SIGNALAF

COMPARE

Operator A
vs
Operator B
8METRIC WINS2
#1 · 976.9KRANK · Υ#2 · 299.0K
88.2K/dTHROUGHPUT—

OPERATOR A LIVEGLOBAL #1
Karthikeyan KKkk_spartans
IGNITER I976.9KΥ YIELD
◆ LEADING · 8–2
  • Leverage · closed loop · 16.8K×
  • Υ Yield · leads 3.3× · 976.9K
  • Cost · cost leader · $0.52/1M
  • Velocity · fast · 58.2 o/i
  • Efficacy · strong · 4428.17
Karthikeyan KK · composite
first: #1 Karthikeyan KKmiddle: #11 Georgelast: #22 DirtyDataYield: 21 of 22 · 976,944.2293Leverage: 21 of 22 · 16,785.7105SNR: 22 of 22 · 0.9831Velocity: 22 of 22 · 58.200910xDEV: 21 of 22 · 4.2249Efficiency: 22 of 22 · 4,428.1743Cost per 1M: 14 of 22 · 0.5177Yield: 21 of 22 points, value 976,944.2293Υ YieldYield: value 976,944.229321Leverage: 21 of 22 points, value 16,785.7105LeverageLeverage: value 16,785.710521SNR: 22 of 22 points, value 0.9831SNRSNR: value 0.983122Velocity: 22 of 22 points, value 58.2009VelocityVelocity: value 58.20092210xDEV: 21 of 22 points, value 4.224910xDEV10xDEV: value 4.224921Efficiency: 22 of 22 points, value 4,428.1743EfficiencyEfficiency: value 4,428.174322Cost per 1M: 14 of 22 points, value 0.5177$/1MCost per 1M: value 0.517714

#1 Karthikeyan KK#11 George#22 DirtyData

Points among 22 live operators (1 to 22); outward = more. Cost per 1M: cheaper is outward. Token-flow shape, not a quality score.

Karthikeyan KK · composite: points among 22 live operators (1 to 22, higher is further out; cost per 1M inverted so cheaper is higher)
MeasureValuePoints
Yield976,944.229321
Leverage16,785.710521
SNR0.983122
Velocity58.200922
10xDEV4.224921
Efficiency4,428.174322
Cost per 1M0.517714
OPERATOR B LIVEGLOBAL #2
plaetteplaette91
IGNITER I299.0KΥ YIELD
  • Leverage · closed loop · 11.8K×
  • Υ Yield · trails · 299.0K
  • Cost · cost leader · $0.38/1M
  • Velocity · fast · 25.3 o/i
  • Efficacy · strong · 3001.82
plaette · composite
first: #1 Karthikeyan KKmiddle: #11 Georgelast: #22 DirtyDataYield: 20 of 22 · 299,028.1982Leverage: 20 of 22 · 11,799.9359SNR: 21 of 22 · 0.962Velocity: 21 of 22 · 25.341510xDEV: 20 of 22 · 4.0719Efficiency: 21 of 22 · 3,001.8192Cost per 1M: 22 of 22 · 0.3835Yield: 20 of 22 points, value 299,028.1982Υ YieldYield: value 299,028.198220Leverage: 20 of 22 points, value 11,799.9359LeverageLeverage: value 11,799.935920SNR: 21 of 22 points, value 0.962SNRSNR: value 0.96221Velocity: 21 of 22 points, value 25.3415VelocityVelocity: value 25.34152110xDEV: 20 of 22 points, value 4.071910xDEV10xDEV: value 4.071920Efficiency: 21 of 22 points, value 3,001.8192EfficiencyEfficiency: value 3,001.819221Cost per 1M: 22 of 22 points, value 0.3835$/1MCost per 1M: value 0.383522

#1 Karthikeyan KK#11 George#22 DirtyData

Points among 22 live operators (1 to 22); outward = more. Cost per 1M: cheaper is outward. Token-flow shape, not a quality score.

plaette · composite: points among 22 live operators (1 to 22, higher is further out; cost per 1M inverted so cheaper is higher)
MeasureValuePoints
Yield299,028.198220
Leverage11,799.935920
SNR0.96221
Velocity25.341521
10xDEV4.071920
Efficiency3,001.819221
Cost per 1M0.383522
MATCHUP · BAR STATISTICS
Karthikeyan KKVSplaette
8–2
HEAD-TO-HEAD · METRIC WINS
Karthikeyan KK

Raw · Flow · Total

plaette
18.2KInput34.8K
1.1MOutput882.2K
305.2MCache-read410.8M
15.8MCache-write6.3M
322.1MTotal418.0M
$0.52Cost$0.38
88.2K/dThroughput—
322.1M∑ Total418.0M
Karthikeyan KK

Metrics

plaette
976.9KΥ Yield299.0K
98.3%SNR96.2%
16.8K×Leverage11.8K×
58.2Velocity25.3
4.2210xDEV4.07
$0.52Cost per 1M$0.38
4428.17Efficiency3001.82
16786:1:58.2Op Ratio11800:1:25.3

Head to head

Cascade stages points in the field

Karthikeyan KKplaette

first: #1 Karthikeyan KKmiddle: #11 Georgelast: #22 DirtyDataKarthikeyan KK: Transmission = 58.2 · 22 of 22Karthikeyan KK: Commitment = 14.9 · 20 of 22Karthikeyan KK: Reuse = 19.3 · 7 of 22Karthikeyan KK: Leverage = 16.8K · 22 of 22Karthikeyan KK: Υ Yield = 977K · 22 of 22plaette: Transmission = 25.3 · 21 of 22plaette: Commitment = 7.18 · 11 of 22plaette: Reuse = 64.8 · 19 of 22plaette: Leverage = 11.8K · 21 of 22plaette: Υ Yield = 299K · 21 of 22Transmission = output ÷ inputTransmissionKarthikeyan KK: Transmission = 58.2 · 22 of 22 points (more)● 22plaette: Transmission = 25.3 · 21 of 22 points21Commitment = cache write ÷ outputCommitmentKarthikeyan KK: Commitment = 14.9 · 20 of 22 points (more)● 20plaette: Commitment = 7.18 · 11 of 22 points11Reuse = cache read ÷ cache writeReuseKarthikeyan KK: Reuse = 19.3 · 7 of 22 points7plaette: Reuse = 64.8 · 19 of 22 points (more)● 19Leverage = T × C × R = cache read ÷ inputLeverageKarthikeyan KK: Leverage = 16.8K · 22 of 22 points (more)● 22plaette: Leverage = 11.8K · 21 of 22 points21Υ Yield = Leverage × Transmission = cache read × output ÷ input²Υ YieldKarthikeyan KK: Υ Yield = 977K · 22 of 22 points (more)● 22plaette: Υ Yield = 299K · 21 of 22 points21

#1 Karthikeyan KK#11 George#22 DirtyData

Where a Yield gap comes from. Transmission × Commitment × Reuse = Leverage (cache read ÷ input); Υ = Leverage × Transmission. Dot = more points on that axis.

Cascade stages: points among 22 live operators (1 to 22)
AxisFormulaKarthikeyan KKplaetteMore
Transmissionoutput ÷ input58.2 (22)25.3 (21)Karthikeyan KK
Commitmentcache write ÷ output14.9 (20)7.18 (11)Karthikeyan KK
Reusecache read ÷ cache write19.3 (7)64.8 (19)plaette
LeverageT × C × R = cache read ÷ input16.8K (22)11.8K (21)Karthikeyan KK
Υ YieldLeverage × Transmission = cache read × output ÷ input²977K (22)299K (21)Karthikeyan KK
Token mix share of own total

Karthikeyan KKplaette

first: #1 Karthikeyan KKmiddle: #11 Georgelast: #22 DirtyDataKarthikeyan KK: Input ↓ = <0.01% · 22 of 22Karthikeyan KK: Output = 0.33% · 5 of 22Karthikeyan KK: Cache write = 4.9% · 17 of 22Karthikeyan KK: Cache read = 95% · 13 of 22plaette: Input ↓ = <0.01% · 21 of 22plaette: Output = 0.21% · 1 of 22plaette: Cache write = 1.5% · 4 of 22plaette: Cache read = 98% · 22 of 22Input ↓ = input share of own total (less is outward)Input ↓Karthikeyan KK: Input ↓ = <0.01% · 22 of 22 points (more)● 22plaette: Input ↓ = <0.01% · 21 of 22 points21Output = output share of own totalOutputKarthikeyan KK: Output = 0.33% · 5 of 22 points (more)● 5plaette: Output = 0.21% · 1 of 22 points1Cache write = cache write share of own totalCache writeKarthikeyan KK: Cache write = 4.9% · 17 of 22 points (more)● 17plaette: Cache write = 1.5% · 4 of 22 points4Cache read = cache read share of own totalCache readKarthikeyan KK: Cache read = 95% · 13 of 22 points13plaette: Cache read = 98% · 22 of 22 points (more)● 22

#1 Karthikeyan KK#11 George#22 DirtyData

How each operates, not how much: each pillar's share of the operator's own tokens, ranked against the field. Input share is flipped, so less input is outward.

Token mix: points among 22 live operators (1 to 22)
AxisFormulaKarthikeyan KKplaetteMore
Input ↓input share of own total (less is outward)<0.01% (22)<0.01% (21)Karthikeyan KK
Outputoutput share of own total0.33% (5)0.21% (1)Karthikeyan KK
Cache writecache write share of own total4.9% (17)1.5% (4)Karthikeyan KK
Cache readcache read share of own total95% (13)98% (22)plaette
Υ YIELD · OVERTIMEKarthikeyan KKplaette0.0467.2K934.5K1.4M1.9M976.9K299.0K07-2407-26
Υ YIELD · OVERTIMEKarthikeyan…plaette0.0934.5K1.9M976.9K299.0K07-2407-26

FAQ · 3 questions

How do I compare two AI operators on SigRank?

Visit signalaf.com/compare and select two operators using the dropdown selectors. The comparison shows Yield (Υ), SNR, Leverage, Velocity, 10xDEV, and blended cost side by side, plus each operator's own raw and metric radars and a shared Yield-over-time chart. You can share the comparison via URL.

What metrics does the SigRank comparison show?

The head-to-head comparison shows six cascade metrics: Yield (Υ = cache_read × output / input²), SNR (signal-to-noise ratio), Leverage (cache reuse efficiency), Velocity (output per token), 10xDEV (log-normalized value above replacement), and blended cost ($/1M tokens). Individual raw and metric radar charts describe each live operator, with their recorded Yield history beneath.

Can I compare myself against another AI operator?

Yes. If you are enrolled on SigRank, the Compare Against Me feature lets you select yourself as one of the comparison operands. You can compare your cascade metrics against any other operator on the leaderboard to see exactly where and why you differ.