cite as: F. Bi. 2008. An algorithm based on AlGorithms. *Intl. J. Inact.*, 1:5–6

Via Deltoid (indirectly), a blog post on Rabett Run about these things called AlGorithms:

There is a long history of Al Gore being right on many issues and being ridiculed for it. Such things are known as AlGorithms and are recited at night by good little girls and boys in the right households.

An interesting question to ask is this: is it possible to use AlGorithms as a basis for algorithms for doing stuff — like, say, proving various propositions about the real world (say, about global warming)? Here’s a possible first attempt at such a thing:

**procedure***DecideWrongly*(*P*:*Proposition*,*G*:*Goracle*)- (*
*P*is a proposition, for example, “the earth is warming” *) - (*
*G*is, of course the Goracle *) - (* we want to know whether
*P*is true or false based only on the information provided by the Goracle *) **var***S*:**set**of*Statement*;*m*,*n*:**integer****begin**- . . .
*S*:= all statements ever made by*G*about*P* - . . .
*m*:= number of statements in*S*which affirm*P* - . . .
*n*:= number of statements in*S*which deny*P* - . . .
**if***m*>*n***then** - . . . . . .
**output**“*P*is true” - . . .
**else if***m*<*n***then** - . . . . . .
**output**“*P*is false” - . . .
**else** - . . . . . .
**output**“don’t know” **end**

Unfortunately, sometimes Gore does make erroneous statements without correcting them in time, so if for example *P* happens to be the “frogs in boiling water” thing, the above procedure will give erroneous results. To fix this, a possible idea is to make the procedure affirm only statements which Gore has stressed on several times (say, 10). This gives us

**procedure***DecideNotSoWrongly*(*P*:*Proposition*,*G*:*Goracle*)**var***S*:**set**of*Statement*;*m*,*n*:**integer****begin**- . . .
*S*:= all statements ever made by*G*about*P* - . . .
*m*:= number of statements in*S*which affirm*P* - . . .
*n*:= number of statements in*S*which deny*P* - . . .
**if***m*–*n*> 10**then** - . . . . . .
**output**“*P*is true” - . . .
**else if***m*–*n*< -10**then** - . . . . . .
**output**“*P*is false” - . . .
**else** - . . . . . .
**output**“stop being a lazy bum, just do your own research already” **end**

Now, the only thing that’s lacking is a proof of correctness for this algorithm. Unfortunately, since Gore hasn’t (yet) made any statements about it yet, therefore according to the algorithm, we don’t know whether the algorithm is always correct or not. However, given that Gore is correct most of the time, and the algorithm can’t make any more mistakes than Gore does, it follows that the algorithm’s accuracy must be quite good indeed.

Related reading: Paul Krugman on the Gore Derangement Syndrome.

ARrrrrrrrghhhhhhhhhhhhhhhhhhhhhhhh!!

Comment by Eli Rabett — 2008/04/24 @ 17:28 |

Um yes, I still need to work on the correctness proof. ðŸ™‚

Comment by frankbi — 2008/04/24 @ 17:44 |