informs ny

Feb. 1, 2006 Second Student-Practitioner Evening Forum
Sponsored by: NY Metro INFORMS Chapter and Pace University


On February 1, 2006, the NY Metro Informs Chapter and Pace University
conducted a Student/Practitioner (Evening) Event on the theme
"A Day in the Life of an Analytical Problem Solver"
It took place in the grand auditorium at Pace's downtown campus and was
attended by a hundred or more of distinguished practitioners, students and faculty.

In the first round the panelists presented their personal perspectives.

Vadim Konstantinovsky, Author of book on "Quantitative Management of Bond
Portfolios", and a Senior VP with Lehman's Fixed Income Quantitative Portfolio
Group stressed the importance of solid educational background in math, finance,
presentation skills and computer science. Lehman has tons of historic data and
is a large provider of benchmarks for fund managers, maintains Lehman index
of some 15k securities. There are adequate methods and software available on
the market, however factor methods were not successful in bond markets. FI
research is a cost center for Lehman.

Leon Schwartz, President of Informed Decisions, reflecting how one is going
to see themselves in 10-15 years.  Al went through three careers (so far); a
mathematician at Pitney Bowes, consultant at Informed Decisions and presently
also Academician.
      Corporate = hectic. Stress is good. Network, stay connected. Spend 60%
of the day to market your business. Don't live in your office. Art of
"Opportunity Identification and Analysis".

Maggie Belknap of West Point highlighted the challenges facing the military;
multibillion budgets to acquire technology for warfare 10-20 years in the
future, base closures, ROTC closures, removing nukes from Europe, professional
force, declining prestige of joining the profession, training thousands of
lieutenants to grow few four-star generals 30 years later. SAS - a successful tool,
statistical cost/benefit analysis, regression, optimization, no agreement on
criteria.

Warren Mitofsky of Mitofsky International, statistician and founder of a
survey research firm, conductor of election exit polls worldwide since 1967.
Typical Presidential Election Day planning takes two years. In 2004, first exit
polls 10 am, shared with clients by 1pm, leaked as win for Kerry, not wrong in
any races (county, precinct).. Exit polls; incomplete data used to obtain
complete results (despite developments like provisional ballots in Ohio). Election
day results are well behaved, highly correlated; Bayesian methods a waste of
time.

Next, a break planned for 15 minutes turned into a much longer forum for
networking.

The meeting concluded with a free format Q&A sessions; on two topics:

How to better connect with theoretical OR, Statistical models?
  • Decide what the real problem is, then use the methods as tools.
  • A model for sales allocation was intended as decision aid. Times changed but model still used.  Three Ps (peoples, processes, politics.)
  • Backtrack from business problem. Optimize within constraints.
  • Its useful to undergo a sales training - it also teaches listening.
  • Understand the customer, then pick OR tools. Mentor on results to better understand the customer,
What are the skills important in (OR) careers today?
  • Creativity, work with outside clients,
  • Big Picture, trends, globalization, predictive models, creativity in better models, data driven insights,
  • Big Picture boom/bust, CRM, desired customer profiles
  • Search, Google, data quality first before data mining,
  • Data mining not to be misused to show what you want not what is.
  • Telephone conversation records, complete with time stamp, what, who; to maintain a high quality Data Warehouse (future of business); used to construct  analytic models, factors,..
  • Transitioning from symptoms to solutions.
  • Dive deeper into data, do quick fixes, then a bigger stuff.
  • 6 weeks: deliver something of value or you are gone.

Concluding remarks on the role of OR
  •   It's big time stuff/important,
  •   Although some areas require little customer experience,
  •   Customer satisfaction is important, but data are superficial and splintered, requiring polls and the right questions
  •   Cost experience important, retention is real money.
  •   OR is a discipline that helps one deal with many problems at once.

The meeting was attended by the largest group yet of students, faculty,
practitioners and students who recently became practitioners.

Notes by Pawel Radzikowski



Home