class RooStats::HypoTestInverterResult: public RooStats::SimpleInterval

 default constructor

 copy constructor

 constructor

 destructor
 no need to delete explictly the objects in the TList since the TList owns the objects

 Merge this HypoTestInverterResult with another
 HypoTestInverterResult passed as argument
 The merge is done by combining the HypoTestResult when the same point value exist in both results.
 If results exist at different points these are added in the new result
 NOTE: Merging of the expected p-values obtained with pseudo-data.
  When expected p-values exist in the result (i.e. when rebuild option is used when getting the expected
 limit distribuiton in the HYpoTestInverter) then the expected p-values are also merged. This is equivalent
 at merging the pseudo-data. However there can be an inconsistency if the expected p-values have been
 obtained with different toys. In this case the merge is done but a warning message is printed.

 Add a single point result (an HypoTestResult)

 function to return the value of the parameter of interest for the i^th entry in the results

 function to return the value of the confidence level for the i^th entry in the results

 function to return the estimated error on the value of the confidence level for the i^th entry in the results

 function to return the observed CLb value  for the i-th entry

 function to return the observed CLs+b value  for the i-th entry

 function to return the observed CLs value  for the i-th entry

 function to return the error on the observed CLb value  for the i-th entry

 function to return the error on the observed CLs+b value  for the i-th entry

 function to return the error on the observed CLs value  for the i-th entry

 get the HypoTestResult object at the given index point

 find the index corresponding at the poi value xvalue
 If no points is found return -1
 Note that a tolerance is used of 10^-12 to find the closest point

 return the X value of the given graph for the target value y0
 the graph is evaluated using linear interpolation by default.
 if option = "S" a TSpline3 is used

 interpolate to find a limit value
 Use a linear or a spline interpolation depending on the interpolation option

 if mode = 0
 find closest point to target in Y, the object closest to the target which is 3 sigma from the target
 and has smaller error
 if mode = 1
 find 2 closest point to target in X and between these two take the one closer to the target
 if mode = 2  as in mode = 1 but return the lower point not the closest one
 if mode = 3  as in mode = 1 but return the upper point not the closest one

std::cout << "finding point with cl = " << (1-ConfidenceLevel())/2 << endl;

 Return an error estimate on the upper(lower) limit.  This is the error on
 either CLs or CLsplusb divided by an estimate of the slope at this
 point.

 need to have compute first lower limit

 get the background test statistic distribution

 get the signal and background test statistic distribution

 get the expected p-value distribution at the scanned point index

 get the limit distribution (lower/upper depending on the flag)
 by interpolating  the expected p values for each point

 Get the expected lower limit
 nsig is used to specify which expected value of the UpperLimitDistribution
 For example
 nsig = 0 (default value) returns the expected value
 nsig = -1 returns the lower band value at -1 sigma
 nsig + 1 return the upper value
 opt = "" (default) : compute limit by interpolating all the p values, find the corresponding limit distribution
                         and then find the quantiles in the limit distribution
 ioption = "P" is the method used for plotting. One Finds the corresponding nsig quantile in the p values and then
 interpolates them

 Get the expected upper limit
 nsig is used to specify which expected value of the UpperLimitDistribution
 For example
 nsig = 0 (default value) returns the expected value
 nsig = -1 returns the lower band value at -1 sigma
 nsig + 1 return the upper value
 opt is an option specifying the type of method used for computing the upper limit
 opt = "" (default) : compute limit by interpolating all the p values, find the corresponding limit distribution
                         and then find the quantiles in the limit distribution
 ioption = "P" is the method used for plotting. One Finds the corresponding nsig quantile in the p values and then
 interpolates them

 get expected limit (lower/upper) depending on the flag
 for asymptotic is a special case (the distribution is generated an step in sigma values)
 distringuish asymptotic looking at the hypotest results
 if option = "P" get expected limit using directly quantiles of p value distribution
 else (default) find expected limit by obtaining first a full limit distributions
 The last one is in general more correct

 number of entries in the results array

 set the size of the test (rate of Type I error) (eg. 0.05 for a 95% Confidence Interval)

 set the confidence level for the interval (eg. 0.95 for a 95% Confidence Interval)

 set CLs threshold for exclusion cleanup function

 flag to switch between using CLsb (default) or CLs as confidence level

 query if one sided result

 query if two sided result

 same in terms of alt and null

 get expected lower limit distributions
 implemented using interpolation
  The size for the sampling distribution is given (by default is given by the average number of toy/point)

 get expected upper limit distributions
 implemented using interpolation

 set the interpolation option, linear (kLinear ) or spline  (kSpline)