The Arena Crowns Burners. The Score Crowns Operators.

The short version: ccwarriors

ccwarriors is a live burn-rate leaderboard covering 15 agents plus an "Other" bucket — whatever ccusage can read counts. GitHub OAuth verifies profiles, collectible warrior cards and README badges make ranks shareable, per-model uploads are priced server-side, and org boards let communities run their own skinned leaderboard — Network School runs the first one. It is the most game-complete implementation of the burn board, and the org-board model in particular is a genuinely good idea.

SigRank reads the same telemetry and asks the other question: not how much burned, but what the burn compounded into. Υ Yield = cache_read × output / input² ranks the efficiency of the cascade itself. The arena tells you who burns hottest; the score tells you who drives best.

Feature comparison

Burn rate is a crowd metric; yield is an operator metric

Burn leaderboards answer a social question: who is going hardest right now. ccwarriors does that well — hourly refresh, verified identity, collectible cards, org skins. What burn cannot express is efficiency: the leaderboard's top spot goes to whoever ran the most tokens through the most sessions, and two operators at the same burn can be an order of magnitude apart in what they got back.

Υ Yield makes the invisible part the ranking. Cache reuse and output compound the score; fresh input penalizes it quadratically. Add Leverage, SNR, Velocity, archetypes, and class tiers and the burn becomes a description of skill — the difference between watching the arena and reading the scouting report.

From the arena to the score

Already burning on the board? The same telemetry produces a signed efficiency score:

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

Or paste token counts into the /score calculator for an instant Υ Yield and class tier.

Frequently asked questions

Is ccwarriors a SigRank alternative?

It is the most game-complete burn board in the category — and the parts it got right are worth naming: collectible warrior cards, GitHub-verified profiles, README badges, server-authoritative pricing, and org boards so communities like Network School can run their own skinned leaderboard. If you want a burn-rate arena with real identity and org support, ccwarriors is it. SigRank is the different question on the same telemetry: not who burned the most, but whose cascade compounded.

What does ccwarriors measure?

Burn rate — tokens consumed per tool, per day, per model, priced server-side and ranked, refreshed hourly, with tiers and collectible cards. It also has an optional consented deep-mode for profile insights. It is honest about what it ranks: burn. SigRank starts from the same four token pillars and derives the efficiency layer instead — Υ Yield (cache_read × output / input²), Leverage, SNR, Velocity, archetypes, and class tiers. The arena measures appetite; the framework measures efficiency.

Can I use ccwarriors and SigRank together?

Yes — same local logs feed both. Keep the warrior card and the README badge for the burn crowd; the org-board model is a genuinely good idea and worth supporting. Then run `sigrank submit` to put a signed Υ score on the global efficiency board. Burn rate for the arena, yield for the record.

Which is better for evaluating operators?

SigRank. Burn-rate boards reward the operator who runs the longest sessions, not the one who drives them best — two warriors with the same burn can be ten-fold apart in output-per-input. Υ Yield makes that gap the ranking: cache reuse and output amplify the score while fresh input penalizes it quadratically. ccwarriors crowns the biggest burner; SigRank identifies the most efficient operator.