI, ASC/AIS, CONI, HPSNZ, Own the Podium, USOPC, Team Danmark), converted to sterling at mid-cycle rates; they cover road, BMX and MTB as well as track and exclude sponsor income, athlete salaries paid through the military or police (Germany) and services in kind. Netherlands and Switzerland publish only a track-programme line and are shown hollow. Grades: A official award, B credible press, C annualised or inferred. France, China, Japan, Poland, Russia and most others publish nothing usable.