Come together with the global Drupal community in Rotterdam, 28 Sept – 1 Oct 2026. Sessions, contribution, connection, and Early Bird savings until 8 June.This is a very cool module and could be expanded in very interesting directions. It's quite unique in what it does. Do you plan on developing this further?
wim
Comments
Comment #1
danieljrmay commentedHello,
Sorry for the very late reply. I haven't checked this modules page for a month or so (as if you haven't guessed) and I didn't seem to get notified of your message.
Yes, this is just a baby module at the moment and I do have plans to develop it further. The first thing I wanted to do was to add the ability to record actual "event instances" which have happened in a log e.g. "Our servers hard drive broke today, it cost us $1000 to replace it when all costs (hard drive, lost buisness, administrator time, restore from backups time) are considered". These "event instances" can then be used to calculate a modified probability of occurrence for their corresponding "risk event class" using Bayesian statistics.
As you say, there are lots of directions which we could go in e.g. pretty graphical representations of the results, incorporating risk mitigation, elimination, transfer & retention.
What do you think?
Best wishes,
Dan
Comment #2
battochir commentedNo worries.
I like your ideas, particularly the pretty graphical representations of the results, that could help more easily explain risk mitigation, elimination, transfer & retention to those who do not understand the numbers.
That is the crux of how I personally view a big strength of this module.
As is, giving a simple standard deviation or pseudo gamma representation would be understood by very few people. I think a module like this should keep the technical side under the hood. A layer in between the user interface and the raw output that "dumbs down/makes raw output easily understood and immediately usable" would be nice and resonate better with users.
I like your idea of instances because risk events can be "personalized". While Bayesian statistics is no panacea, it is powerful enough to be generally applied to most cases dealing with a random distribution of possible outcomes at a lower level.
I personally think it would be nice to allow people to enter the parameters of an event and receive a simple probability of occurrence (an understandable one, preferably in graph form) that can be whittled down to particular instances of events within the distribution itself. This could be similar to option pricing, only applied to random events. Just like the relationship between time, volatility, and asset price movement affects an option's value. Only here the parameters would/could differ such as using only time and a volatility measurement. The volatility measurement itself could be entered by the user under something like "likelihood of occurrence" with 'low', 'medium', 'high' inputs. Each input could be assigned a range of volatiliy, i.e. low = 1% - 33.3%, medium = 33.34% - 66.6%, high = 66.67% - 100% and these ranges would be dependent upon the time-to-occurrence input. The greater the time to occurrence the higher the volatility range within a particular band.
Yes, it's crude. However, it could be quite usable for mass usage. Add the pretty graphs in for easy understanding and it could become quite addictive to use. Will my girlfriend call me back after I was such a clown at her parents? Will the Euro replace the dollar as the world's preeminent currency? Will oil go back to $148.00 US per barrel? Should I buy a house now or wait? Endless possibilities.
What do you think?
cheers,
wim
Comment #3
zeezhao commentedBy the way, I did an internal implementation for recording risk events, doing calcs, graphing etc. Created some custom modules in addition to the usual (cck, views, computed_feld, etc). Let me know if you want to discuss. But it looks like this module is not longer active anyway.
Comment #4
danieljrmay commentedI am not currently maintaining this module. So I am marking this as Postponed.
Comment #5
dwkitchen commented