Risk & Uncertainty

Risk & Uncertainty

1. 'disaster' projects

   • fail 2 out of
     • 1. >40% cost growth
     • 2. >40% time slippage
     • 3. produce <50% than 1st year plan
   • 1/8 major project -> disaster
   • the bigger, >important project gets -> >likely -> disaster

2. problem: business outcome not up to expectations/ possibilities

   • better decisions
     • accurate = unbiased + appropriate range
     • uncertainty assessment & response
     • non-intuitive
     • lead to better performances
     • supported by petrol. geoscientist, engineers, economist
   • Reasons
     • under-estimate: uncertain factors (no. & magnitude), complexity of uncertainties' relationships, consequences
     • systematic biases: mental short-cuts
     • planning fallacy
       • spin success scenarios, overlook potential of mistake & miscalculations - believe in outcomes: unlikely come in budget/time - unexpected returns
     • proceed on beliefs, judgments of P but not uncertainty

3. UNCERTAINTY

   • not remove but accurately assess
   • uncertainty assessment
     • priority
       • 1. accurate (unbiased)
         • 2. uncertainty reduction (> data,time,detail, analysis)
           • 3. stop when further analysis don't change decision
   • knowing when enough is enough
   • defin: not knowing if the statement/ even is true/false
   • predictions: range/list possible outcomes
   • language of uncertainty
     • probabilities
       • scientific way to deal with uncertainty
       • systematically quantify beliefs about uncertainty
       • rules pf probability -> how to reason logically under uncertainty
       • chance of possible outcomes occur
       • views
         • Classical (Theoretical)
           • no. outcomes represent / total no. possible outcomes (3 red, 7 green -> P(red)=3/10)
         • Relative Frequency
           • probability ESTIMATED from sample data ASSUME identical events (15/20 sample wells: dry -> P(dry)=75%)
           • bigger sample -> more accurate P
         • Subjective
           • personal degree of belief of likelihood of a future event occur
           • can based on data, pass similar occurances
           • Oil & industry ( exist: unknowable possible outcomes, distribution, parameters)
   • function of not a system but what known about the system, function of personal state of information
   • RISK
     • possible -ve consequence of uncertainty
   • OPPORTUNITY
     • possible +ve consequence of uncertainty

4. Variability

   • distribution that describes variability of a natural phenomenon is not usually appropriate to describe uncertainty of a single occurance

5. Ambiguity

   • imprecise (unclear) definition of event/ statement/ outcome (wine-drinker)
   • must be removed from uncertainty statement before assign probability

6. often estimate in 80% Confidence Intervals

   • goal is set a range narrow enough to contain actual value >8/10times on ave

7. Bias

   • optimistic on central value (mean)
   • width of distribution