Measured Alongside
Measured alongside
SigRank doesn't measure tokens in a vacuum — it builds on a small ecosystem of token-usage tools, and ranks the architecture of the cascade on top of what they count. Credit where it's due:
Independent project — not affiliated with or endorsed by the tools above; names belong to their authors. Token counts only.
Standing on shoulders
The idea that an operator deserves a stat screen isn't new — it's just never been applied to AI. Every competitive game gives you one: K/D, accuracy, rank, a Hall of Fame. BlitzStars.com, built by sockrobber, showed what a real player stat board looks like for World of Tanks: Blitz — live stats, historical progression, clan rankings, a record of who actually runs the field. Not just “how much you played” but how well you played — the shape of your performance, tracked over time.
That design principle — measure the operator, not the tool — is what SigRank brings to AI coding. The game is different; the idea is the same. Credit where it's due.
blitzstars.com ↗FAQ
What tools does SigRank measure alongside?
SigRank builds on token-usage tools like ccusage, tokscale, and token-dashboard. These tools measure how much you use AI. SigRank ranks how well you use it — turning raw token counts into cascade efficiency metrics like Yield, Leverage, and SNR.
How is SigRank different from ccusage?
ccusage counts your token usage (how much). SigRank ranks your token efficiency (how well). SigRank reads the same session logs but computes cascade metrics — Yield (Υ = cache_read × output / input²), Leverage, Velocity, SNR — that measure whether your tokens compound signal or just burn volume.
What is the difference between measuring and ranking AI usage?
Measuring means counting tokens (input, output, cache reads). Ranking means comparing those counts across operators using efficiency metrics. SigRank does both: it measures your four token pillars locally, then ranks you against other operators on the leaderboard by Yield and related cascade metrics.