🏆 Top Performers

Methodology

Every point places a bike on two axes relative to a fixed reference: the Zwift Carbon frame with Zwift 32mm Carbon wheels at upgrade Level 0, which sits at the origin (0, 0). That reference carries a frame CdA bias of +0.012 m² at 5,050 g and wheels at +0.000 m² / 2,350 g — a combined 0.012 m² bias and 7,400 g. Up and to the right is faster — more aerodynamic and lighter than the reference.

Where the numbers come from

Each frame and wheel in Zwift carries an intrinsic CdA bias (an aerodynamic drag offset applied to the rider's frontal-area model) and a mass. Those values are extracted directly from the game's own data files, so the chart reflects the physics the game actually simulates rather than field-test estimates.

• Frame & wheel CdA bias and mass are read from the extracted game_frames.json and game_wheels.json.
• Road & TT bikes use separate frame selections and defaults. The Bike Type toggle switches the chart and selection lists while retaining your choices for each type and the selected upgrade level. TT frames use each wheel's TT-specific CdA bias; the road reference bike stays unchanged.
• Upgrade levels — each frame has a five-step upgrade ladder. The Level 0 (stock) figures are read straight from the game assets, but the per-level upgrade deltas are not stored there; they are estimated by matching each frame against ZwiftInsider speed-test data (see below). The toggle switches every plotted frame between Level 0 (stock) and Level 5 (fully upgraded); the reference bike is always held at Level 0.
• Halo bikes (Tron, R4000 Roller Blade, Project 74, Pinarello Espada) have permanently attached wheels, so they appear as a single point using their built-in wheelset rather than one point per selectable wheel.

How each axis is computed

For a given upgrade level, a bike's totals are the sum of its frame and wheel components. The reference totals are computed the same way, always at Level 0. The plotted reduction is simply the reference minus the bike, so positive means better:

CdA reduction (m²) = ref_CdA − (frame_CdA + wheel_CdA)

For TT frames, wheel_CdA is the wheel's TT-specific bias, including built-in wheelsets.

Weight reduction (g) = ref_weight − (frame_weight + wheel_weight)

Because the reduction is measured relative to a common reference, the rider's own baseline frontal area cancels out — the chart is independent of rider size and only reflects equipment differences.

Upgrade deltas & validation

The per-stage upgrade effects are cross-checked against ZwiftInsider's published speed tests. A 2-equation steady-state solver back-solves each tested bike's CdA and mass from its flat and climb average speeds, and those measured stage deltas become the per-level upgrade estimates shown here. The frame-to-test mapping used for that validation lives in frame_zi_match.json.

Frame & wheel selection

The frames and wheelsets shown here mirror the curated set on the ZwiftInsider Top Performers road and TT charts, so each bike type can be compared against its chart directly.

Download the data

The raw extracted datasets behind this chart: