The Copilot Isn't the Pilot. The Score Is.

The short version: notch-pilot

notch-pilot is an AI coding assistant — a copilot that helps you write code faster. It does its job well: it suggests, completes, and generates. But it tells you nothing about how efficiently you're operating it. Are you compounding cached context or burning fresh input every turn? Is your yield high or low? Where do you rank against other operators driving the same or different tools? notch-pilot can't answer any of those questions.

SigRank reads the token telemetry your copilot generates and scores the operator behind it. The headline metric, Υ Yield = cache_read × output / input², measures whether your cascade is compounding or burning. notch-pilot is the plane; SigRank is the lap timer. Both matter. Only one tells you whether you're winning.

Feature comparison

The tool isn't the operator

notch-pilot answers "can it write code?" Yes. But that's the wrong question for anyone who wants to know whether they're getting better at driving AI. Two operators can use the same copilot and get wildly different outcomes. One reuses cached context efficiently and produces 30K output tokens; the other re-sends the same context every turn and produces 3K. Same tool, ten-fold difference in signal. notch-pilot can't see that gap.

SigRank's headline metric, Υ = cache_read × output / input², measures exactly that gap. It rewards the operator who compounds cached context into output and penalizes the one who burns fresh input without leverage. notch-pilot gives you the copilot; SigRank tells you whether the cascade it produces is compounding or burning.

From copilot to scored operator

If you already use notch-pilot, you're generating the telemetry SigRank needs. Add the scoring layer the copilot can't provide:

npm install -g sigrank
sigrank enroll      # create your operator identity
sigrank submit      # reads logs, scores, signs, publishes

Prefer to inspect before you submit? Run sigrank me --dry-run to see your scored payload locally, or paste your token counts into the /score calculator to compute your Υ Yield, class tier, and compression ratio instantly, no account, no submission, just the numbers.

Frequently asked questions

Is SigRank a notch-pilot alternative?

They solve different problems. notch-pilot is an AI coding copilot — it helps you write code. SigRank scores the operator driving the copilot. You don't choose between them; you use notch-pilot (or any copilot) to code, then run `sigrank submit` to score how efficiently you drove it. The tool is not the operator; the operator is what gets ranked.

Why would I need SigRank if I already use notch-pilot?

notch-pilot helps you write code but tells you nothing about how efficiently you're operating it. Are you compounding cached context or burning fresh input every turn? Is your yield high or low? Where do you rank against other operators? SigRank answers all three by reading the token telemetry your copilot generates and scoring it. The copilot is the plane; SigRank is the lap timer.

What does notch-pilot not measure that SigRank does?

notch-pilot doesn't measure anything about the operator — it's a tool, not a measurement system. SigRank reads the four token pillars (input, output, cache-read, cache-write) from your AI coding sessions and derives the cascade architecture: Υ Yield (is signal compounding or burning?), compression ratio (output per input), SNR (signal density), Leverage (how much cached context amplifies your input), and Velocity (tokens per unit time). notch-pilot writes code; SigRank scores the writing.

Can I use both notch-pilot and SigRank?

Yes — that's the intended setup. Use notch-pilot (or Claude Code, Cursor, Copilot, or any of 15+ supported tools) to write code. Run `sigrank submit` to read the token telemetry from those sessions, score your cascade yield, sign it with ed25519, and publish to the leaderboard. The copilot and the scorecard are complementary, not competitive.

Which is better for improving my AI coding skill?

SigRank. notch-pilot can make you faster at writing code, but it can't tell you whether you're getting better at operating AI. Yield tracks that: an operator whose Υ is rising is compounding cached context more efficiently over time. An operator whose Υ is flat or falling is burning more without producing more signal. The copilot can't see that; the scorecard can.