EVShowcase
Donate

Methodology

EVShowcase Research Model.

V1.2· current release

The Research Model is the algorithm + curated data tables that turn raw EPA fuel economy and EIA price data into the savings number you see on the calculator. It’s not the data — that comes from public government sources, cited on every result. It’s the interpretation layer: which gas car to compare against, how to translate “how often you charge” into miles per week, how to mix city vs highway efficiency, and a half-dozen other small decisions that shape the final dollar figure.

We publish the model with a version number for the same reason scientific papers do: so the answers are reproducible and improvements are auditable. Every component below has a way it gets better — usually powered by simple contribution missions that owners and curious folks can do in 30 seconds.

What’s in V1.2

Six components. Each is independently testable and replaceable.

  1. 01

    Gas-equivalent comparator table

    Pairs each EV with the gas car a buyer is most likely to cross-shop (e.g. BMW i4 → BMW 330i, Tesla Model Y → BMW X3). Used to compute the fuel-cost difference.

    Source

    Hand-curated by the EVShowcase team plus weighted votes from the Verify-a-Gas-Equivalent contribution mission.

    Upgraded by

    The Verify a Gas-Equivalent mission on /contribute. Five votes against a pairing flag it for review. Three suggestions of the same alternative replace it on the next release.

  2. 02

    Class-default MPG fallback

    When an EV isn't in our curated table, the model falls back to a class-average gas MPG (e.g. 'compact SUV → 27 mpg'). Applied to roughly 8% of vehicle queries today.

    Source

    EPA fueleconomy.gov class averages, updated annually.

    Upgraded by

    Each new comparator added via contributions reduces fallback usage. We're aiming for under 2% by V1.5.

  3. 03

    Charging-frequency → miles per week

    Translates a categorical 'how often you charge' answer (Daily / 4–5x / 2–3x / Weekly / Less) into a miles-per-week estimate, using EPA range and a usable-charge fraction (currently 80%).

    Source

    Default heuristic. The Annual Miles override input bypasses it for power users who know their exact mileage.

    Upgraded by

    Anonymized real-world annual mileage from opted-in users (the share-data checkbox on the calculator). Once 50+ owners per model have shared, we'll switch to a per-model empirical multiplier.

  4. 04

    City / highway efficiency blending

    Mixes the EV's city and highway kWh/100mi values based on the user's % highway slider. Same for the gas comparator's MPG.

    Source

    EPA dynamometer test cycles (city and highway), pulled from fueleconomy.gov.

    Upgraded by

    Coming in V1.3: regional bias (e.g. dense-urban ZIPs default to 30% highway, rural ZIPs default to 70%).

  5. 05

    Public DC fast-charging cost premium

    Mixes home-rate electricity with a $0.45/kWh public-DC-fast estimate based on the user's % DC fast slider.

    Source

    Composite of Electrify America, EVgo, and Supercharger network public rates as of model release.

    Upgraded by

    The Charging Cost Survey mission on /contribute (planned for V1.3) will gather real receipts to calibrate per-network rates.

  6. 06

    Real-world winter range

    Not yet in the savings number — currently shown only as a forum aggregate when 5+ owners report. Will affect range-anxiety and trip-cost estimates in V1.3.

    Source

    The Real-World Winter Range survey on /contribute.

    Upgraded by

    Direct: every survey response moves the per-model average. Surveys are anonymized and aggregated only above the 5-owner k-anonymity floor.

How the model gets better

Each contribution mission feeds a specific component. Nothing here requires expertise — most missions take under a minute and earn XP toward your trust level.

Verify a gas-equivalent

+10 XP

Component 01 — comparator table

Start mission →

Real-world winter range

+25 XP

Component 06 — winter range (V1.3 release)

Start mission →

Tag a news headline

+5 XP

News filters (separate from the savings model)

Start mission →

We publish a new minor version (V1.x) every 6–10 weeks once enough contributor data has accumulated to move at least one component. Major versions (V2.x) when the methodology itself changes shape — for example, adding total-cost-of-ownership.

See active missions →

Version history

  1. V1.22026-04
    • Added city/highway blending via slider for both EV and comparator
    • Added public DC fast-charging cost mix slider
    • Refined fallback class-default MPGs from 6 buckets to 14 buckets
    • Curated comparator table grew from 32 to 50 EVs
  2. V1.12026-02
    • Replaced single national gas/electricity averages with EIA regional rates
    • Added EIA_API_KEY graceful fallback when key isn't set
    • Cookie-bound result share links
  3. V1.02026-01
    • Initial release: 32 hand-curated EV→gas pairings
    • Charging-frequency dropdown → miles/week multiplier
    • Hard-coded national gas + electricity averages

What’s coming

  1. V1.3 (planned)2026-Q3
    • Per-model winter range applied to savings (after 50+ owner survey responses)
    • Charging Cost Survey mission to calibrate per-network DC fast rates
    • Regional default for highway % based on ZIP density
  2. V2.0 (researching)TBD
    • Battery-degradation curve per model (8-year/100k-mile expected capacity)
    • Total-cost-of-ownership: incentives, insurance, maintenance
    • Used-EV pricing comparison

Open methodology

The savings math itself lives in a single ~80-line file: lib/savings.ts. The comparator table is lib/comparator.ts. If you spot a bug or want to suggest an improvement, file an issue on the repo or email contact@qxmedia.us.

The model is owned by EVShowcase but published openly. Researchers, journalists, and educators are welcome to cite it. Suggested citation: EVShowcase Research Model V1.2 (2026).