Welcome to LandS (Limits and Signifcance)

LandS is a statistical tool for calculations of exclusion limits and significances, taking into account

  • multiple channels (or, more generally, binned distributions),
  • all contributing backgrounds channel-by-channel without lumping them together within individual channels,
  • a detailed breakdown of systematic uncertainties and the associated systematic errors on event rates,
  • and a complete map of correlations between errors within and accross different channels.

The current version of LandS allows one to

  • Pick any of the predefined systematic error pdf's individually for each source of errors:
    • truncated Guassian (not recommended, but included to allow for compartisons with other tools)
    • log-normal
    • flat
    • Gamma-function (forthcoming, in development)
    • Gamma-function convoluted with log-normal (forthcoming, in development)
    • ABCD method pdf (forthcoming, in development)
  • Calculate exclusions limits for a given experimental observation or make projections for future measuremnts using:
    • (pseudo)-Bayesian approach (currently implemented priors: flat in signal cross section or "correlated")
    • Modified Frequentist approach (a.k.a. CLs) with no profiling of systematic errors
  • Evaluate significance of an event excess for a given experimental observation or make projections for future measuremnts:
    • Bayesian-Frequentist Hybrid method with no profiling of systematic errors
For more details, go to the LandS documentation link.


Other commonly used statistical tools of similar verstality in handling multiple channels (bins) with a detailed breakdown of signal and individual background rates, sources of uncertainties and associated systematic errors, and a complete breakdown of correlations:

This list is not exhaustive; more references will be added as we go. The differences and similarities between all these implementations in their interfaces, their level of versatality, speed, etc. So in general one should pick tools based on how well they match a given task. One can find more details on differences and similarities between all these tools in the LandS documentation note. Noteably, there is a significant overlap between all these tools, which allows for crosschecking the obtained results (none of the tools is fool-proof).