Ed,
Nice, clear, concise linear statistical break-down and very useful as a platform from which to base one's forecast.

The only thing I would build upon that platform is consideration for the Atlantic Multidecadal Oscillation, and the Quasi-Biennial Oscillation phase states, which both have an apparent "split" correlation with the frequency of intense Atlantic Basin hurricane.

This may be important to some individuals looking at this information more discretely.

1) During warm phases of the AMO, the number of major hurricanes is significantly greater than during cool phases. The AMO flipped into its warm phase ~ in the mid 1990s, and since we have observed major hurricanes (111mph+) occurring with greater frequency. Thus, there could conceivably arise a small issue in the construction of the aggregate time series as they pertain to the analog years in question:
1971 (13/6/1), 1985 (11/7/3), 1950 (13/11/8), 1974 (11/4/2) and 1999 (12/8/5); relative to the current warmer AMO phase state. The reason for this is because during the period of the 1970s through ~ the mid 1990s, the cool phase of the AMO was in place; therefore, a potential error in interpretation in less than an a-priori analysis, exists.

A potential counter-point to the above paragraph, however, is the very fact that you mentioned; there has been an apparent recent cooling of the Atlantic tropical SSTs - showing rather pervasive, albeit weak, negative anomalies stretching between 80 and 45W by ~ 30N. We shall have to see if these modest cool signals withstand the higher solar irradiance interval of April 15 - June 1. My hunch is they do not and we see a bit of progression into the +AMO resume, which is heavily statistically favored in the background.

2) The QBO is a strong motivator on the variations in the amount of vertical wind shear between the upper troposphere and the lower stratosphere over the hurricane genesis regions of the Atlantic Basin. Moreover, the QBO has a very regular oscillatory periodicity, such that successful long range extrapolations of the mean stratospheric zonal wind anomalies can be made with greater than average accuracy for atmospheric phenomenon; and because there is a downward (loss of altitude) propagation of the stream, it begins to effect positively or negatively upon the upper tropospheric critical levels for those QBO events that time accordingly.

A brief primer for those who have not yet learned:
An east QBO phase (typically lasting 12 to 15 months) forces strong easterly winds in the lower stratosphere/upper troposphere over the TC genesis regions of the central Atlantic and Caribbean; this is a positive shear domain. The opposite is true for west QBO phase states. Weak easterly winds and negative shear domains pervade the TC genesis regions.

C/O Dr. William Gray, et al; Colorado State Univeristy
"Years of west QBO phases typically have 50% more named storms, 60% more hurricanes, and 200% more intense hurricanes than occur in east QBO phases within the Atlantic basin."

Here are the measured QBO values - per monthly - since 1990:
1990 -8.71 -6.74 0.95 +5.72 11.46 12.90 12.54 12.63 13.21 12.39 11.55 10.68
1991 8.71 8.44 9.00 10.69 3.82 -3.34 -10.36 -14.69 -13.96 -12.21 -12.27 -12.68
1992 -13.96 -14.33 -16.84 -17.79 -15.96 -15.34 -12.05 -4.75 1.30 3.94 6.33 8.19
1993 9.63 10.81 11.36 12.60 13.56 6.55 0.38 -1.20 -1.14 -4.04 -5.76 -6.00
1994 -7.64 -9.84 -11.29 -14.71 -18.03 -23.09 -28.65 -27.02 -19.07 -10.29 -0.30 5.93
1995 8.38 8.01 8.79 11.79 14.92 15.62 11.74 9.53 6.98 3.43 -0.77 -4.57
1996 -5.79 -6.90 -9.92 -11.08 -14.88 -17.03 -23.93 -25.85 -26.02 -23.40 -18.08 -9.86
1997 -3.57 1.94 4.77 9.74 12.37 14.50 14.85 11.69 11.64 9.91 5.74 0.78
1998 -0.85 -2.96 -4.92 -7.82 -14.08 -18.57 -22.97 -24.70 -22.12 -18.77 -12.22 -3.96
1999 3.09 5.84 8.59 13.51 15.56 15.23 14.11 11.91 11.18 10.62 6.01 6.43
2000 4.85 4.20 5.51 3.98 -0.99 -7.83 -13.13 -15.31 -15.52 -14.04 -15.07 -14.56
2001 -15.69 -15.53 -15.99 -17.73 -20.99 -23.31 -24.45 -21.67 -14.29 -10.81 -3.88 1.48
2002 4.64 8.00 9.32 14.03 14.16 13.26 10.05 10.60 8.90 7.66 4.46 -0.50
2003 -1.39 -1.44 -3.30 -8.57 -13.94 -18.01 -22.99 -24.64 -22.51 -20.34 -17.86 -11.38
2004 -4.84 2.61 5.45 10.46 12.97 11.75 9.96 8.82 7.22 7.84 4.41 2.27
2005 -0.45 -0.88 0.06 -6.64 -15.09 -20.59 -24.20 -25.87 -27.80 -28.76 -29.55 -25.04
2006 -18.83 -11.24 -0.38 5.00 10.36 11.47 10.75 9.10 10.20 10.86 10.10 6.21
2007 2.61 2.43 1.24 -5.18 -14.07 -21.34 -24.93 -27.41 -28.14 -29.05 -27.61 -19.48
2008 -12.43 -4.70

This elucidates rather nicely the very regular periodicity that exist between the negative and positive phases of the QBO. This allows for a higher degree of extended lead predictability of the QBO, based on past performance. What is interesting about this is that January and February are indicating a rising QBO modality, which said predictive measure strongly argues for a continuance and probable rise into positive phase states this March/April...and onward into summer.

*The QBO method may be more important for Cape Verdi systems. My own findings seem to suggest that activity in the Bahamas-Gulf-western Caribbean triangulate seems to have a separate statistical correlation to the QBO than intercontinental trafficking region. Why this is - a case for further study.

Conclusion:
The inclusion of the QBO factor, combined with the positively differentiated AMO, may actually argue for an increased frequency than the linear method would suggest. How much is a fantastically uncertain question, but my prediction is that we will see perhaps ~4 more per category than those predictions you leveled; which isn't too drastically different. 17/12/7 (2 category 5 storms).

There is no statistical derivation from the Sahal soil moisture in this analysis.