Military software predicts Greek default | Hacker News Reader

Military software predicts Greek default

business.financialpost.com

54 pointschailatte28 comments

not to be too cynical, but any military, for lack of a better category, "fire-control" software with an 85% success rate probably got that rate through very conservative target selection and heavy-handed audit of results.

Full disclosure: I was a Navy Tactical Action Officer, the watchstander with weapons release authority, on an aircraft carrier. It's pretty deeply inculcated to not trust the machine. Is that blind job protectionism or professionally sound paranoia? I don't know.

That sounds like proper training. In life-and-death situations where the luxury of at least some time is available, why would anyone trust the machine without first manually evaluating the evidence. If our early warning system for a nuclear attack from Russia gave us a warning, while it would be prudent to automate some defensive measures like scrambling fighters and alerting our forces, I doubt we would allow the defense system to have the authority to retaliate on its own. It would need a human (several, actually) to verify the information to the best of their ability and make the judgement calls necessary to order any retaliation.

Given that this isn't a life-or-death situation, I think we can afford to go with the machine's results until we see more evidence either supporting or rejecting the Greek default.

BTW, that sounds like an incredibly cool job, although I imagine it was probably more staying up all night with nothing going on. Still cool though.

the issue of job protectionism comes up because there's some pretty decent evidence that unmanned drones and missiles are better for certain jobs on the offensive side, though we may come to that conclusion because it's not our people dying in the blast (usually).

True, but the 'unmanned' drones and missiles usually still have a man remotely piloting or pulling the trigger. I doubt a drone will ever be given permission to select and fire on targets.

I'm sure if we disregard ethics, and loss of human life can be taken as an error calculation in a piece of software, then computer systems can be far more efficient. However, I hope we're not at that point.

I mean if we were truly rational we could see that collateral damage could be a factor in whether a campaign was successful, taking into account the attitude of the conquered nation and how it may or may not affect insurgency or resistance. Though I don't want to believe that's the reason we try to preserve life.

Easy to scoff and funny too. However, note the input data used by the software..

A release from the company explained that Senturion analyzed the positions of leading EU officials, private financial institutions and Greek political factions in the week following a critical parliamentary vote at the end of June.

No balance sheet analysis, just verbiage. Combined with what?

It probably simulated a battle between these factions, if I'm understanding it correctly.

It's probably something similar to Bruce Bueno de Mesquita's game-theory system, which is used by the CIA claims has a 90% success rate, even when the experts who input the data get it wrong...

It only needs 4 types of input:

  1. Who has a stake in trying to shape the outcome.
  2. What they *say* they want.
  3. How focused they are on the issue compared to other issues.
  4. How much persuasive influence could they exert if they chose to engage on the issue.

http://www.ted.com/talks/lang/eng/bruce_bueno_de_mesquita_pr...

In other news, military software predicts sun to rise tomorrow.

Genuine Greek sarcasm. :)

At least we still have that :P

LMAO at the photo caption: "Wherever Senturion is housed it probably looks a lot like this, but more ominous and with blinkier lights."

Thanks for this, completely missed it... :)

In theory, every issuer will eventually default. Question is whether it is going to happen within the next 1, 5, 10 or 100 years.

What I'd like to know is at least what is their estimated probability of default and the time horizon of this forecast e.g. "There's a X% chance that Greece will default within the next Y months."

The CDS-implied probability of default (PD), as of July 8th, is roughly 41% within the next 12 months, 66% within the next 2 years, 93% within the next 5 and 100% within the next 10 years.

So yeah, they're very likely to default soon(ish).

I want to know about what this software predicts about whether the US government will default next month...

I stopped reading after glancing at the caption underneath the article's image:

"Wherever Senturion is housed it probably looks a lot like this, but more ominous and with blinkier lights."

That this statement is less than rigorous doesn't speak well for the remainder of the article.

The remainder of the article is equally tongue-in-cheek.

I wonder if the program was written by the guys behind the Patriot software... http://sydney.edu.au/engineering/it/~alum/patriot_bug.html

> It’s official: Greece is going to default, and there’s nothing anybody can do about it.

> That is, if you believe the word of a predictive software program

So that's how we define official now?

Blinkier lights. You can't explain that.

From the markets point of view, I think the large majority of investors have already assumed its going to happen.

Yes, from everything I have read it seems to be a foregone conclusion.

Yep, hence the need for intervention by other governments. The private sector will no longer finance their debts (at least not with an interest rate they can afford to pay).

You don't need military software for that. A bit of common sense and a healthy skepticism of what officials try to tell you is enough.

They can delay the inevitable, but not save Greece' finances. You don't go out of debt by taking on more debt while selling out your best assets at the same time.

What about "common sense" predicts Greek default?

How is that this software do not predicts USA default too? USA is in the same situation that Greece is and those Geniuses continue expending trillions in Irak and Afghanistan.

Greece, USA, Japan, UK, Spain, Italy... those countries are going to default one way or another because math is math and the magic of compound interest exponential growing sometimes do not happen. They could use massive inflation or face default, one way or another they are not going to pay what they indebt.

  > How is that this software do not predicts USA default too?

Because it's working for the USG, of course. Smart contractors don't bite the hand that feeds.

"USA is in the same situation that Greece is"

Because that's not true. Check the debt loads vs GDP, for one. There is a stark difference. The US is far better off.

The claims of any firm in the pay of the US military have to be taken with a grain of salt (this may be propaganda to undermine the EU). That being said, the prediction is likely correct.

Reply on news.ycombinator.com