Net Values of Our Information (includes associated papers 18563 and 18580 )
- John Lohrenz (consultant)
- Document ID
- Society of Petroleum Engineers
- Journal of Petroleum Technology
- Publication Date
- April 1988
- Document Type
- Journal Paper
- 499 - 503
- 1988. Society of Petroleum Engineers
- 6.1.5 Human Resources, Competence and Training, 5.7 Reserves Evaluation, 1.6 Drilling Operations, 5.4.1 Waterflooding, 4.2 Pipelines, Flowlines and Risers, 4.1.2 Separation and Treating, 5.4.9 Miscible Methods, 7.1.9 Project Economic Analysis, 5.7.5 Economic Evaluations
- 0 in the last 30 days
- 203 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 12.00|
|SPE Non-Member Price:||USD 35.00|
Summary. Petroleum engineers' basic product is information. We are continuously deciding what information to produce. This paper discusses how to determine net values of information, recognizing that uncertainties are always present. Bias toward information that is "good news" arises when uncertainties are not considered. Surprisingly, some information can have higher net value when oil prices are lower.
Petroleum engineers are producers primarily of information, and they flourish according to the value of the information they produce. In a very real sense, the most important decisions petroleum engineers make are choices of which information to produce. These decisions are unavoidable. The thought and consideration given in making these decisions ranges from none to turmoiled anguish. The number of decisions is unarguably huge.
A method to guide these decisions on the basis of net value of information produced or purchased is presented here. A simple decision tree rationalizes each decision. The information obtained can turn out to be either favorable(good news) or unfavorable (bad news). In either case, the information may turn out to be wrong. Recognition of this makes the method consistent with the real world.
Only information gathered to guide subsequent decisions is considered here. Information may often be produced or purchased for other purposes. For example, information may be gathered to create an image rather than to improve the chance of a better subsequent decision. A specific example is implicit in a complaint articulated by a professional colleague that managers were spending money on "high tech," but were paying no attention to the results and making the decisions on intuition. These intuitive decisions, when shown to be flawed, can always be rationalized by noting that high tech supported them. The colleague noted that there was nothing wrong, per se, with using intuition to make a decision: what is suspect is paying for high tech that will not be used to make a decision. One cannot change the extant syntax: however, it would be preferable if the use of the term "information" implied only work undertaken to make a better decision. This is the only use of the term implied here.
Here, a simple decision tree is used to quantify net information values with four petroleum engineering examples. For each example, the method guides the decision to obtain (or not to obtain) information. The maximum that can be paid for information is shown. When the possibility of wrong information is neglected, there will be a bias toward purchasing favorable rather than unfavorable in-formation. This bias may correlate with actual performances that tend to be lower than projections.
Finally, the effect of oil price on net values of information is treated. A lower oil price can actually increase the value of information.
Information Value Decision Tree
The decision tree in Fig. 1 puts the decision to gather information pertinent to a subsequent decision into a rational framework. Information can be purchased (or gathered at cost) that, when known, can be characterized as good or bad news. Good news is synonymous with a more favorable view of doing what is proposed; bad news with a less favorable view. Any information, however, whether good or bad news, may turn out to be wrong. The decision tree defines the requisite frequencies and expected value outcomes used in Eq. 1 to define Cmax, the maximum amount that the information might be worth:
It would be reasonable to spend up to Cmax for information, but no more: if Cmax were negative, the information considered would have no positive value.
The simple decision tree of Fig. 1 embodied in Eq. 1 captures the essence of the decision to obtain or not to obtain some information to make a better decision in petroleum engineering and other disciplines involved in oil/gas extraction. The decision tree can, in fact, be applied universally;however, the vagaries of oil/gas extraction projects make one more cognizant of the frequency of bad news, incorrect information, and their consequences. Understandably, venders of information do not emphasize the possibility that their product may be incorrect. It is not surprising that the possibility and frequency of incorrect information have received scant attention in our literature. Ricks is an exception, and Ref. 2 specifically treats seismic studies that may point to the wrong decision. Quite recently, Campbell et al. presented an example that considers errors in seismic information. No doubt other treatments that consider the possibility of wrong information exist but remain hidden.
The more common approach to the value of information is to assume that, given sufficient expenditure for information, "truth" is unimodally asymptotic. This is the usual approach in management science. Ref. 4 is an excellent example. That paradigm may be fruitful sometimes, but not in petroleum engineering and other oil/gas extraction information decisions that characteristically differ. Characteristically, our information does not come close to truth, even with unlimited budgets. For example, no engineer can have information about a reservoir's rock and fluid properties that approaches truth. That is why the common approach to the value of information does not suffice in our work.
The decision tree used here can be applied in our business precisely because it recognizes good and bad news and information that may be correct or incorrect. To be sure, there may be gradations between good and bad news, as well as information that is between completely right and completely wrong, but the bifurcations of information with these qualities on an either/or basis robustly model the real world of our information.
Information Value: Four Examples
Suppose you have completed an economic evaluation that concludes that a certain development well has a value to you of 20. You may have measured value as an appropriately discounted present value in dollars, but that is just one definition. Here it suffices to impute arbitrary units to an economic value.
Suppose further that you have two subordinates who are equally respected. Both propose that more geophysical and geological studies be completed before drilling. Their judgments regarding the outcome of the studies differ, however.
|File Size||708 KB||Number of Pages||9|