The Level 3 cross sections can be converted into numbers of events assuming an integrated luminosity and efficiencies. We arbitrarily assume a 50% efficiency and that 50% of the data survives offline cuts.
of
delivered luminosity, assuming an average luminosity of 4E31 (an
average luminosity of 6E31 would reduce this number by about 40%).
The world's database of tagged diffractive dijet events with
is measured in the hundreds and only for low 
jets.
With this huge data sample we will easily be able to make cross
section measurements in many 
and |t| bins and have the
first precision measurements of the pomeron structure from hard
diffraction. If some of the assumptions are pessimistic and the
accepted cross section is too high,
we can always have a prescaled
lower trigger and an unprescaled higher trigger. Also
note that the physics rate will still include some background from
non-pomeron exchanges (such as the 
meson).
One of the goals of the FPD will be to
attempt to understand this issue better.
From Ref. [21], the
inclusive non-diffractive background is estimated to be 
for
very low , about 15% for 
and 
for 
.
The non-diffractive backgrounds to hard diffraction are likely to
be lower than this if the does not produce jets as
copiously as the pomeron does [44].
will be
more than adequate for unambiguous observation of this class of
events, and should allow a detailed study in a few and |t|
bins. It is clear that the number of events, not rates or backgrounds is
the major concern for this physics topic. Efficiencies can be improved
by relaxing the Level 1 cut somewhat. It will also be possible to
combine a track in the 
spectrometer with a rapidity gap on
the proton side to increase the statistics by a factor of a hundred
for some hard double pomeron studies. About
1% of these events would be ``gold-plated''
with tracks in the p spectrometer as well, and could
be used to study the detailed behavior of this process.
, perhaps
1% of which are diffractively produced.
Of these 6,000 events, about 1,500 would have tagged 's
(50% of the events) and 30 would have tagged protons.
A couple hundred of these would be single interaction events.
The Z boson statistics are expected to be about 10 times smaller.
We should also obtain 10's of thousands of tagged
diffractive b and c events.
As discussed earlier, these processes will not require special triggers
or additional bandwidth as they will simply involve reading out the FPD
for all events.
These unique large data samples will allow DØ to revolutionize the field of hard diffraction.