Internal · not indexed · not linked from primary navigation

Component fixtures

Chart and table components rendered against synthetic values so their layout can be checked at realistic ranges. Labels are anonymised Model A through Model H and correspond to no real system. The public site renders every one of these components in its pending state.

ILLUSTRATIVE FIXTURE Public leaderboard (pending states) Release artifacts

Capability frontier — populated

Fig. F1 FARTScore against generation cost — illustrative fixture

FARTScore 0–100 · USD per generated sample

Positions are synthetic and exercise the plot at realistic ranges only.

Frontier membership does not imply production-grade flatulence. These positions are invented.

Source Synthetic fixture data ILLUSTRATIVE FIXTURE

Leaderboard rows — populated

Fig. F2 Ranked comparison with intervals — illustrative fixture

FARTScore 0–100 · higher is better

Illustrative fixture rows. Not measurements.
#	Label	FARTScore	95% CI	n	FCR	A	T	L	P	I	S	C	Cost	F/$	Result
1	Model A	71.4	[66.2, 76.1]	27	74%	3.1	2.9	2.6	3.0	2.8	2.5	3.2	$0.62	1.19	Fixture
2	Model B	64.8	[59.1, 70.2]	27	63%	2.8	2.7	2.2	2.9	2.5	2.3	3.0	$0.28	2.25	Fixture
3	Model C	58.3	[51.7, 64.4]	24	55%	2.6	2.3	1.9	2.7	2.4	2.1	2.8	$0.15	3.67	Fixture
4	Model D	52.1	[45.0, 58.6]	24	48%	2.4	2.0	1.7	2.4	2.1	1.8	2.6	$0.44	1.09	Fixture
5	Model E	47.6	[40.2, 54.3]	21	41%	2.1	1.8	1.5	2.3	1.9	1.6	2.4	$0.09	4.56	Fixture
6	Model F	39.2	[31.4, 46.7]	21	29%	1.7	1.5	1.2	2.0	1.6	1.3	2.1	$0.71	0.41	Fixture
7	Model G	31.8	[24.1, 39.0]	18	22%	1.4	1.2	0.9	1.7	1.3	1.0	1.8	$0.20	1.10	Fixture
8	Model H	20.0	[14.6, 20.0]	18	11%	1.1	0.9	0.7	0.6	0.8	0.9	1.2	$0.33	0.33	Fixture

Composite FARTScore with its 95% bootstrap interval, sample count, Flatulatory Compliance Rate, per-dimension means, generation cost, and Farts per Dollar.

Source Synthetic fixture data ILLUSTRATIVE FIXTURE

Fig. F3 Model A — dimension profile (illustrative fixture)

Rubric scale 0–4 per dimension

A Acoustic 22%

3.1
T Timing 16%

2.9
L Attribution 16%

2.6
P Physical 14%

3.0
I Instruction 12%

2.8
S Social 10%

2.5
C Coherence 10%

3.2

Rubric scale 0–4 · 0 failure · 1 poor · 2 mixed · 3 good · 4 excellent

Mean rater score on each of the seven weighted FARTScore dimensions, on the 0–4 rubric scale. Hatched fill marks a fixture value.

Source Synthetic fixture data ILLUSTRATIVE FIXTURE

Fig. F4 Failure-code incidence — illustrative fixture

Observations per 100 clips · multi-label

Codes are multi-label, so incidences are not expected to sum to 100.

Source Synthetic fixture data ILLUSTRATIVE FIXTURE

Scoring arithmetic

Worked with the same computeFartScore function the benchmark uses, so the displayed arithmetic cannot diverge from the implementation.

Model A · all seven dimensions applicable

72.2

Σw = 1.00 · D = A T L P I S C

Model B · all seven dimensions applicable

65.9

Σw = 1.00 · D = A T L P I S C

Model C · all seven dimensions applicable

60.0

Σw = 1.00 · D = A T L P I S C

F10 catastrophic cap

20.0

raw 71.5 → capped at 20

FB-C06 · silent condition, T and L not applicable

71.3

Σw = 0.68 · renormalized

Status vocabulary

Pending

No blinded evaluation has been locked. Used everywhere on the public site.

Illustrative fixture

Synthetic value used to develop a component. Confined to this page.

Official

Derived from locked blinded evaluations under a frozen protocol version. None exist.

Critical

Failure severity. Never the only encoding — every badge carries its own text.

Major

Failure severity.

Minor

Failure severity.