Guitar pedals are deeply subjective — but what if the collective taste of the community could surface a real ranking? Pedal Elo puts two pedals head-to-head and asks you to pick the better one. Every vote updates a shared leaderboard using the Elo rating system, the same math that ranks chess grandmasters, competitive video game players, and football clubs worldwide.
The app presents a random matchup from a library of guitar pedals. You click the one you prefer. Behind the scenes, both pedals' ratings update based on how surprising the result was — a nobody beating a top-ranked pedal earns a lot of points; expected wins earn almost nothing. Over enough votes from enough people, the leaderboard converges on a genuine community consensus.
Three voting modes let you slice the data different ways:
- Global — all pedals in the pool, one shared leaderboard
- Brand Pool — filter to specific brands you care about
- Brand Battle — force every matchup to pit one pedal from Brand A against one from Brand B, with a completely separate Elo pool so the results don't pollute the global rankings
An Analysis panel tracks the leaderboard in depth: a searchable rankings explorer, a brand-level report, the most hotly contested pedals (win rates nearest 50%), a live match feed, and an upsets tracker for when underdogs pull off a win against the odds.
An About panel explains the Elo system from first principles — the expected score formula, the rating update equation, and the tiered K-factor schedule the app uses.
Traditional 5-star ratings collapse — everything ends up between 3.8 and 4.6 and the ordering is meaningless. Elo forces a choice. You can't hedge by giving everything four stars. Every vote is a direct comparison, and the system rewards accurate predictions over time. A pedal that keeps beating higher-rated opponents will climb; one that keeps losing to lower-rated opponents will fall, regardless of how many glowing reviews it has elsewhere.
The app uses a tiered K-factor specifically tuned for crowdsourced voting:
- K = 64 for a pedal's first 30 matches — high volatility so new entrants find their level fast rather than getting anchored near the 1200 starting rating by a handful of early mismatches
- K = 32 for matches 31–100 — the standard FIDE rate, responsive but stable
- K = 16 for 100+ matches — a well-evidenced rating that resists manipulation from sudden vote streaks
- React + Vite — frontend
- Vercel — hosting and serverless functions
- Upstash Redis — shared persistent storage for rankings and vote history
api/redis.js— a thin Vercel serverless proxy that keeps Upstash credentials server-side; the browser never sees the token- PedalPlayground — open source pedal data and images, fetched fresh on every build
Built by Jeremy Abramson. Pedal data and images from PedalPlayground. Elo methodology based on the work of Arpad Elo, as standardised by FIDE for competitive chess.