Baseline NSNF plan centered around NEAT:

 Filtered search to z < 0.06. Peak magnitude for s = 1 at this redshift
 is B = 17.8, so NEAT would have the sensitivity to find these SNe about
 one magnitude below peak (assuming improvements in NEAT PSF to compensate
 for filter losses - or maybe even LBL CCD for NEAT).

 Only target fields which will be observable (X > 1.5) at the end of
 twilight 3 months on. X = 1.5 is 3 hrs of RA, and 3 months observability
 is 6 hours of RA, so western most field will be at X = 1.5 in the *east*
 at the end of twilight. Since NEAT trys to observe on the meridian, the
 first 3 hours of each night are not used by NSNF.

 NEAT requires 20s of read out for each 20s exposure, and it takes 3
 exposures of each field. So, the rate is 2 min/field, or 30 fields per
 hour. Assuming an average of 9 hrs/night, (9-3)*30 = 180 fields = 432
 sq. deg would be covered each night.

 NEAT will search over an 18 night span each month, with a nominal
 time gap of 6 days. This gives a maximum search area of 2592 sq. deg.
 per month. This is about 6% of the sky.

 Scaling from our value of 1 pre-max Type Ia per 50 sq. deg out to z = 0.1,
 gives 1 pre-max Type Ia per 230 sq. degree out to z = 0.06. This translates
 to 11 pre-max Type Ia SNe with z < 0.06 per month. It is probably fair to
 estimate that there will be a similar number of SNe Ia with z > 0.06 and
 SNe II, giving more than 30 candidates per month.

 Note that if fields were observed in the same order the next month,
 there would be a 19 day gap - still useable for searching, but
 presenting more old SNe. Alternatively, the nightly order of fields to
 be mirrored, or shifted. Mirroring would give gaps from 10 days to 28
 days. Shifting would give gaps of (19 - shift) days for (6 - shift)
 nights. For instance, a two week gap would require a 5 day shift,
 allowing 4 more nights of searching per month (with the longer gap).

 In order to track the meridian, the starting point would shift by
 2 hrs each month, so such longer baseline searches would have less
 area by a factor of (9-3-2)/(9-3) = 2/3.

 In summary, at 100% capacity NEAT would provide at least 11 pre-max
 SNe per month, and there would be at least 30 candidates per month to
 screen. Cross-month searching would add about 30% more good SNe Ia 
 (giving a total of 22), and probably 50% more candidates (giving a
 total of about 50).

 Now, NEAT will not work at 100% capacity. One could argue that a lost
 reference night translates into a lost search night. If the duty
 cycle is 50% due to weather, then there would be between 5 and 11
 good SNe Ia, and 15 to 30 candidates to examine each month.