SPE 110191

The Cost of Errors in Estimates Used in Concept Selection

C. Jablonowski, SPE, and C. Wiboonkij-Arphakul, SPE, The University of Texas at Austin, and M. Neuhold, SPE, Clough
Engineering

Copyright 2007, Society of Petroleum Engineers

This paper was prepared for presentation at the 2007 SPE Annual Technical Conference and
Exhibition held in Anaheim, California, U.S.A., 1114 November 2007.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, Texas 75083-3836 U.S.A., fax 01-972-952-9435.

Abstract
This study examines the loss in project value incurred when
concept selection decisions are based on erred estimates of
input variables. Estimates of the magnitude of such losses are
provided, along with an analysis of which input estimates
matter most in determining value loss. A procedure for
concept selection is defined to model the decision making
process and is used in conjunction with a simplified asset
development optimization model to estimate project values.
The analysis compares project values resulting from concept
selection decisions based on erred estimates and decisions
based on an alternate hypothesis. Results suggest that cost of
erred estimates for initial costs, expansion costs, and the
timing of future expansion projects are comparable in
magnitude to the cost of erred reserve estimates. Also, the cost
of underestimating expected reserve volume tends to be larger
than the cost of overestimating reserve volume; aggressive
cost estimates are more destructive to value than conservative
estimates; and conservative schedule estimates for the timing
of expansion projects are generally more destructive to value
than aggressive schedule estimates.
Introduction
During the concept comparison and selection phase of
exploration and production (E&P) capital projects, decision
makers estimate the value of competing development
concepts. These estimates are used to rank options and to
select one option to carry forward to the next project phase.
The importance of these estimates cannot be overstated, they
determine which concept is selected, and have a strong
influence on field architecture, initial capacity of facilities,
well counts, production rates, and project schedule. Decisions
in concept selection have a large impact on the value
ultimately derived from the asset.1,2 These estimates are also
used for other important analyses and decisions during
concept selection such as value of information (VOI) analysis.

A variety of input variables are required to estimate the

value of competing development concepts. These input
variables include estimates for the subsurface (e.g. reserves,
flow rates, decline rates), estimates for the surface facilities
(e.g. CAPEX, OPEX, schedule, reliability), and estimates for
exogenous factors such as commodity price. The true values
of these input variables are almost always unknown, and
estimates are developed based on the current information set
available to the decision maker.
The objective of this study is to examine and compare the
loss in value incurred when concept selection decisions are
based on erred estimates of input variables. Errors can occur in
estimates of expected values and in estimates of variance. The
conclusion that erred estimates of input variables destroy
project value can be made using common sense. What this
study attempts to provide are original estimates of the
potential magnitude of such losses, and an analysis of which
input estimates matter more or less than others.
In practice, one does not know if a current estimate for an
input variable is erred, but one can estimate the impact of an
alternate hypothesis being true, and this is the framework
adopted here. A procedure for concept selection is defined to
model the decision making process and is used in conjunction
with a simplified asset development optimization model to
estimate project values. The analysis compares project values
resulting from concept selection decisions based on erred
estimates and decisions based on an alternate hypothesis; in
both cases, the alternate hypothesis is taken to be true. The
difference in value observed, if any, is caused by sub-optimal
initial facility capacity (note, the difference in value can also
be interpreted as the maximum willingness to pay to confirm
the alternate hypothesis). The approach is similar in form to
standard VOI analyses.3-6
The number of input variables normally required to
estimate the value of competing development concepts is
immense, and estimates are required from all project
disciplines. This study examines the loss in value associated
with errors in estimates of three key input variables: (i)
reserves, (ii) facility initial cost, and (iii) facility expansion
cost and schedule. These three variables have a large impact
on project value. The reserves estimate drives most major
development decisions, including the depletion plan, well
counts, and facility design. Facility cost estimates influence
concept type, initial facility capacity, and plans for future
expansion. Estimates for the cost and timing of future facility
expansion affect the initial facility capacity decision and the
value that can be captured given upside realizations of

SPE 110191

subsurface scenarios. The results of this study provide insight

on the magnitude of losses associated with erred estimates for
these input variables, and enable a relative ranking of these
inputs based on the impact on project value.
Estimating the Cost of Errors
In this section, a procedure for concept selection is defined to
model the decision making process, and a simplified asset
development optimization model is specified. Using an
integrated asset development optimization model to assess the
impacts of uncertainty on decisions and project value is widely
viewed as a best practice,7-16 and thus is the method adopted
here.
It is assumed that a steel piled jacket is already selected as
the preferred concept. Therefore, the decision process and
model are designed to inform selection of the initial facility
capacity, a decision typically made during the latter part of the
concept selection phase. (Note, this definition of the problem
is by choice, and there is no restriction on defining a problem
where the fundamental concept selection decision is open, e.g.
subsea or local host, spar or TLP, etc.)
The analysis focuses on the loss in value associated with
errors in estimates of three key input variables: (i) reserves,
(ii) facility initial cost, and (iii) facility expansion cost and
schedule. The specific cases of analysis and associated
assumptions are provided in the description of results.
The Asset Development Optimization Model. In this
section, a simplified asset development optimization model is
defined. The model is used to estimate project value given
estimates for, and realizations of, the input variables of
interest. Where possible, equations are specified based on
empirical research17; where no data is available, the equations
are specified based on anecdotal evidence. The following
simplifying assumptions are made:

The asset is comprised of one oil reservoir;

associated gas is negligible and is not considered in
the optimization
The asset is located in shallow water and a single
steel piled jacket is the selected facility concept; the
facility operates at 100% capacity
All wells are drilled from the jacket
All cash flows occur at the end of the year.

Objective Function. The optimization model assumes a

risk neutral decision maker who maximizes the net present
value, z, of project cash flows defined as follows:

1
z = ( p (t ) * q (t ) opex(t ) capex(t ) ) *
+
(1
r )t
t

(1)

where: z = net present value, $; p(t) = average oil price during

time t, $/bbl; q(t) = total production during time t, bbls; opex(t)
= operating expense during time t, $; capex(t) = capital
expenditure during time t, $; r = discount rate. The oil price is
held constant at $60/bbl, and the discount factor is held
constant at 15%.

Cost Functions. Costs are comprised of operating expense

and capital investment in initial facilities, facility expansion,
and wells. Recall that facility cost is one of the input estimates
that is evaluated for its impact on decisions and project value.
Cases for analysis are defined in the Results.
Operating Expense (OPEX). OPEX represents the variable
cost of production and is a function of production volume. It is
defined as follows:
opex(t ) = vcprod * q(t )

(2)

where: vcprod = variable cost, $/bbl. Variable cost is held

constant at $2/bbl.
Capital Expenditure (CAPEX). CAPEX is composed of
three elements as follows:
capex(t ) = cwell (t ) + cplat (t ) + cexpa(t )

(3)

where: cwell(t) = total well cost during time t, $; cplat(t) =

facility initial cost during time t, $; cexpa(t) = facility
expansion cost during time t, $.
Individual well cost, wellcost(i), includes both drilling and
completion cost. Well cost can vary with the production
capacity of the well. The cost (in $mm) of an individual well i
at any time t is defined as follows:

wellcost (i ) = fcwell + vcwell * qwellmax (i )

(4)

where: fcwell and vcwell = cost parameters; qwellmax(i) =

maximum production capacity, mbopy. Uncertainty in well
cost was not investigated in this study, and in all cases, fcwell
and vcwell are held constant at 5 and 0.005, respectively.
While qwellmax(i) would normally be defined as a control
variable, it is also held constant at 730 mbopy, making
individual well cost a constant at $8.65 million. To produce
over 730 mbopy, additional wells are required. Total well cost
at time t is defined as follows:

cwell (t ) = wellcost (i ) * D (t , i )

(5)

where D(t,i) is a binary variable indicating whether or not well

i is drilled at time t. the well The total number of wells,
wellcount, is a control variable.
Facility initial cost, cplat(t), includes all costs related to
facility design, fabrication and installation. Initial cost (in
$mm) is modeled as a function of initial capacity as follows:

cap0 (t )
cplat (t ) = b0 + b1

365

eplat

(6)

where: b0, b1, eplat = cost parameters; cap0 = initial facility

capacity, mbopy.
Facility expansion cost, cexpa(t), includes all costs related
to capacity expansion and is defined as follows:

SPE 110191

eplat

capadd (t )
cexpa (t ) = m b0 + b1

365

(7)

where: m = cost multiplier; capadd = incremental facility

capacity, mbopy. Observe that the functional form and
parameters are identical to the initial cost function, save for
the cost multiplier which simulates the relatively higher cost
adding capacity via expansion; m is held constant at 1.5. There
are no constraints on capadd(t).
Other Constraints. A variety of constraints are required in
the model. Total production is constrained by the volume of
recoverable reserves as follows:

q(t ) rescap

(8)

where: rescap = recoverable reserves, bbls.

Production at any time t is constrained by the joint
capacity of the individual wells as follows:
q(t ) qwellmax(i) * P(t , i )

(9)

where: P(t,i) is a binary variable indicating whether or not

well i the has been drilled at or prior to time t and is available
for production.
The ultimate recovery per well is primarily a function of
reservoir properties. For modeling purposes, the minimum
number of wells required over the life of the project can be
defined as follows:

wellcount

q(t )

(10)

wellvol

where: wellcount = total number of wells drilled; wellvol =

maximum ultimate recovery, bbls/well. Wellvol is specified as
a random variable and is described further in the discussion of
results.
The daily production rate cannot exceed the current
production capacity of the facility. This constraint is defined
as follows:

A Procedure for Concept Selection and Estimating the

Cost of Errors. In this section, the procedure for estimating
the loss in value incurred when concept selection decisions are
based on erred estimates of input variables is described. The
procedure models a decision making process that would occur
in practice.
The first step simulates the decision makers selection of
initial facility capacity given the current information set and
assumptions. In practice, a decision maker typically evaluates
the NPV for a finite number of cases due to time and resource
constraints. Each case represents a deterministic combination
of initial facility capacity, subsurface realizations, cost and
schedule, and the potential exercise of real options. In simple
cases, this process will be sufficient to dictate the valuemaximizing course of action.
When time and resources permit, or for large complex
projects, the decision maker often defines input variables as
random variables by specifying probability density functions
(pdfs), () , and using simulation techniques to model
project value. For example, if the current information set is
given by , the pdf for a random variable such as reserves,
rescap,
may
be
specified
as
follows:
.
Other
key
input
variables
(rescap | ) ~ N ( rescap , rescap )
can be specified in a similar manner. This probabilistic
approach yields a more complete view of potential project
outcomes and leads to better value capture, on average, as
compared to the finite case deterministic approach.
Step 1. In the first step, the asset optimization model is used to
simulate a probabilistic analysis as it would be performed by
the decision maker to determine an initial facility capacity.
The decision maker specifies pdfs for the key input variables
of interest (here, reserves, cost parameters, and expansion
timing), and the model maximizes NPV given an initial
facility capacity, and realizations of the key input variables. In
practice, realizations for some variables, e.g. reserves, tend to
occur over several years. In this study, it is assumed that the
realizations are observed immediately after initial production.
The result obtained from each run is the maximum NPV
attainable given the initial facility capacity constraint and the
set of realizations. If this process is repeated n times, a pdf of
NPV is generated at each increment of initial facility capacity.
An example of the output at a specific increment of capacity is
depicted in Figure 1.

q (t ) cap (t )
t

cap (t ) = cap 0 (t ) + cap add (t )

(11, 12)

Cumulative Distribution Function of NPV

@ initial platform capacity = 10,000 Mbopy
1.0
0.9
0.8
0.7

where: cap(t) = facility capacity at time t, bbls.

Additional equations are required for flow control, other
programming issues, miscellaneous constraints, and to control
for constraints of specific cases, but these are routine and not
itemized here. In summary, the model is defined with four
endogenous variables: q(t), cap0(t), and capadd(t), and
wellcount (and associated binary variables). The model is
written in the General Algebraic Modeling System (GAMS).

P10 = 635.88
P50 = 797.60
P90 = 945.01
Expected Value = 794.55

P90

0.6
0.5

Expected Value

P50

0.4
0.3
0.2
0.1
0.0
300

P10
400

500

600

700

800

900

1000

1100

1200

NPV, $MM

Figure 1: Distribution of NPV @ Initial Facility Capacity = 10,000

Mbopy

SPE 110191

This process is repeated for each increment of initial facility

capacity and the optimal initial capacity, FAC0, is observed as
depicted in Figure 2. It is assumed that the decision maker
would select this optimum.
Optimal Initial Platform Capacity
850

NPV, $MM

FAC 0 = 11000 Mbopy

800

750

700
4000

6000

8000

10000

12000

14000

16000

Initial Platform Capacity, Mbopy

Figure 2: Comparison of Expected NPV Obtained from Various

Initial Facility Capacities

Step 2. The second step simulates the case where the

project is initiated using FAC0 (as determined in Step 1), but
the key input variables are realized according to alternate
hypotheses. The model maximizes NPV given FAC0 and
realizations on the key input variables. If this process is
repeated n times, a pdf of NPV is generated. The expected
NPV determined by this process is the value that would be
captured if the alternate hypotheses are true. The expected
NPV determined by this process is denoted as NPVB.
Step 3. The third step repeats Step 1, except that the
alternate hypotheses are used to specify the pdfs for the key
input variables and to select an initial facility capacity. The
initial facility capacity determined by this process is denoted
as FAC0*, and the expected NPV is denoted as NPVC.
The cost of using estimates based on the current
information set when the alternate hypotheses are true is given
by NPVB - NPVC. This difference in value represents the suboptimality of initial facility capacity. In a VOI context, this
figure also represents the maximum willingness to pay to
confirm the alternate hypothesis.
Results
This study examines the loss in value associated with errors in
estimates of three key input variables: (i) reserves, (ii) facility
initial cost, and (iii) facility expansion cost and schedule.
These cases are investigated separately, that is, when
evaluating reserve cases, the estimates for costs and schedule
are fixed (either as constants or as random variables with
constant mean and variance), and vice versa.
Reserves. Three cases were developed to examine the loss in
project value when erred estimates of reserves are used in
concept selection. For all cases, reserves are normally
distributed, and parameters are as defined in the model
description except for the following:

The cost parameters are held constant at b0=20,

b1=0.001, and eplat=2.5; these apply to initial and
expansion costs
The expansion date is fixed and occurs in year 3

Wellvol is normally distributed with mean and

standard deviation as follows (mean, standard
deviation): Reserve Case A: (5,1); Reserve Case B:
(20,2); Reserve Case A: (40,8).

The results are best explained via a specific entry in Table 1.

For example, if the decision makers estimate for reserves,
based on the current information set, is defined as a normally
distributed random variable with a mean of 10 mmbbls and
standard deviation of 1 mmbbls (the upper left cell of Table
1), and there is an alternate hypothesis that the reserves are
normally distributed with a mean of 20 mmbbls and standard
deviation of 3 mmbbls (the center cell of Table 1; the center
cell always represents the alternate hypothesis), then the cost
of making decisions based on the current estimate given the
alternate hypothesis is true is computed to be -4.27% of
expected project value (entries are computed as (NPVBNPVC)/NPVC using the three step procedure defined above).
All of the off-center entries are interpreted in the same
manner. Two additional cases representing larger reserve
endowments were also examined; these results are reported in
Tables 2 and 3.
Table 1. Reserve Case A: Cost of Erred Estimates for
Reserves (percent of project value)
Mean (mmbbls)
Std. Dev.
(mmbbls)
1

10

20

30

-4.27

0.00

-1.03

-4.27

0.00

-1.03

-6.68

0.00

-1.03

Table 2. Reserve Case B: Cost of Erred Estimates for

Reserves (percent of project value)
Mean (mmbbls)
Std. Dev.
(mmbbls)
4

40

80

-4.46

-0.13

-0.85

12

-4.46

0.00

-0.85

20

-1.92

-0.20

0.00

120

It is observed in Tables 1 and 2 that the cost of

underestimating reserves is considerably larger than the cost
of overestimating reserves. This suggests that the cost of
unexpected expansion is greater than the cost of unused
capacity, on average. Of course this result is a complex
function of many variables, namely the expansion cost
multiplier, m. This result is not as strong in the larger Reserve
Case C shown in Table 3.
Table 3. Reserve Case C: Cost of Erred Estimates for
Reserves (percent of project value)
Mean (mmbbls)
Std. Dev.
(mmbbls)
8

80

160

-3.24

-0.32

-3.64

24

-5.31

0.00

-3.64

40

-5.31

-0.25

-2.79

240

SPE 110191

Facility Cost. Three cases were developed to examine the loss

in project value when erred estimates of the cost parameters
(b0, b1, and eplat) are used in concept selection. For all cases,
the cost parameters are normally distributed as defined in
Tables 4-6, and parameters are as defined in the model
description except for the following:

It is assumed that rescap is normally distributed with

a mean of 80 mmbbls and a standard deviation of 12
mmbbls
Wellvol is normally distributed with a mean of 20
mmbbls and a standard deviation of 2 mmbbls
The expansion year is fixed at year 3.

The results are presented in an identical format as the

previous results for reserves. In Table 4 for example, if the
decision makers estimate for b0, based on the current
information set, is defined as a normally distributed random
variable with a mean of -10 and standard deviation of 2.5 (the
upper left cell of Table 4), and there is an alternate hypothesis
that the parameter is normally distributed with a mean of 20
and standard deviation of 5 (the center cell of Table 4; again,
the center cell is fixed as the alternate hypothesis), then the
cost of making decisions based on the current estimate given
the alternate hypothesis is true is computed to be -1.37% of
expected project value. The results for the analysis of b1 and
eplat are reported in Tables 5 and 6. In all three cases, the
values of the two parameters not under examination are
specified per the alternate hypothesis for those parameters.
Table 4. Cost Case A: Cost of Erred Estimates for b0
(percent of project value)
Mean (mmbbls)
Std. Dev.
(mmbbls)
2.5

-10

20

-1.37

-0.19

-0.20

5.0

-2.86

0.00

-1.36

10.0

-0.36

-0.58

-1.40

30

Table 5. Cost Case B: Cost of Erred Estimates for b1

(percent of project value)
Mean (mmbbls)
Std. Dev.
(mmbbls)
0.000175

0.00075

0.001

0.002
-0.36

-1.80

-0.24

0.000200

-0.76

0.00

-0.40

0.000250

-1.23

-2.36

-0.52

Table 6. Cost Case C: Cost of Erred Estimates for

eplat (percent of project value)
Mean (mmbbls)
Std. Dev.
(mmbbls)
0.05

2.0

2.5

-11.48

-2.21

-0.76

0.10

-4.04

0.00

-0.39

0.20

-6.59

-0.35

-1.11

2.7

These results suggest that aggressive cost estimates are

more destructive to value than conservative estimates. When
estimates are aggressive, initial facility capacity is larger on
average based on the assumption that capacity is cheap.
Conservative estimates also lead to value loss. An interesting
observation is that conservative estimates also lead to larger
initial facility capacities. The reason for this is that when costs
are assumed to be higher, there is a disincentive to invest in
initial capacity, but there is a stronger incentive to avoid more
costly expansion in the future. The net result is increased
investment in initial capacity. This same structure is observed
in the results for all three parametersunless the cost estimate
is close to perfect, the initial facility will be too large on
average.
Expansion Timing. Three cases were developed to examine
the loss in project value when erred estimates for the year of
expansion are used in concept selection. For all cases,
parameters are as defined in the model description except for
the following:

It is assumed that rescap is normally distributed with

a mean of 80 mmbbls and a standard deviation of 12
mmbbls
Wellvol is normally distributed with a mean of 20
mmbbls and a standard deviation of 2 mmbbls
The cost parameters are normally distributed with
mean and standard deviation as follows (mean,
standard deviation): b0: (20,5); b1: (0.001, 0.0002);
and eplat: (2.5, 0.1); these values apply to initial and
expansion costs.

Again, the results are best explained via a specific entry in

Table 7. For example, if the decision makers estimate for an
expansion project, based on the current information set, is
defined to occur in year 5, and there is an alternate hypothesis
that the expansion will occur in year 2, then the cost of making
decisions based on the current estimate given the alternate
hypothesis is true is computed to be -4.29% of expected
project value. Additional cases representing alternate
hypotheses were also examined; these results are reported in
Tables 8-10 and are interpreted in the same manner.
The results in Table 7 indicate that conservative schedules
are quite destructive to project value. The reason for this is
that initial facility capacity during concept selection is affected
by the assumption of when expansion will occur. If expansion
is assumed to occur later, say in year 7, and the expansion
actually were to occur in Year 2, the initial capacity decision
is observed to be sub-optimal (a case of over-investment in
initial capacity).
Table 7. Expansion Case A: Cost of Erred Estimates
for Expansion Year (percent of project value)Alternate Hypothesis = Year 2
Expansion Year
Cost of
Erred
Estimate

0.00

-4.17

-4.29

-4.53

SPE 110191

Table 8. Expansion Case B: Cost of Erred Estimates

for Expansion Year (percent of project value)Alternate Hypothesis = Year 3

Expansion Year
Cost of
Erred
Estimate

-1.31

0.00

-1.08

-1.30

Table 9. Expansion Case C: Cost of Erred Estimates

for Expansion Year (percent of project value)Alternate Hypothesis = Year 5
Expansion Year
Cost of
Erred
Estimate

-2.36

-2.00

0.00

-2.48

Table 10. Expansion Case D: Cost of Erred

Estimates for Expansion Year (percent of project
value)-Alternate Hypothesis = Year 7
Expansion Year
Cost of
Erred
Estimate

-2.30

-2.71

-2.48

0.00

Conclusions
The objective of this study is to examine and compare the loss
in value incurred when concept selection decisions are based
on erred estimates of input variables. Using a generic
procedure for concept selection decisions in conjunction with
a simplified asset development optimization model reveals
several strategic insights. Errors in estimates cause suboptimal initial facility capacity decisions, resulting in overinvestment, or under-investment coupled with costly future
expansion. Specifically, the results of this study suggest the
following:

Acknowledgements
This study was partially supported by the Cockrell School of
Engineering and the Jackson School of Geosciences at the
University of Texas at Austin. The authors thank Tim Taylor
and Bill Lamport for comments and suggestions. All errors
and omissions remain the responsibility of the authors.
Nomenclature
(mm) (m) bopy = (million) (thousand) barrels of oil per year
$mm = million dollars
References

Table 10 depicts a common situation in industry. In these

cases, the decision makers estimates for an expansion project,
based on the current information set, are aggressive relative to
the alternate hypothesis. In these cases, if the expansion does
not occur until year 7, and planning is done based on the
assumption that expansion will occur in an earlier year, the
initial facility capacity is again sub-optimal (a case of underinvestment in initial capacity). A comparison of Tables 7 and
10 suggest that conservative schedule estimates for the timing
of capacity expansion are more destructive to project value
than aggressive schedule estimates.

Aggressive cost estimates are more destructive to

project value than conservative estimates
Erred cost estimates, whether aggressive or
conservative, create an incentive to over-invest in
initial facility capacity
Conservative schedule estimates for the timing of
capacity expansion are generally more destructive to
project value than aggressive schedule estimates
Generally, the cost of erred estimates is more
sensitive to errors in estimates of means than to errors
in estimates of variances.

1.

2.

3.

4.

5.

6.

7.

The cost of erred estimates for initial costs, expansion

costs, and the timing of future expansion projects are
comparable in order of magnitude to the cost of erred
reserve estimates
The cost of underestimating expected reserve volume
tends to be larger than the cost of overestimating
reserve volume

8.

Walkup, G. and Ligon, J.: The Good, the Bad, and the
Ugly of the Stage-Gate Project Management Process in the
Oil and Gas Industry, paper SPE 102926 presented at the
2006 SPE Annual Technical Conference and Exhibition,
San Antonio, TX, September 24-27.
Evans, S.: A Modern Approach to E&P Asset Valuation,
Development, and Decision Making, JPT (July 2005) p.
36-38.
Coopersmith, E., Burkholder, M., and Schulze, J.: Valueof-Information Lookbacks Was the Information You
Gathered Really Worth Getting? paper SPE 101540
presented at the 2006 SPE Annual Technical Conference
and Exhibition, San Antonio, Texas, September 24-27.
Bickel, E., Gibson, R., McVay, D., Pickering, S., and
Waggoner, J.: Quantifying 3D Land Seismic Reliability
and Value, paper SPE 102340 presented at the 2006 SPE
Annual Technical Conference and Exhibition, San
Antonio, TX, September 24-27.
Gilbert, R., Lake, L., Jablonowski, C., Jennings, J., and
Nunez, E.: A Procedure for Assessing the Value of
Oilfield Sensors, paper SPE 109628 presented at the 2007
SPE Digital Energy Conference and Exhibition, Houston,
TX, April 11-12.
Prange, M., Armstrong, M., Bailey, W., Couet, B.,
Djikpesse, H., Wilkinson, D., and Galli, A. Better
Valuation of Future Information Under Uncertainty, paper
SPE 103028 presented at the 2006 SPE Annual Technical
Conference and Exhibition, San Antonio, TX, September
24-27.
Narayanan, K., Cullick, A., and Bennett, M.: Better Field
Development
Decisions
from
Multi-Scenario,
Interdependent Reservoir, Well, and Facility Simulations,
paper SPE 79703 presented at the 2003 SPE Reservoir
Simulation Symposium, Houston, TX, February 3-5.
Cullick, A., Narayanan, K., and Gorell, S.: Optimal Field
Development Planning of Well Locations With Reservoir
Uncertainty, paper SPE 96986 presented at the 2005 SPE
Annual Technical Conference and Exhibition, Dallas, TX,
October 9-12.

SPE 110191

9.

10.

11.

12.

13.

14.

15.

16.

17.

Bilderbeck, M., and Beck, I.: Using Integrated Project

Models to Evaluate Field Development Options, paper
SPE 93554 presented at the 2005 SPE Middle East Oil &
Gas Showand Conference, Bahrain, March 12-15.
Laughton, D., Joe, G., Paduana, M., and Samis, M.:
Complete Decision-Tree Analysis Using Simulation
Methods: Illustrated With an Example of Bitumen
Production in Alberta Using Steam Injection, paper SPE
103178 presented at the 2006 SPE Annual Technical
Conference and Exhibition, San Antonio, TX, September
24-27.
Guyaguler, B., and Ghorayeb, K.: Integrated Optimization
of Field Development, Planning, and Operation, paper
SPE 102557 presented at the 2006 SPE Annual Technical
Conference and Exhibition, San Antonio, TX, September
24-27.
van Essen, G., Zandvliet, M., Van den Hof, P., Bosgra, O.,
and Jansen, J.: Robust Waterflooding Optimization of
Multiple Geological Scenarios, paper SPE 102913
presented at the 2006 SPE Annual Technical Conference
and Exhibition, San Antonio, TX, September 24-27.
Carreras, P., Johnson, S., and Turner, S.: Tahiti Field:
Assessment of Uncertainty in a Deepwater Reservoir Using
Design of Experiments, paper SPE 102988 presented at
the 2006 SPE Annual Technical Conference and
Exhibition, San Antonio, TX, September 24-27.
Camargo de Abrea, C., Costa Lima, G., and Suslick, S.:
The Timing of Investment in Deepwater Heavy-Oil
Opportunities With Uncertainty in Oil Price and Production
Technology, paper SPE 102891 presented at the 2006 SPE
Annual Technical Conference and Exhibition, San
Antonio, TX, September 24-27.
Begg, S., Bratvold, R., and Campbell, J.: Improving
Investment Decisions Using a Stochastic Integrated Asset
Model, paper SPE 71414 presented at the 2001 SPE
Annual Technical Conference and Exhibition, New
Orleans, Louisiana, September 30-October 3.
Woodhead, T.: Development of a Generic Life Cycle
Modeling Tool, paper SPE 102948 presented at the 2006
SPE Annual Technical Conference and Exhibition, San
Antonio, TX, September 24-27.
Karlik, C.: Parametric Estimating of Oil and Gas
Production Facilities Capital Costs Worldwide, paper SPE
22015 presented at the 1991 SPE Hydrocarbon Economics
and Evaluation Symposium, Dallas, TX, April 11-12.