Designing a Risk Model Michael Schilmoeller Thursday, December 2, 2010 SAAC.

Slides:



Advertisements
Similar presentations
System Analysis Advisory Committee Futures, Monte Carlo Simulation, and CB Assumption Cells Michael Schilmoeller Tuesday, September 27, 2011.
Advertisements

Assessing Uncertainty when Predicting Extreme Flood Processes.
Will CO2 Change What We Do?
10- 1 Chapter Ten McGraw-Hill/Irwin © 2005 The McGraw-Hill Companies, Inc., All Rights Reserved.
Uses and Abuses of the Efficient Frontier Michael Schilmoeller Thursday May 19, 2011 SAAC.
Copyright © 2006 Pearson Education, Inc. Publishing as Pearson Addison-Wesley Slide
Northwest Power and Conservation Council Effects of Alternative Scenarios on Sixth Power Plan Northwest Power and Conservation Council Whitefish, MT June.
Engineering Economic Analysis Canadian Edition
System Analysis Advisory Committee - A New Metric - Michael Schilmoeller Tuesday, September 27, 2011.
The Cost-Effectiveness Premium for Conservation Michael Schilmoeller Thursday May 19, 2011 SAAC.
Session 9b. Decision Models -- Prof. Juran2 Overview Finance Simulation Models Securities Pricing –Black-Scholes –Electricity Option Miscellaneous –Monte.
Chapter 7 Sampling Distributions
CF-3 Bank Hapoalim Jun-2001 Zvi Wiener Computational Finance.
Part III: Inference Topic 6 Sampling and Sampling Distributions
Copyright © 2005 Brooks/Cole, a division of Thomson Learning, Inc Chapter 10 Introduction to Estimation.
Chapter 14 Risk and Uncertainty Managerial Economics: Economic Tools for Today’s Decision Makers, 4/e By Paul Keat and Philip Young.
Probability and the Sampling Distribution Quantitative Methods in HPELS 440:210.
Introduction to ModelingMonte Carlo Simulation Expensive Not always practical Time consuming Impossible for all situations Can be complex Cons Pros Experience.
© Harry Campbell & Richard Brown School of Economics The University of Queensland BENEFIT-COST ANALYSIS Financial and Economic Appraisal using Spreadsheets.
SAAC Review Michael Schilmoeller Thursday May 19, 2011 SAAC.
Sampling Theory Determining the distribution of Sample statistics.
A Choice of Platform: Excel ® and Crystal Ball ® Michael Schilmoeller Thursday, December 2, 2010 SAAC.
Northwest Power and Conservation Council 6 th Plan Conservation Resource Supply Curve Workshop on Data & Assumption Overview of Council Resource Analysis.
1 Regional Portfolio Model and Direct Use of Gas Assessment Michael Schilmoeller NW Power and Conservation Council for the Regional Technical Forum Tuesday,
SAAC Review Michael Schilmoeller Tuesday February 2, 2011 SAAC.
SIMULATION USING CRYSTAL BALL. WHAT CRYSTAL BALL DOES? Crystal ball extends the forecasting capabilities of spreadsheet model and provide the information.
The Council’s Risk Model and The Requirements of the Act Michael Schilmoeller Thursday, December 2, 2010 SAAC.
Climate Change and The NW Power Supply Climate Impacts on the Pacific Northwest University of Washington April 21, 2009.
Net Metering Technical Conference Docket No PacifiCorp Avoided Costs October 21, 2008 Presented by Becky Wilson Executive Staff Director Utah.
Discussion of Resource Plans Michael Schilmoeller for the Northwest Power and Conservation Council Wednesday, June 10, 2009.
10/4/20021 Systems Analysis Advisory Committee (SAAC) Friday, October 4, 2002 Michael Schilmoeller John Fazio.
The Primary Sources of Risk and a New Metric Michael Schilmoeller Thursday May 19, 2011 SAAC.
Preliminary Results with the Regional Portfolio Model Michael Schilmoeller for the Northwest Power and Conservation Council Generation Resource Advisory.
Engineering Economic Analysis Canadian Edition
System Analysis Advisory Committee FAQ Questions Michael Schilmoeller Friday, January 25, 2013.
1 Estimation From Sample Data Chapter 08. Chapter 8 - Learning Objectives Explain the difference between a point and an interval estimate. Construct and.
HYPOTHESIS TESTING. Statistical Methods Estimation Hypothesis Testing Inferential Statistics Descriptive Statistics Statistical Methods.
Copyright © 2005 Brooks/Cole, a division of Thomson Learning, Inc Chapter 12 Inference About A Population.
Sixth Northwest Conservation & Electric Power Plan Draft Wholesale Power Price Forecasts Maury Galbraith Northwest Power and Conservation Council Generating.
The Council’s Approach to Economic Risk Michael Schilmoeller Northwest Power and Conservation Council for the Resource Adequacy Technical Committee September.
The Council’s Regional Portfolio Model Michael Schilmoeller for the Northwest Power and Conservation Council Generation Resource Advisory Committee Thursday,
1 Introduction to the Regional Portfolio Model Michael Schilmoeller NW Power and Conservation Council Thursday, June 10, 2010.
1 Systems Analysis Advisory Committee (SAAC) Thursday, December 19, 2002 Michael Schilmoeller John Fazio.
Chapter McGraw-Hill/Irwin Copyright © 2008 by The McGraw-Hill Companies, Inc. All rights reserved. Risk and Capital Budgeting 13.
Estimation Chapter 8. Estimating µ When σ Is Known.
Jeopardy Hypothesis Testing t-test Basics t for Indep. Samples Related Samples t— Didn’t cover— Skip for now Ancient History $100 $200$200 $300 $500 $400.
How the RPM Meets the Requirements for a Risk Model Michael Schilmoeller Tuesday, February 2, 2011 SAAC.
System Analysis Advisory Committee Assumptions for the Seventh Power Plan Michael Schilmoeller Friday, January 25, 2013.
Basic Business Statistics, 10e © 2006 Prentice-Hall, Inc.. Chap 7-1 Chapter 7 Sampling Distributions Basic Business Statistics.
Introduction to Inference: Confidence Intervals and Hypothesis Testing Presentation 8 First Part.
Introduction to Inference: Confidence Intervals and Hypothesis Testing Presentation 4 First Part.
AP Statistics Section 11.1 B More on Significance Tests.
Uncertainty in the Regional Portfolio Model Michael Schilmoeller for the Northwest Power and Conservation Council Generation Resource Advisory Committee.
Review Normal Distributions –Draw a picture. –Convert to standard normal (if necessary) –Use the binomial tables to look up the value. –In the case of.
Making Decisions about a Population Mean with Confidence Lecture 35 Sections 10.1 – 10.2 Fri, Mar 31, 2006.
1 Proposed Input Assumptions to RTF Cost-Effectiveness Determinations February 2, 2010.
The Nature of Risky Futures Michael Schilmoeller Thursday May 19, 2011 SAAC.
©2009 McGraw-Hill Ryerson Limited 1 of Risk and Capital Budgeting Risk and Capital Budgeting Prepared by: Michel Paquet SAIT Polytechnic ©2009 McGraw-Hill.
Chapter 9: Introduction to the t statistic. The t Statistic The t statistic allows researchers to use sample data to test hypotheses about an unknown.
Sampling Theory Determining the distribution of Sample statistics.
Statistical Inferences for Variance Objectives: Learn to compare variance of a sample with variance of a population Learn to compare variance of a sample.
Making Decisions about a Population Mean with Confidence Lecture 35 Sections 10.1 – 10.2 Fri, Mar 25, 2005.
Sixth Northwest Conservation & Electric Power Plan Draft Wholesale Power Price Forecasts Maury Galbraith Generating Resource Advisory Committee Meeting.
Confidence Intervals Cont.
Challenges, Issues, and Future Direction of the SAAC
Psychology 202a Advanced Psychological Statistics
Key Findings and Resource Strategy
Professor S K Dubey,VSM Amity School of Business
Statistical Thinking and Applications
Presentation transcript:

Designing a Risk Model Michael Schilmoeller Thursday, December 2, 2010 SAAC

2 Overview Scope of uncertainty Decision trees (briefly) and Monte Carlo simulation Implications of cost and risk accuracy to the number of futures The number of possible plans and finding the “best” plan Computational alternatives

3 Sources of Uncertainty Scope of uncertainty Fifth Power Plan –Load requirements –Gas price –Hydrogeneration –Electricity price –Forced outage rates –Aluminum price –Carbon allowance cost –Production tax credits –Renewable Energy Credit (Green tag value) Sixth Power Plan –aluminum price and aluminum smelter loads were removed –Power plant construction costs –Technology availability –Conservation costs and performance

4 Impact on NPV Costs and Risk Scope of uncertainty C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs.xlsm

5 Decision Trees Estimating the number of branches –Assume possible 3 values (high, medium, low) for each of 9 variables, 80 periods, with two subperiods each; plus 70 possible hydro years, one for each of 20 years, on- and off-peak energy determined by hydro year –Number of estimates cases, assuming independence: 6,048,000 Studies, given equal number k of possible values for n uncertainties : Impact of adding an uncertainty: Decision trees & Monte Carlo simulation

6 Monte Carlo Simulation MC represents the more likely values The number of samples is determined by the accuracy requirement for the statistics of interest The number of samples m k necessary to obtain a given level of precision in estimates of averages grows much more slowly than the number of variables k: Decision trees & Monte Carlo simulation

7 Overview Scope of uncertainty Decision trees (briefly) and Monte Carlo simulation Implications of cost and risk accuracy to the number of futures The number of possible plans and finding the “best” plan Computational alternatives

8 Monte Carlo Samples How many samples are necessary to achieve reasonable cost and risk estimates? How precise is the sample mean of the tail, that is, TailVaR 90 ? Implication to Number of Futures

9 Relationship Between the Size of the Sample and the Accuracy Depends on knowledge of the distribution Given the distribution, requires knowledge of how the accuracy depends on sample size Implication to Number of Futures

10 Central Limit Theorem Both our cost and our risk estimates are averages CLT says that as the number of samples used to estimate the mean increases, the distribution of the sample means tends to normal Unfortunately, it doesn’t say how fast it tends to normal or how the shape of the underlying distribution affects the rate of approach Implication to Number of Futures

11 TailVaR 90 Implication to Number of Futures C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm

12 Assumed Distribution Implication to Number of Futures C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm

13 Set Up a Sampler Implication to Number of Futures C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm

14 Dependence of Tail Average on Sample Size Implication to Number of Futures C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm, worksheet “Simulation”

15 Implication to Number of Futures Dependence of Tail Average on Sample Size C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm, worksheet “Simulation”

16 Implication to Number of Futures Dependence of Tail Average on Sample Size C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm, worksheet “Simulation”

17 Implication to Number of Futures Dependence of Tail Average on Sample Size σ=2.040 C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm, worksheet “Samples_50”

18 Implication to Number of Futures Dependence of Tail Average on Sample Size C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs 02.xlsm, worksheet “Samples_75” σ=1.677

19 Chi-Squared (X 2 ) Tests Check the hypothesis that our sample has the variation from normal by chance (p) 50 samples per calculation: p= samples per calculation: p=0.10 Implication to Number of Futures

20 Accuracy and Sample Size Estimated accuracy of TailVaR 90 statistic is still only ± $3.3 B (2σ)!* Implication to Number of Futures *Stay tuned to see why the precision is actually 1000x better than this!

21 Accuracy Relative to the Efficient Frontier C:\Backups\Plan 6\Studies\L813\Analysis of Optimization Run_L813vL811.xls Implication to Number of Futures

22 Conclusion At least 75 samples are needed for determining the value of our risk metric –Known distribution of statistic –The precision of the sample Our risk metric is 1/10 of the total number of futures We need to test our plan under 750 futures to obtain defensible results Implication to Number of Futures

23 Overview Scope of uncertainty Decision trees (briefly) and Monte Carlo simulation Implications of cost and risk accuracy to the number of futures The number of possible plans and finding the “best” plan Computational alternatives

24 Finding the Best Plan Each plan is exposed to exactly the same set of futures, except for electricity price Look for the plan that minimizes cost and risk Challenge: there may be many plans Implication to Number of Plans

25 Avogadro’s Number In the draft Sixth Plan, there were at times nine capacity expansion candidates, not counting conservation and demand response Total number of possible plans: 1.3 x Number of molecules in a mole under standard conditions (Avogadro’s number): 6.02 x Implication to Number of Plans

26 Candidates for the Final Plan In the case of the final study for the Sixth Power Plan, there were a mere 6.7 trillion Implication to Number of Plans Source: C:\Backups\Olivia\SAAC 2010\ SAAC First Meeting\Presentation materials\States_L813.xls

27 Space of feasible solutions The Set of Plans Precedes the Efficient Frontier Reliance on the likeliest outcome Risk Aversion Efficient Frontier Implication to Number of Plans

28 Finding the “Best” Plan C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\Asymptotic reduction in risk with increasing plans.xlsm Implication to Number of Plans

29 OptQuest ® Recommendations The RPM used to produce the portfolio for the Council’s draft Sixth Power Plan has 69 decision variables Our finding of 3500 simulations is consistent with OptQuest guidelines (page 156, OptQuest for Crystal Ball User Manual, © 2001, Decisioneering, Inc. ) Implication to Number of Plans

30 Overview Scope of uncertainty Decision trees (briefly) and Monte Carlo simulation Implications of cost and risk accuracy to the number of futures The number of possible plans and finding the “best” plan Computational alternatives

31 How Many 20-Year Studies? How long would this take on the Council’s Aurora2 server? Implication to Computational Burden

32 Time on Council’s Server Council’s server tech specs: –Xeon W3580 processor –3.33 MHz, L3 Cache 8 –Quad core, 8 Threads per core 20-year, hourly study requires 128 minutes Total time requirement for one study: 2.33 x 10 5 days (639 years, 3 months, 7 days) Implication to Computational Burden

33 Time on a Supercomputer October 28, 2010: China acquires the fastest machine on earth: 2.5 petaflops (floating point operations per second) The Tianhe-1A supercomputer is about 50% faster than its closest rival. Implication to Computational Burden

34 On the World’s Fastest Machine Assume a benchmark machine can process 20- year studies as fast: –Xeon 5365, 3.0 MHz, L2 Cache 2x4, 4 cores/4 threads per core –38 GFLOPS on the LinPack standard –To the extent this machine underperforms the Council server, the time estimate would be longer Total time requirement for one study on the Tianhe-1A: 3.54 days (3 days, 12 hours, 51 minutes) and estimated cost $37,318 Implication to Computational Burden

35 How Do We Achieve Our Objectives? If it takes more that a workday to perform the simulation, the risk of making errors begins to dampen exploration In the next presentation, we consider alternatives and the RPM solution Implication to Computational Burden

36 End