These P90, P50, and P10 exceedance probabilitiy definitions are correct and rigorous when robust probabilistic methods are used, but otherwise the 1P,2P, and 3P definitions become technically incorrect and somewhat arbitrary, and very confusing.

1P, 2P, and 3P are low ("proved"), median ("proved plus probable"), and high ("proved plus probable plus possible") estimates of reserves may be quantified in probabilistic reservoir modeling (and economics), from uncertainties in the descriptive and control (and economic) variables, as exceedance P90, P50, and P10 (or as cumulative P10, P50, and P90) estimates of reserves, respectively. Commercially-recoverable reserves could be defined in units of BOE, by some simple value function of production and injection, or  by NPV.  The latter is more accurate and appropriate, as the definition of what will be "commercially-recoverable" in the future certainly depends on economics, and is specifically indicated by NPV > 0, but it introduces large uncertainties in  economic model variables.  Estimates "made under existing economic and operating conditions" as defined by the SEC (as required for "proved" reserves) generally have no significant meaning, because those conditions are constantly changing, and our projects last for decades.  The conditions "anticipated to be commercially recovered" and "under existing economic and operating conditions" are absolutely contradictory and nonsensical, since anticipated means "to expect in the future".  ABSOLUTELY NO RESERVES are anticipated to be recovered in the future under existing economic and operating conditions.  The assumption that economic and operating conditions are fixed is completely invalid.  One of the main purposes of simulation is to optimize production by determining the optimal process and boundary (operating) conditions as a function of time (see Goals of Reservoir Simulation).  Optimization is a pre-requisite of prediction.

The SEC rules also state:

"(iii) Estimates of proved reserves do not include the following:

(B) crude oil, natural gas, and natural gas liquids, the recovery of which is subject to reasonable doubt because of uncertainty as to geology, reservoir characteristics, or economic factors; "

All production is subject to reasonable doubt, before it occurs, due to guaranteed high levels of uncertainty in all of those and many more factors.

It is virtually impossible to estimate with any degree of certainty what will be "commercially recoverable" in the future under unknown and unpredictable economic and operating conditions.  For example, (1) a major advance in the efficiency of alternative energy sources could eliminate most of the value of all hydrocarbon reserves at any time.  Regulation may also do so, for example (2) the destruction of the US coal industry by the Clean Power Plan that has reduced the value of US coal reserves to near 0.  Other examples are (3) proposed regulation of CO2 emissions (see SRMS and SPE CCUS Technical Section) that would eliminate true reserves and hugely increase the cost of energy to the public for no reason or benefit that can be substantiated, and (4) regulation following the BP Macondo incident that has virtually eliminated the value of all oil and gas reserves in federal waters, and (5) unknown future and present values of the true rates of dollar devaluation (inflation) resulting in misapplied "cost of capital" and discount factor considerations.  There are an infinite number of such uncertainties that cannot be reliably predicted, unless one has or believes in psychic abilities.  The existence and incompetence of reserves definitions and reporting requirements are certainly indications of that "Crystal Ball Syndrome".

Strictly speaking, according to the SPE/SEC rules, the amounts and value of all defined proved, probable, and possible reserves are and will always be exactly 0.

Beyond any reserves reporting requirements that they may rigorously satisfy, probabilistic performance predictions are obviously required to optimize production and our operations, and to minimize the cost of energy.  As a result we have developed some of the most complex and advanced software systems in the world (using Pipe-it^® workflow integration and iterative optimization software applied to Sensor), which are now capable of automated probabilistic optimization for characterization, upscaling, history matching, forecasting, and predictive optimization - all of the components required to achieve our goal of solving the real-time global predictive optimization problem under uncertainty.  Simulators can predict future production, for given values of all the data, but most of that data is usually uncertain.  So we must employ probabilistic methods to make any predictions or optimizations in reserves estimations or in reservoir modeling, using any model making valid assumptions while rigorously accounting for the uncertainties.

These methods, capabilities, optimizations, and predictions are the basic components of business management and (national) free-market competition in the energy industry  Those companies able to better plan their operations and better optimize reserves and production to minimize risk and maximize efficiency and profit will be more successful and others will fail, thus minimizing the cost of energy and maximizing efficiency, technical advancement, corporate responsibility, sustainable development, and growth of the economy (assuming necessary and proper regulation for a sustainable national economy that is a prerequisite of sustainable development (and of the basic sustainability of all individuals and businesses).

Regardless of how reserves may be defined, probabilistic analysis in reservoir modeling provides for the most rigorous possible forecasts and (more importantly, as mainly used historically) for probabilistic predictive optimizations of reserves and reservoir performance.  It can also potentially eliminate or rewrite, into very simple and much more meaningful engineering/economic and mathematical terms, the definitions, methods, and rules for reserves estimation.