The Games We Play: the Future of DFA Models’ Interfaces Aleksey Popelyukhin, Ph.D. Vice President, Information Systems & Technology 2 W INGS R ISK S ERVICES
Dynamic Financial Analysis Studies the company as a whole Takes into account multiple types of risk Compares results of different strategies Generates distribution of possible outcomes but 1
Dynamic Financial Analysis is it really Dynamic 2 ?
Now that you have your distribution 3 questions arise
How did I get there? staying profitable being not so profitable making all sorts of mistakes 3a time
Why didn’t I react? at that point in time 3b time
What could I do? how to avoid getting into “red zone” what levers do I have which scenarios require actions 3b time
Non-dynamic DFA 4 Doesn’t allow analysis of particular scenarios or group of scenarios Doesn’t allow time-variant strategies Doesn’t’ allow scenario-dependent strategies In short, All decisions has to be made before simulations.
DFA can be enhanced 4a If one adds: Ability to run particular scenario (s) Higher level of details Clearly identified “ levers ” Time animation Ability to define strategy “ on-the-fly ” (as a response to scenario behavior)
Should we start from scratch? 5 We need a computing environment that provides: High level of details Time animation Interactivity
Welcome to the world of Games ! 5
Simulation Games 6 Provide a computing environment with: High level of details Time animation Interactivity
Similarity is striking! 7 Empire/City/Railroad Building game vs. Insurance Company management simulation: Visual Metaphor Functional Metaphor “Gameplay”
Visual Metaphor 7a u Zoning Areas u Bulldozing Cost u Tall Buildings u Cultural Influence u Natural Disasters (fires, earthquakes) u etc. u Lines of Business u Brokerage Fees u Large Losses u A.M. Best Rating u Natural Disasters (fires, earthquakes) u …
Functional Metaphor 7b u Emperor/Mayor u Advisors u Progress Monitoring u Orders u Taxes u etc. u Decision Maker u CFO/Actuaries/Claims u Reporting u Actions u Taxes u …
“Gameplay” 7c u Watch Critical Indicators u Check Advisors Reports u Make Decisions u Intervene u Observe Effect in “real” time u repeat
Benefits 8 “ Game-like ” computing environment serves as: Training ground for Executives Fine-Tuning tool for Strategists Self-Learning program for AI Monitoring paradise for Management Modeling Sandbox for Actuaries
Training 8a Thinking with Images Making Decisions without any Damage Behavior in case of Disaster
Fine-Tuning 8b Choosing Success Criteria Refining Strategies Definitions (if..then) Business Processes Assumption Testing
Monitoring 8c Multiple factors at once “Real” time Visual cues
Dynamic Response 8c Time is visualized Non-linear strategies Feedback u Reserves u Prices u Investment
Implementation 9 Starting with the existing game Engine we will get: u Rich Interactive Interface u Flexible Modeling Environment we will not get: u Industry specific Models u Ability to import and constantly update Data
Engine Modifications 9a Simulate Economy Simulate Company Simulate Correlations Simulate the rest: u Competition u Geography u Weather u Catastrophes
Monitoring 9b Even if we were able to: u Design new Models u Collect necessary Data u Implement all Algorithms We still have to decide: u Which statistics to calculate u Which indicators to monitor u Which criteria to use for alerts
Presentation of Analysis Results to the Management of the Company 10 built-in!
Conclusion end Inevitably, DFA interfaces will resemble modern simulation game environments. Everyone will benefit: u Actuaries u Management u Shareholders u Policyholders
Real-life examples P.S. 1. Spatial visualization:
Real-life examples P.S. 2. Time animation:
Real-life examples P.S. 3. First Actuarial Game: