[BUG] ArgumentOutOfRangeException thrown from unit test (`TwoSamples_Less_Than_One_CPU_Cycle_Method`) on CI

[filzrev]

I've encountered following ArgumentOutOfRangeException on GitHub CI. (test-linux)
Log: https://github.com/dotnet/BenchmarkDotNet/actions/runs/15608653226/job/43964042169

I've only encountered it this one time.
Though it seems there is some conditions that cause ArgumentOutOfRangeException with -∞ value.
Because TwoSamples_Less_Than_One_CPU_Cycle_Method argument using fixed test data.

2025-06-12T11:00:32.0101433Z [xUnit.net 00:00:02.34]     TwoSamples_Less_Than_One_CPU_Cycle_Method(workload: [2.0049000000000001, 2.0141, 2.0194000000000001, 2.0253000000000001, 2.0264000000000002, ···], overhead: [1.9963, 2.0036999999999998, 2.0036999999999998, 2.0045999999999999, 2.0051000000000001, ···]) [FAIL]
2025-06-12T11:00:32.0102296Z [xUnit.net 00:00:02.34]       System.ArgumentOutOfRangeException : value=-Infinity, but it should be in range [0;1] (Parameter 'value')
2025-06-12T11:00:32.0103076Z [xUnit.net 00:00:02.34]       Actual value was -Infinity.
2025-06-12T11:00:32.0103576Z [xUnit.net 00:00:02.34]       Stack Trace:
2025-06-12T11:00:32.0104691Z [xUnit.net 00:00:02.34]            at Perfolizer.Common.Assertion.InRangeInclusive(String name, Double value, Double min, Double max)
2025-06-12T11:00:32.0105471Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.Common.Probability..ctor(Double value)
2025-06-12T11:00:32.0106296Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.Common.Probability.op_Implicit(Double value)
2025-06-12T11:00:32.0108096Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.SignificanceTesting.MannWhitney.MannWhitneyTest.PerformGreater(Sample x, Sample y, Threshold threshold, MannWhitneyStrategy strategy)
2025-06-12T11:00:32.0110304Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.SignificanceTesting.MannWhitney.MannWhitneyTest.PerformStrategy(Sample x, Sample y, AlternativeHypothesis alternative, Threshold threshold, MannWhitneyStrategy mannWhitneyStrategy)
2025-06-12T11:00:32.0111974Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.SignificanceTesting.MannWhitney.MannWhitneyTest.Perform(Sample x, Sample y, AlternativeHypothesis alternative, Threshold threshold)
2025-06-12T11:00:32.0113448Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.SignificanceTesting.Base.SignificanceTwoSampleTestBase`1.GetPValue(Sample x, Sample y, AlternativeHypothesis alternative, Threshold threshold)
2025-06-12T11:00:32.0114673Z [xUnit.net 00:00:02.34]            at Perfolizer.Mathematics.SignificanceTesting.SimpleEquivalenceTest.Perform(Sample x, Sample y, Threshold threshold, SignificanceLevel alpha)
2025-06-12T11:00:32.0116851Z [xUnit.net 00:00:02.34]         /home/runner/work/BenchmarkDotNet/BenchmarkDotNet/src/BenchmarkDotNet/Analysers/ZeroMeasurementHelper.cs(30,0): at BenchmarkDotNet.Analysers.ZeroMeasurementHelper.AreIndistinguishable(Sample workload, Sample overhead, Threshold threshold)
2025-06-12T11:00:32.0119239Z [xUnit.net 00:00:02.34]         /home/runner/work/BenchmarkDotNet/BenchmarkDotNet/tests/BenchmarkDotNet.Tests/Analysers/ZeroMeasurementHelperTests.cs(141,0): at BenchmarkDotNet.Tests.Analysers.ZeroMeasurementHelperTests.TwoSamples_Less_Than_One_CPU_Cycle_Method(Double[] workload, Double[] overhead)
2025-06-12T11:00:32.0120419Z [xUnit.net 00:00:02.34]            at System.RuntimeMethodHandle.InvokeMethod(Object target, Void** arguments, Signature sig, Boolean isConstructor)
2025-06-12T11:00:32.0121568Z [xUnit.net 00:00:02.34]            at System.Reflection.MethodBaseInvoker.InvokeDirectByRefWithFewArgs(Object obj, Span`1 copyOfArgs, BindingFlags invokeAttr)

[filzrev]

It's also confirmed this error is occurred on macos-13 (x64) runner.

[filzrev]

Close this PR, because it can't reproduced by running tests 100 times (On ubuntu-latest / macos-13)

As far as I've confirmed.
TwoSamples_Less_Than_One_CPU_Cycle_Method tests don't contains dynamic input data.
And calculated result should be stable value.