Monday March 31, 2008
I am thoroughly enjoying Deepayan Sarkar’s book on Lattice, R’s trellis lookalike.
Trellis was one of those things that I thought looked cool when it came along, but with high start-up costs, so I never took the plunge, preferring lots of custom code for graphs of various flavors.
This book is full of nice examples, that I am very much looking forward to implementing in my own work.
Sunday March 30, 2008
Samuel Arbesman and Steven Strogatz at Cornell have been looking at DiMaggio’s 56 game heating streak with a ton of Monte Carlo simulations, resulting in a piece in the New York Times.
They create a population of coins (players), each with a success probability equal to each player’s batting average (or hits per game I think), on a season-by-season basis. In that diverse population of coins (players), how often do you see a run of 56 heads, or longer? More specifically, how often is the longest streak in any simulated history of baseball longer than the 56 game streak recorded by DiMaggio? Answer, according to the authors:
….streaks of 56 games or longer are not at all an unusual occurrence. Forty-two percent of the simulated baseball histories have a streak of DiMaggioâ€™s length or longer.
Well, ok, except that if streaks of that length that are “not at all unusual”, then why have we only seen one of them? That is, I wonder if the appropriate conclusion here is that the model underlying the simulation is very unrealistic, if it is putting such high probability on events that we’ve never seen?
Thursday March 27, 2008
I am enjoying Skim. Bye bye Texniscope; it hadn’t been updated in a long time, seems a dead project. The Skim wiki has all the details on pdfsync integration.
I’ve been doing some heavy duty R coding for a bunch of projects on the boil at the moment: this is what Spring Break is for, right? Each of these projects involves tossing data from R to JAGS. I am very much enjoying the way Martyn Plummer has got his rjags package handling the go-between: JAGS now runs from R, as it were, with JAGS being able to see objects in your R environment. This takes some of the pain out of MCMC. I think of this as like rbugs, but better (because it works natively on a Mac, because its JAGS, etc).
I upgraded to Leopard a while and in so doing clobbered the very nifty Mail appetizer plug-in. I found some instructions for reviving it in Leopard, but all that did was provide me a the pop-up shaded preview window, with malformed previews and no functionality from the button set. Now it looks like its back. I’m about to give it a whirl. Some discussion here at versiontracker. Apple should hire this guy (Stefan SchÃ¼ÃŸler) and make this plugin part of Mail.app for real.
I have also replaced letterbox with Widemail.
Saturday March 22, 2008
And perhaps the faculty too. See these photos.
Its been a glorious start to Spring here in Northern California. We celebrated the end of Winter quarter with a Happy Hour out on the lawn in front of our building.
An addition to Apple’s OS/X that I think came in with Leopard. Nice touch. You really don’t your laptop battery going dead in the middle of a software update.
Friday March 21, 2008
Spotted in an Annabel Crabb piece in the weekend Sydney Morning Herald:
In a Sydney Town Hall debate in 2000, Gough Whitlam explained that he had never dropped the “C-bomb” in Parliament, despite his widely recognised capacity for colourful language.
“The nearest I came to doing so was when Sir Winton Turnbull, a member of the cavalleria rusticana, was raving and ranting on the adjournment and shouted ‘I am a country member’,” the former prime minister recalled.
“I interjected ‘I remember’. He could not understand why, for the first time in all the years he had been speaking in the house, there was instant and loud applause from both sides.”
Wednesday March 5, 2008
This started out as a comment on my citizen Kate post. But it is worth a separate thread.
Adrian writes [inter alia]:
hereâ€™s a question Iâ€™ve been itching to ask. Where, oh where are the American versions of Antony Green?
This question solidified when watching the Texas results dribble in. In particular:
* Everyone seems to say the Texas Democratsâ€™ primary is horribly complex, and us mortals canâ€™t possibly understand it. Itâ€™s not that difficult, is it? Itâ€™s just a two thirds primary and one third caucus with a couple of minor tweaks in way delegate numbers are allocated to districts, isnâ€™t it?
* All the election coverage Iâ€™ve seen is obsessed with awarding states to various candidates, with an almost cavalier disregard for the delegate count. Yeah, I know that state counts are the only thing that matters for the Electoral College and Presidential election, but why arenâ€™t the big media sites reporting delegate counts for the primaries?
* The media sites were agonizingly slow at declaring the winners of various states. Did they really need to wait until 80% of the vote was counted before making their bold Ohio predictions? Couldnâ€™t an Antony have could have done better?
In Australian elections, Iâ€™m so used to seeing Antonyâ€™s projections and being able to see the raw data on the AEC web site. Is there anything similar in America? Am I asking the wrong questions?
Cute shtick. And she’s using my photos.
And I did another stint on TWT yesterday (“no doubt about it”).