Archive for July, 2009

Simulation of truncated normal variables [reprint]

Posted in Statistics with tags , , on July 24, 2009 by xi'an

As I do get on a very regular basis emailed requests for reprints of my 1995 Statistics and Computing paper “Simulation of truncated normal variables”, I decided to put a reprint of the original version on arXiv. As is (or was), i.e., in the TEX format of 1992… I take the opportunity, though, to recall here that a fundamentally identical solution was proposed by John Geweke in the Proceedings of the 23rd Symposium in the Interface in 1991. Although I was unaware of this paper until John pointed it out to me, it is the one deserving the citation.

Harry Potter and the half-blood prince [debate]

Posted in Books, Kids with tags , , on July 23, 2009 by xi'an

“The film is great, but it was made quickly and without true feelings. But it is a series, so it has good points.” Rachel

I watched the sixth Harry Potter movie in the series with my teenage kids last afternoon and I came back rather disappointed. My daughter does disagree and so here are our opposite views, written together.

On the [lack of] depth of the plot (in the movie rather than the book), we concur that (new) secondary characters are missing (and even a major character like Voldemort only appears through a few childhood memories), that the surroundings have shrunk to three or four places, that the little snapshots on the “usual” life of magicians that made the earlier movies so fun are missing (except for Fred and George’s Weasley’s Wizard Wheezes joke shop), that the students’ studies at Hogwart are almost inexistent, with much less students visible through the movie, and that the story is proceeding at a fairly chaotic pace, lacking a driving line. The central role of the half-blood prince’s book is clear at the beginning of the movie but it disappears too quickly with no clear explanation, except at the end when Rogue claims his crux back. This also makes the train scenes amazingly long given that they involve hardly any action (even though the views of Rannoch Moor are stunning).

The major characters are alas far from convincing in my opinion, including Harry himself who is mostly one-dimensional (with a surprising limited range of expressions and an impossibility to communicate with his friends) but who appears in about every scene, while my daughter thinks Ron and Hermione played well. We both agreed that Draco Malfoy may be the best role in the movie, with much more depth than others. Dumbledore is also mostly convincing, as well as the new professor, Horace Slughorn, especially in his armchair disguise (my opinion!). The obligatory Quidditch scene is rather well-set, although Ron turning into a champion thanks to a placebo is too unrealistic, even for a magician. My daughter enjoyed the special effects surrounding the flight of Ron and Ginny in the field, I did not. My own preferred scene may be the early one in the subway dinner when Harry is treated as a “normal” teenager…

The central issue is the prevalence of the “love scenes” over the action scenes. I find them simply unbearable because poorly managed and badly acted, while my daughter considers the main drawback to be less magic and less action. Both of us think that the relation between Lavender and Ron is completely botched, especially the breakup scene in the infirmary. (In my view, Ron confirms his total lack of acting abilities from previous movies!) And my daughter does not see the point in the magic birds being thrown by Hermione at Ron!

So, overall, maybe not the worst Harry Potter in the series (two was really bad), but a very unexciting one, way below the book

Little bonsai [day 66]

Posted in Kids on July 22, 2009 by xi'an


The outdoor adventures of the single remaining little bonsai are not very promising. As it stayed outside in the shade during our week away in Aosta, it got drenched by heavy rains and some branches got broken… Maybe we should cut them clean. Or buy a fully grown bonsai!

Bayesian Data Analysis for Ecologists [reflections]

Posted in Books, Statistics, University life with tags , , , , on July 21, 2009 by xi'an

The course I gave last week in the Gran Paradiso National Park was certainly one of the most exciting I ever gave! And not only because of the paradisiac location. Indeed, the young twenty ecologists/biologists/geneticists who attended the course were unbelievably motivated to learn about Bayesian Statistics. They did come to the course with a (strong) purpose and with clear problems and real datasets as well.They were thus ready to endure some of my most theoretical slides to obtain indications for progressing in their own work. Another great point was the repeated will to get beyond black box solutions, even Bayesian black box solutions, to understand the available softwares and possibly to modify them. Towards this goal, I slightly modified the way I usually teach Bayesian Core, in order to replace the standard datasets with three “local” datasets obtained from the park biologist, Achaz von Hardenberg (but not available publicly). Here are the modified slides:

which will mostly be useful to those who attended the course. I am thus very grateful to Achaz von Hardenberg, from the Alpine Wildlife Research Centre, Gran Paradiso National Park, and to Antonietta Mira, from the University of Insubria, who invited me to give this course. My only regret is that we could not cover the types of problems met by the attendees more in depth. Given their diversity and richness, this is rather frustrating! It is thus most likely that we will have a follow-up course in a near future with the same participants based on case studies only, where we can study those problems in thematic groups.

Summit fever

Posted in Mountains, Running with tags , on July 20, 2009 by xi'an

Frozen gloves in Vittorio SellaIn connection with the climb of last Saturday, I have had trouble sleeping for three nights: the night before starting, from sheer excitement at the prospect of climbing, the night at the refuge, from noisy neighbours to, once again, sheer nerves on the proximity of the climb and the need to wake up at 1pm, and the last night, from repeating the climb over and over and wondering how much astray alternatives would have gone, and from being back at sea level. (I could somehow feel the weight of those additional 1500 meters of air!) I still waited till 7am to go running, with a terrific impact from spending seven days above 1500m on my breathing!

La Grivola and the Trajo glacier (C.) B. Plessis

This is usually the case for most of my mountain outings so I do not think this is of particular importance, as the psychic involvement represented by an alpine climb is quite different from the requirement of a half-marathon or even of rock-climbing on a cliff. It is not as much the objective dangers of the mountains as the tension resulting from concentrating on every move for a long while and from repeatedly forcing one’s body into unusual positions, like bending ankles to grip the ice with all spikes of the crampons or walking down snow slopes in a duck-like manner, bending forward in order to avoid turning into the ultimate human sledge… The intensity and duration of this commitment explain quite easily why the brain cannot let immediately go, once the “game is over”.

Aosta [hike 3: no Grivola!]

Posted in Mountains, Travel with tags , , on July 19, 2009 by xi'an

pict6607My last lecture over in Cogne, I “ran” up to Refugio Vittoria Sella to meet my guide, to whom I had been introduced by the organiser of the course, Achaz von Hardenberg. This guide happened to be the very impressive mountaineer Abele Blanc, who is one of the twenty men in the World to have climbed thirteen of the fourteen 8000’s. I thus felt a bit shy at wasting his time with my low level climbing experience but he was both very professional and very patient in his guiding. Meeting him was an remarkable moment, for his intense personality is of a quality I rarely met. Climbing with Abele, albeit too briefly, was thus the experience of a lifetime, even though we could not meet our climbing goal, and I can only wish I have the opportunity to repeat this experience in the future…

Col de la Noire

The climb (la course) we intended to make was La Grivola, an almost 4000’s (3969m) whose sharp features are quite attractive and whose (low) difficulty was within my abilities. However, a storm built up on most of the Alps on Friday afternoon and it stroke during the night. We had intended to leave the refuge at 1am, but snow was falling too heavily for us to get out then and we thus started close to 6am when the snow had abated and daylight made handling the more delicate passages possible. (It thus allowed us to get at least some sleep, for some excursioners have had a good time in the hall of the refuge till midnight!) In the first hour, we saw a group of seven chamois, including four yearlings, who had gone down to the altitude of the path because of the snow. The first delicate passage was about climbing the snow corridor visible on the above picture (taken on the previous afternoon), to Colle de la Nera (3491m). There was enough iced snow there to make climbing easy with crampons and one ice-pick, even though we had to rope up for the final steep bit of maybe 50-60 degrees. Once at the pass—it was then 8:30am and we had made a fairly good time so far—, the weather had not improved, with gusts of winds sending snow powder inside every crack of our clothes and bags—as on Ben Nevis in 2006, my Lowe Alpine jacket hood would not close!—and a very poor visibility on the Trajo Glacier. Abele took a look at the glacier and decided that crossing it would be too dangerous for crevices were covered with fresh snow and impossible to detect—unless too late! He then took me to La Punta Nera (3683m), visible on the left of the above picture, that was an easy climb up packed snow, for there is an alternative route to La Grivola that goes through La Punta Bianca with very little differential in altitude (compared with going down on the glacier). Unfortunately the windy conditions were such that attempting this technical ridge walk was also impossible…

Colle de la Nera

We thus went down to the pass, hoping for a miracle for there still was time to cross the glacier in good visibility conditions and to reach La Grivola in about two hours. As miracles rarely happen, Abele then decided to try for another peak on the other side of the pass, La Punta Rossa, to make the most of the day, but after fifteen minutes of walking on the side of the glacier, we hit bare ice where our mixed climbing crampons would not grip, as I quickly found with my first steps… We could thus only head back down to the bottom of the valley. Going down the snow corridor proved to be the hardest part of the climb as, after the 20 meters of steep hard snow, we hit recent snow covering rubble and loose slates that made my going very very slow. It is only when we hit the end of the snow line, at about 10:30am, that the weather started to lift, as seen on the above picture, but it would have been too late anyway. So, after a cup of tea at the refuge—above which we saw a lone chamois—, we rushed back to Valnontey, our starting point, that we reached at 1:30…

From Refugio Vittorio Sella

It is always a disappointment to have to turn back on a climb, but the weather and the difficulty always are the final judges! In the end, this still is a wonderful experience and I dearly hope to be able to come back to this Aosta valley to attempt La Grivola once again.

Trajo Glacier

Aosta [course+hike 2]

Posted in Mountains, Statistics, University life on July 18, 2009 by xi'an

The course on Bayesian data analysis for ecologists is a very interesting moment in that all twentysome participants are there with a specific goal. These young researchers came to the course with their ecology, biology or genetic problem(s) and this really helps in focussing their approach to the general theory. This also makes for exciting intermission talks, with insufficient time to discuss in depth with all! Actually, this made me think of a post-course workshop where we would process case studies with the same participants and get a proper assessment of the true impact of the course… The research topic of the participants to the course range from animal population studies, including capture-recapture experiments, to genetic analyses, with a strong emphasis on plylogenies.


The second day of the course actually was a break of sorts, in that we all hiked to an alpine meadow at 2220m that was like a balcony to the Gran Paradiso peaks. The walk was quite enjoyable as, first, we climbed rather quickly to a high route where the view was terrific and, second, this was the prefect oportunity to get to know more about the participants’ interests and fields of study. Discussing the convergence of MCMC algorithms sitting on a granit slab at 2220m is certainly very close to my ultimate teaching goal! One additional appeal for the hike was to look for ibexes and chamois but, despite the help of a Park ranger, we saw none. However, on the way back, we saw a fox, an animal I have rarely seen in the past. The fact that it was so easily seen means that it lost its natural fear of humans and has unfortunately gotten used to been fed by Park visitors (despite signs explicitely warning against this).