Estimating Mature Plays
51
The number of pools in a play, which is a fi nite number, should
7.
include all small pools that might not be economically viable
at the time of assessment. One might be concerned that if small
pools are included in the assessment, the mean (of the untransformed data) of the pool-size distribution will be substantially reduced. Consequently, the mean would not adequately
describe the economic resources. But from the viewpoints of
exploration and economic analysis, the remaining largest pool
sizes are far more signifi cant than the mean value of the poolsize distribution.
As previously mentioned, the superpopulation concept is used
8.
by PETRIMES. Thus, the predictions made by the system are
of cases that would occur most frequently. A singular case,
for example, is that of the Cardium marine sandstone play
(see Fig. 2.10), in which the largest pool size is about 10 times
larger than the size of the second largest pool. In such a situation, sizes in between the two largest pools may be mistakenly predicted. However, if additional information indicates
that no other sizes would exist, then the information can be
entered into the system as a condition for predicting the individual pool sizes (see the matching process, discussed in the
next section).
The concept of pool-size-by-rank can also be explained using
9.
Monte Carlo simulation. Assume that we have a pool-size
distribution and a number-of-pools distribution. A random
number is generated, and the number of pools, N j (say, 100),
is obtained from the number-of-pools distribution. A total
of N j (= 100 in this case) pool sizes is randomly drawn from
the pool-size distribution (e.g., x 1 , x 2 , … , x 100 ) and is sorted in
descending order. These steps are repeated many times. The
largest size from each simulation trial is then used to construct the size distribution of the largest pool. The sizes for
the second largest, the third largest, and others are similarly
obtained. In practice, the statistical approach for the estimation of individual sizes is more effective and can also provide
various matching options.
The Matching Process: Operation
We will now attempt to use the matching process to estimate individual
pool-size distributions based on point estimates from either LDSCV or
NDSCV. This will assist us in determining N values for immature or
51
The number of pools in a play, which is a fi nite number, should
7.
include all small pools that might not be economically viable
at the time of assessment. One might be concerned that if small
pools are included in the assessment, the mean (of the untransformed data) of the pool-size distribution will be substantially reduced. Consequently, the mean would not adequately
describe the economic resources. But from the viewpoints of
exploration and economic analysis, the remaining largest pool
sizes are far more signifi cant than the mean value of the poolsize distribution.
As previously mentioned, the superpopulation concept is used
8.
by PETRIMES. Thus, the predictions made by the system are
of cases that would occur most frequently. A singular case,
for example, is that of the Cardium marine sandstone play
(see Fig. 2.10), in which the largest pool size is about 10 times
larger than the size of the second largest pool. In such a situation, sizes in between the two largest pools may be mistakenly predicted. However, if additional information indicates
that no other sizes would exist, then the information can be
entered into the system as a condition for predicting the individual pool sizes (see the matching process, discussed in the
next section).
The concept of pool-size-by-rank can also be explained using
9.
Monte Carlo simulation. Assume that we have a pool-size
distribution and a number-of-pools distribution. A random
number is generated, and the number of pools, N j (say, 100),
is obtained from the number-of-pools distribution. A total
of N j (= 100 in this case) pool sizes is randomly drawn from
the pool-size distribution (e.g., x 1 , x 2 , … , x 100 ) and is sorted in
descending order. These steps are repeated many times. The
largest size from each simulation trial is then used to construct the size distribution of the largest pool. The sizes for
the second largest, the third largest, and others are similarly
obtained. In practice, the statistical approach for the estimation of individual sizes is more effective and can also provide
various matching options.
The Matching Process: Operation
We will now attempt to use the matching process to estimate individual
pool-size distributions based on point estimates from either LDSCV or
NDSCV. This will assist us in determining N values for immature or