PREDICTION MARKETS - PowerPoint PPT Presentation

Loading...

PPT – PREDICTION MARKETS PowerPoint presentation | free to download - id: 549eac-M2ZiN



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

PREDICTION MARKETS

Description:

PREDICTION MARKETS Business Intelligence Community March 10, 2009 Tony Tsai (anttsai_at_umich.edu) Matthew Comstock (comstock_at_umich.edu) University of Michigan Health System – PowerPoint PPT presentation

Number of Views:77
Avg rating:3.0/5.0
Slides: 25
Provided by: AmyP154
Learn more at: http://www.bi.umich.edu
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: PREDICTION MARKETS


1
  • PREDICTION MARKETS
  • Business Intelligence Community
  • March 10, 2009

Tony Tsai (anttsai_at_umich.edu) Matthew Comstock
(comstock_at_umich.edu) University of Michigan
Health System
2
Prediction Markets
3
UMHS Overview
  • Mission The University of Michigan Health
    System improves the health of patients,
    populations and communities through excellence in
    education, patient care, community service,
    research and technology development, and through
    leadership activities in Michigan, nationally and
    internationally.

4
UMHS Overview
  • Medical School Established 1850
  • First University Hospital Opened - 1869
  • Today a large and vibrant organization
  • 15,000 staff
  • 2,600 faculty in 27 academic departments
  • 680 medical students, 475 graduate students
    1,000 house officers (residents)
  • 5.6M in total square footage of educational,
    research and clinical space
  • FY2008 total revenues exceeding 2.8B
  • National leader in education, research and
    patient care

5
UMHS Planning
  • Planning Process at UMHS
  • Varied and complex
  • Strategic tactical
  • Occurring at all levels of the organization and
    across all components of our mission
  • UMHS Finance Short- Long-Range Projections
  • Monthly, Quarterly and Annual projections on
    faculty recruitment, key activity drivers,
    operational performance, margin, compensation,
    research productivity, endowment returns, payer
    mix, enrollment, etc.
  • 10-Year Strategic Financial Plans Major Capital

6
UMHS Planning
7
UMHS Planning
  • UMHS Financial Planning - Techniques Deployed
  • Annual Budget Processes (School Hospital)
  • Major Project Planning (capital construction)
  • Structured Discussion Meetings
  • Statistical Modeling Trend Analyses
  • Simulation Models
  • Internal Surveys Ad Hoc Data Requests
  • Expert Opinion Management Decisions
  • Prediction Markets?

8
Prediction Model Pilot
  • Original Intent to explore the use of prediction
    markets to support capital investment decisions
    by confirming operational assumptions and future
    business environment
  • UMHS Finance was chosen to do the pilot because
    of its quantitative ability and manageable size
    of around 80 (also considered MSIS/MCIT)
  • Test usability, predictive ability, and fit with
    organizational managerial culture

9
Potential Organizational Uses
  • Predict likely demand for certain type of
    treatment (support capital investment strategy)
  • Predict likely external economic conditions (e.g.
    reimbursement rates, payor mix )
  • Predict on time launch of organizational
    initiatives (e.g. opening of new facility,
    success of union negotiations)
  • Predict internal cost savings from organizational
    initiatives

10
Prediction Markets
  • Outcomes are the commodities or stocks
  • Insiders and outsiders are invited to participate
    in the market
  • Participants are given virtual currency to trade
    on the most likely outcomes
  • Market lasts until the outcomes resolve
    themselves (6 month window)
  • Market winners often receive incentives

11
Examples of Prediction Markets
12
Examples of Prediction Markets
13
Examples of Prediction Markets
14
Organizations Using Predictive Mkts.
15
Wisdom of Crowds
Good Answer
Expert
Non-Expert
Better Answer
16
Benefits of Prediction Markets
  • Leveraging untapped institutional knowledge
    (incl. external stakeholders and customers)
  • Increase in information awareness and
    transparency
  • More robust predictions to support strategic
    decision-making

17
Prediction Model Pilot
  • Training Round (6 weeks)
  • Will the football team have more than 5 wins?
    (Y/N)
  • How much will the movie Hancock gross in its
    opening weekend? ()
  • Who will win the U.S. Open, Roger Federer, Rafa
    Nadal, or Other? (Choice)
  • Operational Round (6 weeks)
  • Will total outpatient clinic visits year-to-date
    (7/1/08 to 10/31/08) exceed 605,000 visits?
  • What will be the number of total inpatient
    discharges (excl. newborns) year-to-date (7/1/08
    to 10/31/08)?
  • What will the IP average length of stay
    year-to-date (7/1/08 to 10/30/08)?

18
Results
  • Predictability All three markets had relatively
    high levels of accuracy to the final outcome.
  • Usability highly variable in terms of confidence
    with trading. 30 of participants actively
    traded. 35 of participants bought and held. 35
    did not participate. More training needed.
  • Fit with Culture Information out in the open
    would concern many within more conservative
    managerial culture. The game aspect was hard to
    process for many participants.

19
Results (Question 1)
  • Will total outpatient clinic visits year-to-date
    (7/1/08 to 10/31/08) exceed 605,000 visits?
  • Starting Value 50
  • Final Predicted Value 25
  • Actual volume 602,339

20
Results (Question 2)
  • What will be the number of total inpatient
    discharges (excl. newborns) year-to-date (7/1/08
    to 10/31/08)?
  • Starting Value 14,000
  • Final Predicted Value 14,050
  • Actual volume 14,259

21
Results (Question 3)
  • What will the IP average length of stay
    year-to-date (7/1/08 to 10/30/08)?
  • Starting Value 8.3 (6.25-6.3)
  • Final Predicted Value 49.9
  • Actual LOS 6.29

22
Pilot Decision
  • Interesting technology, but barriers to
    implementation
  • Lack of familiarity and perception
  • Work load change management
  • Information management decision making
  • Hold on implementation but look for future
    opportunities

23
Lessons Learned
  • The market medium is one that is foreign to
    many people. Lack of familiarity leads to
    perceptions of inequity
  • Building engagement requires intrinsic incentive
    (confidence that someone will do something with
    the information) as well as extrinsic incentive
    (prizes and rewards)
  • Despite advantages of the tool, it has major
    organizational cultural implications which may
    affect its implementation
  • Potential a viable alternative for decision making

24
Q A
Tony Tsai (anttsai_at_umich.edu) Matthew Comstock
(comstock_at_umich.edu)
About PowerShow.com