The lack of evidence for superparticles at the CERN LHC, along with the rather high value of the Higgs boson mass, has sharpened the perception that what remains of supersymmetric model parameter space suffers a high degree of electroweak fine-tuning (EWFT). We compare three different measures of fine-tuning in supersymmetric models. 1. ∆ HS measures a subset of terms containing large log contributions to m Z (and m h ) that are inevitable in models defined at scales much higher than the electroweak scale. 2. The traditional ∆ BG measures fractional variation in m Z against fractional variation of model parameters and allows for correlations among high scale parameters which are not included in ∆ HS . 3. The model-independent ∆ EW measures how naturally a model can generate the measured value of m Z = 91.2 GeV (or m h ) in terms of weak scale parameters alone. We hypothesize an overarching Ultimate Theory (UTH) wherein the high scale soft terms are all correlated. The UTH might be contained within the more general effective SUSY theories which are popular in the literature. In the case of ∆ HS , EWFT can be grossly overestimated by neglecting additional non-independent terms which lead to large cancellations. In the case of ∆ BG , EWFT can be overestimated by applying the measure to the effective theories instead of to the UTH. The measure ∆ EW allows for the possibility of parameter correlations which should be present in the UTH and, since it is model-independent, provides the same value of EWFT for the effective theories as should occur for the UTH. We find that the well-known mSUGRA/CMSSM model is fine-tuned under all three measures so that it is unlikely to contain the UTH. The non-universal Higgs model NUHM2 appears fine-tuned with ∆ HS,BG 10 3 . But since ∆ EW can be as small as 7 (corresponding to 14% fine-tuning), it may contain the UTH for parameter ranges which allow for low true EWFT. *