AttemptGroupContrastScenario
Data scenarios · Implementation: mlsysim/engine/scenario_definitions/data.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
attempts independent attempts per fixture at each of pass_rates.
Notes
Equation: The calculation members below define success_yield, mixed_fraction, attempts_per_success.
Assumptions: Attempts have identical independent pass probability; each fixture consumes the full attempt count.
Excludes: Correlated attempts, early stopping, and unequal attempt costs.
References
Derivation: attempts independent attempts per fixture at each of pass_rates. Each member records the corresponding algebra or named physics function.
Examples
Evaluate the named configuration and display a modeled result.
>>> from mlsysim import Scenarios
>>> case = Scenarios.EightAttemptGroupContrast
>>> value = case.success_yield(case.pass_rates[0])
>>> print(f"{value:.3g}")
0.337Attributes
| Name | Description |
|---|---|
| attempts | |
| pass_rates |
Methods
| Name | Description |
|---|---|
| attempts_per_success | Expected attempts spent per fixture yielding any success; requires positive success yield. |
| mixed_fraction | Dimensionless probability of both pass and failure, via calc_mixed_outcome_fraction. |
| success_yield | Dimensionless probability of at least one pass, via calc_independent_success. |
attempts_per_success
engine.scenario_types.AttemptGroupContrastScenario.attempts_per_success(
pass_rate,
)Expected attempts spent per fixture yielding any success; requires positive success yield.
Attempts spent per fixture that yields a success.
Equation: attempts / success_yield(pass_rate).
mixed_fraction
engine.scenario_types.AttemptGroupContrastScenario.mixed_fraction(pass_rate)Dimensionless probability of both pass and failure, via calc_mixed_outcome_fraction.
Chance of at least one success and one failure.
Calculation: mlsysim.physics.agents.calc_mixed_outcome_fraction.
success_yield
engine.scenario_types.AttemptGroupContrastScenario.success_yield(pass_rate)Dimensionless probability of at least one pass, via calc_independent_success.
Chance of at least one success.
Calculation: mlsysim.physics.agents.calc_independent_success.
attempts
Type: int.
Declared default: required.
pass_rates
Type: List[float].
Declared default: required.