| public Execute ( ) : ABsoluteMaybeStatisticsResult | ||
| return | ABsoluteMaybeStatisticsResult | |
public override string ToString()
{
var retval = new StringBuilder();
ABsoluteMaybeStatistics.ABsoluteMaybeStatisticsResult result;
try
{
result = _statistics.Execute();
}
catch (ABsoluteMaybeException e)
{
return(e.Message);
}
if (result.InsufficientSampleSize)
{
retval.Append("Take these results with a grain of salt since your samples are so small: ");
}
if (result.BestOption.ConversionRate == result.WorstOption.ConversionRate)
{
retval.AppendFormat(
@"
Both options are currently converting at the same rate ({3}),
with [{0}] having {1} conversions from {2} participants. " ,
result.BestOption.Name,
result.BestOption.Conversions,
result.BestOption.Participants,
result.BestOption.ConversionRate.ToString("#0.##%")
);
}
else
{
retval.AppendFormat(
@"
The best option you have is: [{0}], which had
{1} conversions from {2} participants
({3}). The other option was [{4}],
which had {5} conversions from {6} participants
({7}). " ,
result.BestOption.Name,
result.BestOption.Conversions,
result.BestOption.Participants,
result.BestOption.ConversionRate.ToString("#0.##%"),
result.WorstOption.Name,
result.WorstOption.Conversions,
result.WorstOption.Participants,
result.WorstOption.ConversionRate.ToString("#0.##%")
);
}
if (result.ConfidenceLevel == 0)
{
retval.Append("However, this difference is not statistically significant.");
}
else
{
retval.AppendFormat(
@"
This difference is <b>{0} likely to be statistically significant</b>, which means you can be
{1} that it is the result of your options actually mattering, rather than
being due to random chance. However, this statistical test can't measure how likely the currently
observed magnitude of the difference is to be accurate or not. It only says ""better"", not ""better
by so much"". " ,
result.ConfidenceLevel.ToString("#0.##%"),
Descriptions[result.ConfidenceLevel]
);
}
return(retval.ToString());
}
