v2.0: Bayesian analysis based on data available in September 2014
Menu
 Data summary (standard χ^{2} analysis)
 Parameter ranges
 Credible intervals: θ_{23}
 Credible intervals: δ_{CP}
 Model comparison: θ_{23}
 Model comparison: δ_{CP}
 Twodimensional regions: Normal Ordering
 Twodimensional regions: Inverted Ordering
 Twodimensional regions: Mixed Ordering
 Onedimensional likelihoods: θ_{23}
 Onedimensional likelihoods: δ_{CP}
 Onedimensional likelihoods: J_{CP}
If you are using these results please refer to JHEP 09 (2015) 200 [arXiv:1507.04366] and JHEP 11 (2014) 052 [arXiv:1409.5439] as well as NuFIT 2.0 (2015), www.nufit.org.

Comparison of the results of χ^{2} and Bayesian analysis in the framework of threeflavor oscillations, using the global data available up to September 2014. Note that Δm^{2}_{3ℓ} = Δm^{2}_{31} > 0 for NO and Δm^{2}_{3ℓ} = Δm^{2}_{32} < 0 for IO. 

Standard deviations, credible intervals, and χ^{2} intervals for sin^{2}θ_{23}. NO, IO and MO denote Normal Ordering, Inverted Ordering, and no assumption on the mass ordering, respectively. 

Measures of dispersion, credible intervals, and χ^{2} intervals for δ_{CP}. Here [...]^{c} is the complement of the interval [...], i.e., all values of δ_{CP} not contained in [...]. NO, IO and MO denote Normal Ordering, Inverted Ordering, and no assumption on the mass ordering, respectively. 

Model comparison for different assumptions on sin^{2}θ_{23}. We list the logarithms of Bayes factors, the comparable differences in the Akaike Information Criterion (AIC), and the differences in χ^{2} minima. The sign is chosen such that positive values correspond to preference for first mentioned assumptions in each case, i.e., the 2nd octant and nonmaximal mixing, respectively. 

Model comparson for different assumptions on δ_{CP}: M^{1}_{CPC} ≡ {δ_{CP}=0°}, M^{2}_{CPC} ≡ {δ_{CP}=180°}, M_{CPC} ≡ {M^{1}_{CPC} ∨ M^{2}_{CPC}} with equal priors, and M_{CPV} ≡ {δ_{CP} ∈ [0°, 360°] ∖ M_{CPC}} with prior π(δ_{CP}) = 1/360°. We list the logarithms of Bayes factors relative to M_{CPV}, the comparable differences in the Akaike Information Criterion (AIC), and the differences in χ^{2} as Δχ^{2} = χ^{2}(M_{CPV})  χ^{2}(M^{i}_{CPC}). For all variables, positive values would indicate preference of the corresponding M^{i}_{CPC} over M_{CPV}. 
Twodimensional regions: Normal Ordering
pdf jpg 
Onedimensional posterior distributions (black full lines) and twodimensional 1σ, 2σ and 3σ Bayesian credible regions (black void contours) assuming Normal Ordering. The figure also shows the onedimensional profile likelihoods (red dashed curves) and twodimensional χ^{2} regions (colored filled regions) from the standard χ^{2} analysis. 
Twodimensional regions: Inverted Ordering
pdf jpg 
Onedimensional posterior distributions (black full lines) and twodimensional 1σ, 2σ and 3σ Bayesian credible regions (black void contours) assuming Inverted Ordering. The figure also shows the onedimensional profile likelihoods (red dashed curves) and twodimensional χ^{2} regions (colored filled regions) from the standard χ^{2} analysis. 
Twodimensional regions: Mixed Ordering
pdf jpg 
Onedimensional posterior distributions (black full lines) and twodimensional 1σ, 2σ and 3σ Bayesian credible regions (black void contours) without any assumption on the mass ordering. The figure also shows the onedimensional profile likelihoods (red dashed curves) and twodimensional χ^{2} regions (colored filled regions) from the standard χ^{2} analysis. 
Onedimensional likelihoods: θ_{23}
pdf jpg 
Bayesian posterior/marginal likelihood (black solid) and profile likelihood (black dashed) from the standard χ^{2} analysis, both normalized to their maximal value. We also plot the number of σ's (red solid) and √Δχ^{2} (red dashed) functions. The vertical lines show the posterior mean (yellow), the median (green), and the maximum of the marginal likelihood (cyan). The different panels correspond to Normal Ordering (top left), Inverted Ordering (top right) and Mixed Ordering (bottom). 
Onedimensional likelihoods: δ_{CP}
pdf jpg 
Bayesian posterior/marginal likelihood (black solid), profile likelihood (black dashed), number of σ's (red solid) and √Δχ^{2} (red dashed) as a function of δ_{CP}. The vertical lines show the posterior mean (yellow), the median (green), and the maximum of the marginal likelihood (cyan). The different panels correspond to Normal Ordering (top), Inverted Ordering (middle) and Mixed Ordering (bottom), both in cartesian (left) and polar (right) coordinates. 
Onedimensional likelihoods: J_{CP}
pdf jpg 
Bayesian posterior likelihood (black solid), profile likelihood (black dashed) and marginal likelihood (blue dotdashed), together with the number of σ's (red solid) and √Δχ^{2} (red dashed) as a function of J_{CP} (right) and its maximum value (left). The different panels correspond to Normal Ordering (top), Inverted Ordering (middle) and Mixed Ordering (bottom). 