Data and Fit (log y-scale)
5 10 15 20 25 30 35
E=0.06186, (A,B,C)=(0.73848,4.7959e-005,0.26084)
Difference Plot (diff of logs)
0.5 K
40 45 50
5 10I I I I I I I I
5 10 15 20 25 30 35 40 46 50
Figure 4-12.
Data analysis and data fitting for M 12 simulations. We show here the
fitting procedure for the D&K comparison with M 12 and weak coupling.
At low coupling the signal to noise ratio is unfavorable. The fit is still rather
good and is achieved by fitting to two exponentials. By doing so we are are
taking into account possible contributions from the next gap which would
come out as the exponent of the second exponential which is what happened.
This makes the read-off for the first gap better (E in the figure is scaled to
a gap of 2.9693 which is consistent with the low coupling limit of 3). The
difference plot shows interesting oscillation indicating some remnant behavior
in the correlations which we have not studied so far.
acceptance rate, or equivalently the sampling rate, is low and approaches zero. This
however should not be a reason to abandon Monte Carlo approaches since this regime is
precisely the one where perturbative results are accurate.