Art Thou Bitchun?

krotty @ 2010-02-10T15:11:00

2010-02-10T23:11:34Z

"Permutation City" by Greg Egan

2009-05-25T08:29:44Z

This hit too close to home (jaanix = absolutely nothing):

"Optimizing anything to do with Copies is a subtle business. You must have heard about the billionaire recluse who wanted to run as fast as possible -- even though he never made contact with the outside world -- so he fed his own code into an optimizer. After analyzing it for a year, the optimizer reported THIS PROGRAM WILL PRODUCE NO OUTPUT, and spat out the optimized version -- which did precisely nothing."

In the news

2009-03-12T17:54:17Z

CNN Money on News Alpha.

Who Moved My Stock?

2008-10-22T22:42:37Z

Working on Who Moved My Stock?. Lets you figure out what was behind stock price change. Was it market or sector turning? Was it news breaking out? Is it something to worry about or was it totally expected?

What do you think? Need your feedback.

jaanix update

2008-07-17T08:16:13Z

Jaanix imports RSS from other sites, traffic more than doubled after adding reddit. What other feeds do you want to read through jaanix?

jaanix update

2008-06-19T21:36:33Z

Jaanix just got 300% better, lets you "tune" on people, got more stories, and doesn't look ugly any more. What do you think?

Lots of small design and usability improvements on jaanix

2008-04-01T02:18:06Z

The most important ones are that now you can tune pages with the help of the toolbar button (same old one, just more functionality) and that you can show / hide stories that you have seen with one click.

New "post to jaanix" widget is available too, this is how it looks:

post to jaanix

See my "tune" on jaanix

2008-03-11T01:20:33Z

Jaanix is a social news site that learns what matters to you. Once learned this "state of mind" can be shared with your friends. Here's my:

jaanix update

2008-02-14T06:05:40Z

jaanix got a brand new backend with lots of new features:

same url posts - aggregated, but you can always have your own version saved and edited

individual stats per post - see how many people you've reached with your post

new system sliders - you can control more aspects of the recommendation algorithm

customizing list of topics - you can list topics as your personal interests to tune and highlight them under posts

tagging of any post - if you do not like how somebody has tagged a post but do not care enough to save it you can just tag it

no registration required - new users get access to all the features of the system without registration (and registration just got easier too)

new post button - combined all the tabs into one, please update your bookmarks in tools, old url will be gone soon

new urls for topics - topics have urls of their own, works in your personal jaanix too - you.jaanix.com/topic

import of links from del.icio.us - the backend is more scalable, high quality links are welcome

new account settings - hide posts, control image preview size or hide image previews completely for more even layout

switched from rss to atom - please update your subscription, if you are using your browser as a reader - feeds are personalized now

inline editing of tags, comments, posts - more ajax for seamless and efforless editing

send your feedback to feedback@jaanix.com or contact "jaanix" on skype.

krotty @ 2008-01-19T11:52:00

2008-01-19T19:54:41Z

Notice that peaks are at 60%, 70%, 75%, 80%, 85%, 90%, 95%. The peak at 65% is less visible, but is still there.

krotty @ 2007-12-14T23:09:00

2007-12-15T07:14:03Z

Why are you still here? Go see jaanix now. Let me know if something doesn't work for you. The quality of links is amazing, for example check out Russia on jaanix.

A Model for the Observed Election Data

2007-12-09T07:30:22Z

The model of election adopted from slobin:

There is one control parameter X - the degree of control over election outcome, that varies from region to region. It includes control over the minds of the people with propaganda, forced attendance, forced vote for a given party, a ballot filled for a person by somebody else, or a ballot posted for a person that never came to elections. The end result is the same - X% of the voters will vote for our party no matter what. This parameter varies from region to region. In some places it is 100%.

The rest of the population (1-X) % are not affected by this and vote the way they want. Most of them just stay out of the whole thing for obvious reasons, but say 50% of them actually go and vote. The distribution of their votes across different parties doesn't change much in different regions.

The important point is that given only the results there is no way to look inside of X and say how specifically this X was achieved.

Thanks Everybody

2007-12-08T09:12:19Z

In case you've missed: tons of great comments on Slashdot on who's to blame and what should be done about Soviet Russia.

Detailed: Putin's Party Share and Invalid Ballots vs. Turnout (%)

2007-12-08T08:56:54Z

Source

Lies, Damn Lies, and Statistics

2007-12-07T20:10:25Z

A lot of you have asked me if there's an explanation for this graph and if it proves or disproves something.

As usual with statistics a single argument will not prove or disprove anything by itself, this is not logic where if you find a contradiction - you're done, or if you've proved something, you've proved it no matter what other arguments you can find.

This graph is the single most convincing graphical argument that I can find that "proves" that the elections were not free and that that they were rigged. Of course there are other arguments for and against that. If you follow the news will will find plenty of them, and you have to find them all, combine them with your prior assumptions to come up with your own conclusions.

As for the actual method of how this was done, I'm not so convinced that it was just ballot stuffing and nothing else. It seems to me that different portions of the graph show different methods of fraud, and it is just a coincidence that they all lined up like that.

The top 80-100% turnover portion is the most clear example when people have no voice at all, like in the soviet union. No matter if they showed up because they were forced to, or they didn't show up and their voices were counted anyway. Chechnya and Ingushetia had almost 0% real turnout and they've reported ~90% - ballot stuffing at its worst.

The dark left cluster shows a different story to me - a story of coercion and massive propaganda. People were forced to attend, people were forced to show their vote, people were given the idea that there is no other choice. And everybody not voting for the main line was discouraged to attend - there's no reason to go if your vote will end up counted against your will because of the 7% barrier (if you voted for a party that didn't pass it, your vote will go to the top contenders proportionally). Ballot stuffing is still there, but it probably plays a minor role here.

Party's Share vs. Voter Turnout (%) in 2007 Parlament Elections in Russia

2007-12-07T04:09:08Z

Guess which party is Putin's (hint - correlation coefficient is 90%). X axis: percent voter turnout at a given territory, Y axis: percent voted for a given party. Each dot is an election territory.

via russian blogs: original data, explanation?.

Update: my interpretation

Update 2: Uploaded to Swivel:

Update 3: For comparison - the same graph for 2006 elections in Canada:

Update 4: Much more detailed version:

Update 5: A simple model of elections that explains the data.

Update 6: Final election results are available, now if you plot party share out of all registered voters vs turnout you get 97% correlation:

Update 7: Excellent slashdot discussion

Update 8: More evidence via

x axis - turnout percent, y axis - number of elections commissions reporting this turnout, notice the peaks are exactly on the round numbers.

privacy - shrivacy

2007-12-06T19:39:32Z

Are you still using livejournal? Do you care about your privacy? (In Soviet Russia) SUP is already tracking your visits and is showing this information to the author of the journal you've visited. Currently there's no way to turn this "feature" off.

krotty @ 2007-12-03T08:13:00

2007-12-03T16:19:17Z

jaanix status:

comments - working,
rss - working,
tags equalizer - working

todo:

comment ratings
people finder
stats

overall:

posting on jaanix is fun
finding other stuff on jaanix is fun
people discovering your stuff is fun

What are you waiting for?

krotty @ 2007-11-24T12:28:00

2007-11-24T20:42:34Z

Once jaanix has comments, I'll be out of livejournal. You can follow me on my jaanix blog, or official jaanix blog

krotty @ 2007-11-18T18:22:00

2007-11-19T02:51:09Z

jaanix.com update:

Added delete account option, now you can wipe out all your posts and ratings

Added tuning in the personal blogs, now you can filter based on the author and still be able to tune the stories.

krotty @ 2007-11-16T01:23:00

2007-11-16T09:28:21Z

jaanix is fast again, give it a try - no registration required to tune the web

jaanix.com update

2007-11-14T03:26:47Z

Next rev is up and running (response time is slow, sorry), no tune-in or registration required to use it.

jaanix - your personal internet tuner: with a twist of a few knobs sort and filter the huge onslaught of stuff on internet demanding your attention.

jaanix update

2007-11-02T21:03:18Z

added topic tuner on the right, let me know how it works for you.

krotty @ 2007-11-01T13:28:00

2007-11-01T20:40:06Z

It is 10x easier to post on jaanix, especially with images and video, and now I am posting things 10x faster than I ever did with livejournal.

Scary thought, if this happens with others too, how I'm going to keep up with that? Recommender algorithm will help, but still...

jaanix update

2007-11-01T02:42:33Z

Added rss feeds, now you friends that do not use jaanix can subscribe to your jaanix mini-blog.