Most
of the "E" part of the E&P business has, by now, embraced the
main principles of probabilistic risk analysis:
- Reserves distribution
forecasts for prospects.
- Prediction of chance
of completion.
- Probabilistic NPV
forecasts, given success.
- Chance-weighted economic
yardsticks, allowing comparison of projects.
Many companies
are now trying to adapt such probabilistic risk methods to the "P"
part of the E&P business, recognizing its inherent power, which
is further increased because development projects typically involve
more trials (wells) and reduced uncertainties of key geotechnical
parameters, therefore, better predictability.
This is
additionally enhanced because development ventures generally require
relatively large investments in production facilities and infrastructure,
making it even more critical that such projects be assessed appropriately
prior to committing to develop.
There are
two important differences that must be addressed in order to successfully
adapt Exploration risk analysis methodology to the Exploitation
and Production aspects of the upstream oil and gas business.
The first
difference has to do with how engineers and geoscientists deal with
uncertainty. My friend, John Campbell, a highly respected consulting
economist and reservoir engineer, expressed it this way:
"Engineers
are trained to see data as constraining; geoscientists perceive
data as departure-points."
Another
way of characterizing these two philosophies is to recognize that
most scientific and technical education trains us to spend increasing
amounts of time and money to get closer and closer to "The Answer."
This naturally leads to deterministic estimates, false precision
and frequent surprises.
It also
leads to conservative estimates, guided by the outlook that "it's
okay to be wrong on the low-side, but it's not okay to be wrong
on the high side."
Probabilistic
risk analysis has an alternative approach: the Earth is a coarse
filter, so varying degrees of practically irreducible uncertainty
surround most key geotechnical parameters.
Accordingly,
a superior, much more efficient approach is to:
- Spend enough time
and money to capture and express appropriate probabilistic ranges
for the key parameters.
- Use the substantial
time and money saved to generate more prospects.
The most
visible expressions of these different approaches are the two prevailing
estimating conventions:
- Deterministic
— commonly used by engineers in exploitation and production work.
- Probabilistic
— employed by most geoscientists in exploration.
So, a part
of the shift to probabilistic expression of key exploitation and
production parameters lies in finding effective ways to help engineering
folks get comfortable, thinking probabilistically.
Figure
1 is an example
problem. Traditional and conservative decline-curve analysis might
have led to a deterministic forecast of field abandonment in January
2004, at a cumulative future production of 1.513 million barrels.
Probabilistic
methodology, recognizing the inherent uncertainty of outcomes and
the lognormality of remaining reserves, and employing variations
in percentage decline slopes and economic limits, might have predicted
cumulative future production like this:
- Ninety percent confidence
in 1.41 million barrels or more, and abandonment after February
2003.
- Ten percent confidence
in two million barrels or more, and abandonment after October
2006.
- Mean recoverable
reserves of 1.69 million barrels, and abandonment in October 2004.
The second
important issue arises from characteristic dissimilarities between
development and exploration projects, which cause development projects
to have more "key drivers" (i.e., influential parameters) that have
differing relative "weights" in overall project evaluation.
Most of
these development "key drivers" express lower variance (i.e. uncertainty),
shorter investment periods relative to onset of production revenues,
and lower profit to investment ratios. Chance of success is higher
(70 — 95 percent) for development than for exploration projects
(10 — 50 percent), hence chance is less influential in project
evaluation.
In exploration
ventures, key drivers include ultimate recoverable reserves (especially
the productive area forecast), chance of success, producing rates
and cost of finding.
By contrast,
development ventures are characterized by lower variance in recoverable
reserves (and productive area), producing rate, percentage decline
and much lower P/I ratios. Additional parameters take on increased
significance, such as well-head price, development and operating
costs, timing (production onset after investment) and consequences
of political and economic risk.
We generally
see relatively lower variance in production rates and percentage
decline in development projects, simply because more is known about
a discovered petroleum accumulation after delineation than before
discovery. Because onset of production usually occurs closer to
development investment than exploration investment, development
projects are relatively more sensitive to near-term fluctuations
in well-head prices.
Moreover,
because capital investment for development projects is large relative
to the PV of the recoverable reserves, costs also take on increased
importance in project evaluation.
The consequence
of these inherent characteristics is that probabilistic risk analysis
for development projects involves a new set of key drivers, having
different weights than for exploration projects, and having typically
lower variance, and greater sensitivity to capital investment, process
efficiency, well-head price and political or business risk.
Finally,
some development key drivers may not follow the lognormal distribution,
necessitating conventional Monte Carlo simulation and complex decision-trees
involving multiple scenarios.
Bringing
probabilistic risk analysis to the "P" part of the E&P business
will be a service to our investors!
Recommended
Reading: "Bionomics," by Michael Rothschild (Basic Books 1991).
The myriad
tiny interactions and adjustments inherent in a free-market economy
may well be better understood by comparing them with the evolving
myriad complex and interactive biologic sub-communities of the integrated
biosphere. Diverse developing new firms (often "niche businesses"
at first) may be analogous to detached small communities, so important
to "punctuated equilibrium" in evolutionary theory.
A very
thought-provoking and multidisciplinary book.
Read it
— you'll like it!