h a l f b a k e r ycarpe demi
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
The Dubit
Unversal measurement of certainty/believability | |
I think it would be nice to quantify how believable (or, conversely, how questionable) a statement, idea, etc. is -- hence, the dubit. The more questionable something is, the more dubits -- something that is a known certainty (e.g. "You will die someday") would rate zero dubits.
What are the chances?
http://www.halfbake..._20the_20chances_3f
related idea [krelnik, Oct 17 2004]
[link]
|
| |
something to ponder on, yes, mr/ms/miss banana. |
|
| |
I may rename my bullshit meter - my dubit meter. |
|
| |
In Sweden I talked to Ms. Inger,
Who said her elbow to long finger
Was a unit called cubit.
Didn't doubt it a dubit,
Her arms were like Notre Dame's bell ringer. |
|
| |
Then there was his twin brother Pseudomodo who fell to his death, a dead ringer. |
|
| |
I don't think you can quantify a completely subjective estimation. It's all about how credulous the reader / listener is... |
|
| |
I think a better conversion function is -log2(p). This has the advantage that if you're x dubious about A happening, and y dubious about B happening, then you're x+y dubious about both of them happening, which seems intuitively correct. |
|
| |
Why log2 instead of ln? Because that way a dubit is basically a bit of information. "This fair coin will come up heads" rates exactly one dubit. |
|
| |
The concept is "baked" in information theory (handwavingly, if an event is 100% certain, then if it happens you get 0 bits of information; if an event is extremely unlikely, then knowing that it happened is a lot of information), though I'm not sure if I could come up with a single name that means the same thing. Possibly "bit of entropy". |
|
| |