The $242 Million Software Error | Hacker News Reader

The $242 Million Software Error

spectrum.ieee.org

28 pointsptrckryn19 comments

Whereas, if the 'risk model' had failed for any other reason, there'd have been no consequences. What's to say their model was correct even on paper? Funnily, nobody would have said anything if the mistake had made money. So, caveat emptor!

This site is inaccessible on mobile browsers (or at least iPhone)

It "helpfully" redirects to the IEEE Spectrum mobile homepage. If you are making a mobile version of a site, please don't redirect blindly. I followed a link, I expect to go there.

EDIT: thankfully, the article was recent and a link to the mobile story shows up on the page we get redirected to.

Yeah, hate it when the websites do that. It's even more fun when you're on a desktop and they 'detect' you're on mobile anyway!

According to their SEC filing, the source of the error was that they were using both percentages and decimals in the software, and someone thought that a decimal was a percentage. It's surprising how little your risk model does for you when you multiply it by 0.01.

I wonder if it was a formula in an Excel spreadsheet.

"... an independent consultant with expertise in quantitative investment techniques who will review disclosures and enhance the role of compliance personnel."

I assume this is press-release-based and so pretty thin on actual information, but if someone told me they had a $242 million software error and the plan for the future was to "review disclosures" I would not be reassured.

$242M? Meh. http://en.wikipedia.org/wiki/Ariane_5_Flight_501

That wasn't a software error in the same sense, it was more of a design error.

A very good point. Let's settle for a software engineering error :)

Guess they shouldn't have written those unit tests.

There are no software errors, only programmer errors. The computer did exactly what it was told to do.

[Generated by HNCommentGenerator Vers. 2.04 2011-02-11 17:01:44 GMT]

My failing hard drive would like to have a word with you.

A failing harddrive is a hardware problem.

Software is a property of hardware. I can yell at my computer all day and it won't do a thing. Sometimes, there is a software error. Sometimes, there is a programmer error. The computer doesn't "know" or "care" the difference.

Software and hardware have significantly different risk mitigation strategies. No matter how much high quality code you write, you can't keep a harddrive from failing. And when a harddrive fails, you don't immediately blame it on software (of course, software can drive hardware beyond its limits and cause a failure of the hardware, but that is software abusing the hardware or not knowing the limits of the hardware, and thus a software problem).

While software risks should ideally not be included in part of those risks, investing has some inherent level of risk.

Once society finally grasps that, we'll hopefully be better off. Not condoning what these guys did by any means, but caveat emptor for sure.

I can't wait until the finance industry is subject to risk. I mean, yeah, you can lose your job... but you'll always retire on piles and piles of money as a finance exec.

Speaking of software errors, that link redirects me to the mobile site home (not the article page), so I guess I won't read this story.

Copy the url, search for it on google, under the article click "options," then "mobile formatted."

Reply on news.ycombinator.com