Hacker News new | comments | show | ask | jobs | submit login
Competing Hypotheses - Made for the CIA. Now free to the public. (competinghypotheses.org)
180 points by _pius on Aug 25, 2010 | hide | past | web | favorite | 33 comments

Hi everyone. If you're interested in contributing code to the project, we'd love to have you. See the Developer FAQ at http://competinghypotheses.org/docs/Developer_FAQ.

Glad to see an open-source project that wants help and is written in PHP :)

I should be able to knock a couple items off your bug list. Thanks for the code!

Hello, can you please remove the OSX "dot underscore" meta files from the archive?

1. Beautiful site

2. To those who are interested in its background, you should read Heuer's original masterpiece on analysis (what this software package is based on I believe). It's very easy to read and elucidate many rationales you think is "obvious" but can't explain why.


Also, believe it or not I once wrote the CIA asking where I could find a paper copy of this book and they were nice enough to send me a copy free of charge

mass market book by one of his colleagues: http://www.amazon.com/Thinkers-Toolkit-Powerful-Techniques-P...

A way to implement these ideas reddit-style would be amazing.

Maybe a 2D reddit, so you can add (and vote on) evidence and hypotheses; and then also vote on each intersection (a choice of 5 levels of consistency). Not sure about nesting; would it mean sub-evidence/hyp, commentary, alternative expression, or related evidence/hyp. Perhaps users could work it out, in effect adding dynamic untyped fields, like twitter's @name, with voting for significance. Let the users decide! The UI would look messy, but collapsing fixes it.

To avoid group-think, you could do pseudo A/B-testing, with different votes. Though I don't think users would go for this - and when would the results become "official"? It becomes more stately (stateful); one of the beautiful simplicities of reddit is that the model types/components are constant.

No shortage of HN "hypotheses" too. :)

Hypothesis: Lisp/Python/Haskell/Ruby is the best programming language ever made. Evidence: PG recommends/PG criticizes; Libraries for X available/unavailable; Supports/Omits 1st class functions, etc. etc.

Hypothesis: You need a cofounder to succeed at your startup. Evidence: YCombinator funds/does-not-fund solo founder startups; Fail rate higher/lower for cofounder tech startups; etc. etc.

I really like this idea. I'm not sure how useful it would be but it would definitely be cool!

I love software like this: software that not only models data structures, like a spreadsheet, but attempts to model thought structures and patterns of reasoning, too. It's so much harder than building some dinky wizard that steps you through a series of fixed questions to answer. Instead it attempts to quantify the process of hypothesis and disputation in an open-ended way—yet an intuitive one, which attempts to interface with the ways that a professional will already think about a problem. I'd love to see more examples; I guess certain CRM/Project Management tools try to accomplish the same type of thing.

The software looks cool, and I totally get the demo scenario. I'm having trouble coming up with situations in which i would be useful to me, though. Do any of you all see other uses for it?

When it comes to your personal life, it's probably very limited. That's because most personal decisions involve preferences, and it's completely normal for you to be subjective when making such decisions. The point of ACH is to help you think objectively about objective facts, not keep you from thinking subjectively about things that are inherently subjective.

Argument-mapping software would be particularly useful for controversial topics where it's hard to judge just based on the appearance.

For example, take cryonics. An 'outside view' would be that it's largely scorned, associated with nutcases, etc.; but its partisans make an 'inside view' of specific technical arguments and say the facts overcome the outside view.

A mapped out inside view would be easier to follow for non-technical folks, and at least make clear what technical points are disputed. (eg. if the cryobiologists' counter-arguments are basically 'yes, it's possible except for one small issue with synaptical vesicles', an outsider might think that's soluble issue and decide to believe in the whole thing)

Of course, so far I know of no one who has put together a useful argument map like this (and all applications seem to be in niches - like the CIA), which sort of suggests that perhaps it's not a good tool at all or is too heavy-weight or something.

It seems to work well as a decision matrix. Try using it to choose your next car.

Investing based on macro-economic trends.

To clarify.

You could have (for example).

Chinese growth will continue. China will fall into a large recession. America will have high-inflation. America will have deflation.

Then as evidence you could take demographic data. Monetary policy decisions that are taken in the past & their reasoning...

I might have to do this.

Be sure that when you do, you choose hypotheses that are mutually exclusive: only one of them should be correct. http://competinghypotheses.org/docs/ACH,_Step_By_Step http://competinghypotheses.org/docs/User_manual#Hypotheses

I'd also love to see any real-world examples (without spies, maybe in startup/business/management context?).

If you have a scenario, I'd be happy to work with you to build it. We need more case studies, even if they're hypothetical. Write to me. http://www.competinghypotheses.org/docs/About_The_Project

We made a previous version of the software in collaboration with Richards Heuer that you can get for free from PARC:


The setup is a lot easier with that version if you just want to try it out. It is just a little java program. Unfortunately, I don't think it is open source.

Isn't this flawed in that it ignores Bayes' theorem? The "object was a UFO" conclusion from the video is a good example of this. As compelling as the evidence is for that hypothesis, it's still virtually certain to be untrue.

The best possible reason for liking this: the how-to video has three icons when explaining that the method was developed by the CIA for managing national security; the third icon is Godzilla rampaging among skyscrapers.

Oddly, I've been working on this project for years and didn't see the video (created internally at Howcast) until yesterday.

Actually I thought that this would be very useful for Medical diagnosis, and I vaguely recall a similar app being used thusly...

Very interesting possible uses for legal work.

I wonder if it would be acceptable in the US for a jury to use this software. For a judge to require a jury to use this software?

Unlikely and no way.

A jury is supposed to rely only on the evidence presented at trial and ruled admissible, no matter how obvious or compelling other evidence might seem. The only exceptions are for fundamental matters of common sense, like the fact that the sun does not shine brightly at night or that gravity usually causes things to fall down rather up or sideways.

Any kind of external research by a juror is right out - that's like bringing in additional testimony after the trial has taken place, and undermines a defendant's right to cross-examine and dispute the evidence. If an attorney's omission of some information seems like a glaring oversight, that's not the jury's problem. Maybe the client has an incompetent attorney (lots of cases involving that) or maybe the two sides agreed to a partial deal before the trial which included staying off that topic...there's no 'right to know' and juror's are not allowed to introduce any new information to others in the jury room. Any personal background knowledge which gives a juror extra insight, he is supposed to keep to himself. Usually that is considered juror misconduct.

Now here you're only looking at putting pieces of evidence into a matrix and scoring them, which seems totally neutral - but unless it's a very simple case and you can fit all the evidentiary arguments on there verbatim, then the losing side would probably claim the presentation of evidence in such simplistic fashion gave rise to bias.

A recent appeal of a murder conviction examined the use of software by a juror in depth. Although the judge concluded it was acceptable, it was only because the juror used the software alone and it simply served a note-taking function, rather than letting him manipulate or explore the paramaters in any novel way: http://scholar.google.com/scholar_case?case=1103407335837753... Scroll down to section B2 for the software use challenge.

In another case from this year, video playback software was questioned but ultimately authorized: http://cyb3rcrim3.blogspot.com/2010/03/jurors-experimentatio...

I think it will be a long time before any kind of data-organizing tool will be allowed. I am a bit skeptical about allowing even the frame-by-frame video playback in the second case, because not all video codecs are created equal and it's possible that artifacts in the display of a video image might distort it enough to sway a person's mind about whether it resembled a defendant (cf http://en.wikipedia.org/wiki/Pareidolia). IANAL, mind.

Usually a judge will err on the side of caution if any technological innovations are proposed, since being reversed for not conducting the rial properly is bad for one's future career. Please note, IANAL and this is an amateur opinion.

I do think there are some judges who would like to use tools like this for their own analysis of a complex case, without a jury. Richard Posner springs to mind, he's very much into Bayesian reasoning and formal methods.

Software aside, wouldn't the output of the software simply stand on its own as particularly insightful reasoning? What's to prevent a juror from using it privately and then mentioning the output to the rest of the jurors, who, if rational, would be hard pressed not to agree.

I think that would be OK for an individual, no different from using a notebook to organize one's thoughts. You'd still have the benefit of the systematic circumstance/condition comparison, although you'd lose the benefit of the voting aspects.

California criminal jury instructions say jurors can take notes, but cautions that they are only for individual use.

hmm.. explicitly uses myisam engine in mysql. Not good.

What would you suggest this be changed to?

Anything else?

PostgreSQL (with ADODB for PHP) would be my suggestion.

Applications are open for YC Winter 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact