|
|
 |
Software
principles : gambling theory and statistics
Our software products
mainly rely on two mathematical models :
- the theory of games,
used in many decision making softwares
- Statistical models
that can explain current behaviors
The theory of games is
used in our softwares to enforce business
models and compute business functions. Statistical models
are used to build models in order to compute figures.
|
The
software estimates every risk as
figures measuring probability of the risk to occur and consequences of
its occurrence in concrete terms. The evaluation of the risks
is computed according to answers to a questionnaire. The evaluation of
the
consequences is computed as mathematical expectancies.
Computations lean on probability algebra, and are stored
with the questionnaires in what we call a "knowledge base".
Probability
algebra is the frame and strength of the solution
:
- we can use actual
statistical data within the knowledge base,
- risks can be easily computed
as figures with statistics
- we can put all parameters
outside the software product
- it can evolve as "Scoring"
models will never be able to
Moreover, each response, each declared incident becomes a statistical
information : we can therefore feed the model with figures provided
from the model and improve accuracy of the data.
|
|
In practice : learn with
no risks
It may look funny to view
risks as in a Casino, imagine the
chief officer of a
company, a project leader or a product manager "betting" on risks :
however, there is a clear link between risks and choices we make within
our
environment.
In fact, many executive
has very few visibility on the
impact of choices they make : this is why our products provide
comprehensive figures on risks and their consequences, therefore
helping
to understand the stakes.
|
One
of the side effects of our software is
to allow decision makers to estimate " at no risk " their choices and
decisions: the product computes and feedbacks the consequences of
these choices before they become real. A study on the behaviour of
users
of "SelfScore Créateur d'entreprise" (french product for
"new business" analyses) shows that
the user
begins by answering all the questions " without arising any "
and
starts to take care of risks only when they get alerted by the results
of the analysis.
|
|
Competing
with experts
Our software is regularly
confronted with experts in risk
management in the various domains in which we build products : these
confrontations always showed the capability of our solution to compete
with experts.
In many cases, the
results of our analyses seems very similar to results produced by human
experts. But the main difference it
that our software produces figures, while experts are very reluctant to
produce any, and it never forgets any of the subjects to analyse.
|
However, various studies show that decision makers require simple
results : it is why
we also produce color codes which notify directly
on the
priorities. We can also produce specific quotation depending on the
subject our software are managing.
|