h a l f b a k e r yNeural Knotwork
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,
|
|
|
Being bullishly insecure, I just checked
my
profile page and discovered, to my smug
glee, that I was in a net croissant
situation
- taking the average across my paltry six
inventions.
Imagine my crestfalling, then, when I
discovered that almost *all* halfbakers
are in a positive patisserie
situation:
croissants outnumber fishbones by
something like 2:1 overall.
Now, this is great and it means we are all
brilliant and positive people, but it also
means that, on average, our ideas are
better than average. I am not sure where
this will lead, but no good will come of it.
I therefore tentatively suggest that the
threshold for croissant/fishbone
imparting
should be dynamically coupled to the
voting pattern, overall, such that the
average result across all ideas is neutral
(neither fish nor flakiness).
Croissignificance_2...voting_20technology
My similar idea; yours lumps the halfbakery element into one "big" user. [phundug, May 14 2007]
[link]
|
| |
Hm. Much more accurate, methinks, to take the inverse square root of the logarithms of the net votes, minus the checksum of the annotations. |
|
| |
That would be very silly. Imagine if there
were zero votes. Think, for heaven's sake
man. |
|
| |
Looks to me like you need some bones to help balance you out. Here you go! [-] |
|
| |
Thanks, 'ear hobo. Can't beat a little
pragmatism. |
|
| |
[-] I'll contribute, too. (saw the word *normal*) |
|
| |
Thanks one and all. I feel there may be
some mis-understanding of the word
"renormalization" here, but I guess half
an understanding is better than no bread. |
|
| |
A quick look shows 5 out of your 7 ideas are bunned. If you were to delete 3 of these you would be in perfect harmony with the halbakery universe you propose. |
|
| |
This is getting very much out of hand, I
fear. And alas it would not address the
problem for the entire Bakerverse, only for
the Buchanananiverse. |
|
| |
A true leader leads by example. |
|
| |
//Bakerverse, only for the
Buchanananiverse.// |
|
| |
A + for the laugh, but a - for the
idea=net gain/loss of nothing. Which is
somehow more important than
something, at times. Depending upon
your audience, of course. |
|
| |
<pedant> You will be unable to renormalise the bakery unless someone has an infinite amount of buns that can be cancelled out against someone else's infinite amount of bones. </p> |
|
| |
expert advice from [wagster]
"Get thee to a bonery" |
|
| |
//A true leader leads by example// |
|
| |
This week, I am delegating leadership. |
|
| |
//unable to renormalise unless...an
infinite
amount of buns// True in some
contexts, especially where
renormalisation is used to provide a
finite answer. However, in this case, we
are aiming for a zero sum - I believe
this may make it possible to
renormalize sub-infinite opposing
quantities. |
|
| |
There are many ways to normalize data. For example: |
|
| |
Total Buns for you / Total buns for whoever has the most |
|
| |
This ratio would be one for one person, and less than one for all others. The distribution of ratios would be gaussian between 0 and 1. Or not, hmm. |
|
| |
//The...would be....or not// A syntax
for all seasons. |
|
| |
Relativistically, of course, there is no
fixed frame of reference. This is why I
keep spilling my drink. |
|
| |
It makes most sense, therefore, to take
(as a reference) either the fixed stars, or
the average of the entire bakerverse. |
|
| |
This started out as a simple concept.
Thanks to all those amongst you who,
by dint of sterling efforts, have made it
difficult. |
|
| |
I feel entitled to a say in how normalization is defined. I am currently engaged in some influential work in the field. |
|
| |
Hang on: I've got an engine for that... |
|
| |
...Anyhow, wouldn't this idea be implemented more simply if [MaxwellBuchanan] just sat perpendicular to his profile page? |
|
| |
Not to mention the hopes that someone had invented a way to make us like other people. |
|
| |
Most of my [+] votes were lost in the Great Crash of September '04... |
|
| |
//if [MaxwellBuchanan] just sat
perpendicular to his profile page// I have
an LCD screen. Therefore, if I sit
perpendicularly to it, my votes (and indeed
everything else) vanish from view. This is
therefore an instance of the "Schrodinger
Renormalization Problem". But with fewer
cats obviously. |
|
| |
[hippo] I have always felt bad for those who lost +votes in the crash, but then again there is a most inspirational aura surrounding those ideas that I truly respect and always see them as 2 1/2 bun ideas even if they never were. |
|
| |
//and always see them as 2 1/2 bun ideas even if they never were// |
|
| |
It always amuses me that despite the perfectly normal obsession with collecting votes, people here are even more obsessed with the statistical methods of handling the votes. Then again, I am easily amused. |
|
| |
mb: sorry to take so long to get back to you, but what could be more half-baked than a voting system that results in values of log(0) ? |
|
| |
Dr C - well, OK. But I prefer ln(0) rather
than log(0). |
|
| |
Just out of curiosity, fellahs, if I'd
bemoaned the devaluation of the fishbone
would you all have chipped in with a
helpful bun? |
|
| |
//Wouldn't renormalisation just be a
different view of the voting actually
collected ?// Exactly so, [Bigs]. The
original aim was not to in any way alter
the actual votes cast, but rather to
dynamically shift the fishbone/croissant
transition point in response to the
current average. |
|
| |
Hence, if the Bakerverse is in a generally
benevolent mood and is awarding more
+ than - votes, one might find that an
idea with a net vote of +1 would be
below the average, and hence would
warrant a fishbone. Conversely, if the
Bakers age and become a bunch of
grumpy old men who habitually award
more - than + votes, the threshold
would shift such that an idea with a net
vote of -1 might, by being relatively
good, find itself sporting a croissant. |
|
| |
I can't help but feel that the concept is a
lot simpler than people seem to be
finding it. |
|
| |
I often delete my fishboned ideas, unless I strongly believe in their validity. I don't want to clutter up the halfbakery with crap. Maybe other posters do this too. |
|
| |
Furthermore, some bad ideas get [m-f-d] and delete themselves. This leaves a net surplus of croissants. |
|
| |
It's hard to tell how many votes *cast* were croissants and how many were fishbones. I do wonder. We could possibly display it on each user's profile. But, I think the number of croissanted ideas the person was able to come up with is evidence enough. |
|
| |
I guess crap is in the eye of the beholder, because I often find myself casting bones on ideas others adore. |
|
| |
Deleting one's own fishy ideas makes the Bakerverse a poorer place. Any idea can serve as an inspiration, and some of the more heavily boned ones do this best. Also, it seems to me that the fishiest ideas are often some of the most heavily annotated. |
|
| |
Like this one, for instance. |
|
| |
and Vagina Jam. Can't forget Vagina Jam. Anything that gets 170 votes is worth keeping. |
|
| |
I love the way that this idea is single-handedly renormalising the bakery. Not that this idea is that bad... but it does seem to be the solution to the problem it poses. Is that zen? |
|
| |
If the Halfbakery *was* normalized, wouldn't we all have to go somewhere else? |
|
| |
Can't help you with that one. Every place I go is normalized. |
|
| |
I'm sure it would be possible to calculate the total bun/bone ratio using views. Once you have that ratio you can easily 'renormalise' vote counts, either individually or by dynamically interpreting results returned by the bakery. |
|
| |
I'm damned if I can be bothered. |
|
| |
A not-so-quick analysis shows that the current bun/bone ratio is 18,792/5,755 (3.265/1). |
|
| |
This shows that either [MaxwellBuchanan] is correct, and that we are on the whole all brilliant and positive people OR that we are a self-deluded bunch with prescious little ability to discriminate between stupid and really stupid ideas (assuming of course that there is a difference). |
|
| |
The other thing this shows is that a bone is worth more than three buns, so one should be proud and honored to collect these valuable treats. |
|
| |
[Jutta], is the number of votes required to generate a whole fish the same as the number of votes per whole croissant? Or is the scale asymmetric? |
|
| |
//I'm damned if I can be bothered.//
I'm damned if I can blame you. |
|
| |
//any stats for all ideas discounting
those containing the words - custard,
trebuchet or pirate ?// You could be
onto something here. It should be
possible to datamine the bakery and
obtain, in effect, a bun quotient for
every word on the site. |
|
| |
Obviously, "custard" and "trebuchet"
would have a high BQ, since ideas
containing them are frequently bunned.
Words like "the" and "eventual" will have
a BQ close to zero (at least after
renormalisation; these words are no
more likely to be associated with a
bunned idea than with a boned one).
Words like "renormalization", however,
would have negative BQ's, as they are
associated predominantly with boned
ideas. |
|
| |
Having derived the BQ of every word, it
should then be a trivial matter to
generate an uber-bunnable idea,
essentially by selecting only those
words with positive BQs (plus a few
neutral "the"s and "and"s). |
|
| |
In a preferred embodiment of this idea,
hidden Markov models are used to
further optimise the choice and order of
high-BQ words. |
|
| |
I think the problem lies in that each individual vote can only have three states: positive, negative, or neutral. Thus, bun status is not really a measure of quality, simply a measure that states that an idea is good rather than bad. More buns does not mean "really good idea", it means "really popular idea that is also somewhat good". |
|
| |
As such, the nun imbalance doesn't mean we all have fantastic ideas--just that we have more good ideas than bad ones. Factor in that many people delete their really bad ideas, removing bones from the system, and a bun surplus seems fine to me. |
|
| |
I was tempted to make a pun here, but I'm trying to get out of the habit. |
|
| |
//bun status is not really a measure of quality, simply a measure that states that an idea is good rather than bad// |
|
| |
I don't equate bun status with the quality of an idea itself but rather with the enjoyment it delivers to halfbakers. A well-bunned idea may be a good idea, but if it is truly halfbaked then it can never be good in a practical sense. Thus a high BQ imples an idea that a majority of voting halfbakers approve of for other reasons, such as humor, irony, presentation, etc., assuming of course that halfbakers in general agree with this precept. |
|
| |
[imaginality] gets the prize! I actually noticed that typo, but was so amused I left it in to see if anyone would notice. |
|
| |
However, the prize is revoked for the habit pun. >:-) |
|
| |
This from the help file: "Positive and negative votes are summed up, compared with the votes for all other inventions, and normalized into a numeric range of -2.5 to +2.5." |
|
| |
I don't understand how this idea is any different from what already happens. |
|
| |
How did you come across your findings Maxwell? If there is a majority of people with a bun surplus*, might there be a bony minority of people who are making it look as though you have skewed results? |
|
| |
*there is the question of how sensible it is to measure the bun/bone ration at the user level, rather than at the individual idea level - after all, it's the ideas that attract the articles, not the users. |
|
| |
//Positive and negative votes ...
normalized ....// |
|
| |
Aha. Yes. Well, I think my idea still
stands, but with the caveat that it is now
totally redundant. |
|
| |
The votes are normalized for each half separately; I thought you wanted to normalize them together. (So that, if an idea has 1 vote in favor, but the average idea has 2 votes in favor, that +1 idea would display with half a fishbone.) |
|
| |
Another way to normalize would be to automatically apply 3.265 bones to every idea. |
|
| |
//The votes are normalized for each half
separately// Aha! In which case I am less
redundant than I thought I was. Yes, I was
mooting (wanting is maybe too strong a
word, in the cold light of day) an overall
normalization, along the lines you
described. |
|
| |
[MaxwellBuchanan], if you collect a few more bones with this idea, you may succeed in accomplishing your objective without the need for administrative action. |
|
| |
Let's try halfnormal rebakerisation instead - it sounds more interesting. |
|
| |
I did not read the annotations yet, but this art-tickle made me laugh; I really couldn't help it :) |
|
| |
Data. I want rock hard, throbbing data. For example, I see [Antegrity] has resurfaced - no fear of the bone there. Old [myclob] is another who springs to mind as an ossuary keeper. [Max], I think you need to take a random sampling of HB users and determine exactly what the bone and bun count is. |
|
| |
Write up a nifty little robot to do this, because if robotized it can actually be comprehensive instead of just a sample. |
|
| |
Report back with that diamond-hard data. |
|
| |