Comments on: Will Computers Ever Help With Decisions? (No) http://timoelliott.com/blog/2007/12/will_computers_ever_help_with_.html Timo Elliott's Business Analytics Blog Fri, 10 Feb 2012 13:24:00 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Timo Elliott http://timoelliott.com/blog/2007/12/will_computers_ever_help_with_.html/comment-page-1#comment-49 Timo Elliott Thu, 21 Feb 2008 05:13:00 +0000 http://192.220.58.236/blog/?p=81#comment-49 This post is not about technology, it's about people. At least since the abacus, people have always technology to give them an edge. In the big scheme of things, there's nothing special about tomorrow's new techniques. They will help do things that people can't do today. But then people will take that for granted and move on to the next level of complexity. And they will remain just as unhappy with their computer systems' ability to help them make decisions.... This post is not about technology, it’s about people.
At least since the abacus, people have always technology to give them an edge. In the big scheme of things, there’s nothing special about tomorrow’s new techniques.
They will help do things that people can’t do today. But then people will take that for granted and move on to the next level of complexity.
And they will remain just as unhappy with their computer systems’ ability to help them make decisions….

]]> By: Data. Knowledge. Wisdom. http://timoelliott.com/blog/2007/12/will_computers_ever_help_with_.html/comment-page-1#comment-50 Data. Knowledge. Wisdom. Mon, 10 Dec 2007 18:07:30 +0000 http://192.220.58.236/blog/?p=81#comment-50 <strong>Can Systems Decide?</strong> The topic of how/if systems can make (or help make) decisions has been getting increasing discussion in the blogosphere. One of the latest posts came on Tim Elliot’s BI Questions Blog, where Tim argues that the answer is “No.” I shall...

Can Systems Decide?

The topic of how/if systems can make (or help make) decisions has been getting increasing discussion in the blogosphere. One of the latest posts came on Tim Elliot’s BI Questions Blog, where Tim argues that the answer is “No.” I shall…

]]> By: Dave Dixon http://timoelliott.com/blog/2007/12/will_computers_ever_help_with_.html/comment-page-1#comment-48 Dave Dixon Wed, 05 Dec 2007 16:32:48 +0000 http://192.220.58.236/blog/?p=81#comment-48 I agree with your point that a "decision" implies a lack of information. With perfect information, you just apply a control algorithm, e.g. the automatic choke example. However, a lack of information does not exclude the help of computers. Imperfect information implies uncertainty about the future. Consider blackjack: if you count cards, you have a running distribution in your head of what cards are left in the deck. This allows you to make betting decisions which maximize your expected winnings. You don't know precisely which card will come up next, but it's not completely unknown either. Now suppose you program a computer with the card-counting/betting algorithm. The best a person could hope to do over time is match the computer's performance. In all likelihood the machine will win more over time, if only because it doesn't make mistakes. Further, you could supply the computer with a more complex but superior algorithm, beyond the capacity of most humans to manage in their head. Most business decisions involve uncertainties and future decisions with complexity well beyond that in a game of blackjack, so we would expect the computer to be of even greater service. Take a look at the <a href="http://blog.provisdom.com/?p=27" rel="nofollow">case study we recently posted at our blog</a>, which arguably still constitutes a simple example. I agree with your point that a “decision” implies a lack of information. With perfect information, you just apply a control algorithm, e.g. the automatic choke example.
However, a lack of information does not exclude the help of computers. Imperfect information implies uncertainty about the future. Consider blackjack: if you count cards, you have a running distribution in your head of what cards are left in the deck. This allows you to make betting decisions which maximize your expected winnings. You don’t know precisely which card will come up next, but it’s not completely unknown either.
Now suppose you program a computer with the card-counting/betting algorithm. The best a person could hope to do over time is match the computer’s performance. In all likelihood the machine will win more over time, if only because it doesn’t make mistakes. Further, you could supply the computer with a more complex but superior algorithm, beyond the capacity of most humans to manage in their head.
Most business decisions involve uncertainties and future decisions with complexity well beyond that in a game of blackjack, so we would expect the computer to be of even greater service. Take a look at the case study we recently posted at our blog, which arguably still constitutes a simple example.

]]>