Monday, November 21, 2005

Digital tug-of-war

Tug of war, Philip Greenspun Last Wednesday I got home just in time to see a rather unusual invention introduced on The New Inventors called the "Family Winner", a piece of software which assists a win-win resolution after a marital breakup.

At first I thought this would be such a useless piece of software. How could a software algorithm - which is a glorified mathematical formula - be truly useful in terms of negotiating with the infinite variability of human minds and human emotions? Then I gradually came to accept its value. Software is already used extensively for decision support and risk modelling. Why not for negotiating? You can't bribe it and can expect it to be 100% objective and impartial. In fact limited software for conflict resolution support already exists.

The concept on which the Family Winner rests is that in any negotiation the two parties don't necessarily have the same aims and conflicting aims don't always perfectly overlap. The two parties typically place higher values on particular aims and lower values on others.
Through a series of trade-offs and compensation strategies, disputants can often achieve 70-80% of what they require, rather than the traditional 50-50 approach to resolving disputes. The system takes a rational approach to negotiation and forces disputants to focus upon their interests.

It is based on game theory techniques and subsequent algorithms developed by Nobel Laureate John Nash. Basically, the disputants are advised to use the Family Winner system as the 3rd step in a computerized conflict resolution process (the first 2 steps already exist in different software packages, but will help the disputants be rational by the time they hit Family Winner).
Apparently this has not just been trialled as a replacement for using a lawyer in a domestic breakup but also for assisting negotiation of industrial relations disputes. It makes me wonder what the potential for "negotiation support" software could be. Is it just a niche or is there a widespread need for software-assisted negotiation? Could it be useful in international-scale conflicts? Even if it could be useful, could two parties ever agree to a computer-modelled resolution which is mathematically the best case outcome or will they prefer human-negotiated resolutions because they feel they can work these more to their advantage?

I know I'd feel that way. It is a bit like betting on those dogs in the picture above versus the roulette wheel; certain dogs have a distinct advantage, whereas each pocket of the roulette wheel has an equal chance of being the winner regardless of human/animal emotion!

Image courtesy of Philip Greenspun
Technorati Tags: , , ,


Anonymous djfoobarmatt said...

Maybe we can use this software for negotiating contracts under the new IR laws?

November 21, 2005 11:25 pm  
Blogger Lisa said...

Nice one :) but doubt the employers will want it. I think they'll like the current (as in the new) system just as it is...

Good IR negotiation software would constantly demand employers make more concessions!

November 22, 2005 10:22 am  

Post a Comment

<< Home