Archive for stochastic dominance

Un Diario del Año de les Tapes

Posted in pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , , , on November 29, 2025 by xi'an

On my five-day stay in Sevilla, I got reacquainted with the small streets of the old town, and those of Cadiz for a sunny if short day outing, both that share some similarity with Venice!, as long as the bank of the Guadalquivir where I sent training every early morning. (If missing the opportunity of running the Cádiz media maratón with a last-minute bib a colleague of my wife had procured, due to a lack of favourable public transportation. I was further unsure running even relaxedly a half-marathon a week before my favourite half-marathon was such a great idea.)) I also enjoyed very much the company of the young statisticians and operation researchers at 5SYSORM, including discussions with Antía Enríquez about network scale-up methods, which I had heard off previously (but wondered at the matter of over-counting), alas missing the plenary by Rosa E. Lillo, and with  Adam Olivares about modelling stochastically ordered pairs through a mixture representation (above) I was unaware off. (Also kudos to all participants for sticking to inglés while I was the only participant not fluent in español!)

Foodwise, I (over?!) sampled half-a dozen tapas bars, both in Sevilla and Cádiz, with varying returns, but came back with several great experiences and the (obvious) rule that back-street, low-key, bars, with some pensioner patrons, were to be preferred. Not much greens though, apart from a nearby farmers’ market! (Sorry for the catalán tapes, as I could not find an Andalusian word that sounded like Peste, but with a positive vibe!!!)

Unsurprisingly, I did not have much time to read, except in trains and planes, but finished Les Chaînes de Markov by Noham Selcer, whose title is more a pretence than a driving line. As the story painfully unrolls along the uninteresting couple issues of a (former) maths teacher and a French literature teacher. With artificial dialogues and an overall whining tone that gave me toothaches! There is no redeeming character in the novel, which furthermore describes rural places in Normandie in the worst possible terms. I also tried to complete reading Juice by Tom Winton, but I gave up on this post-apocalyptic story where the climate crisis does not stop radical eco-warriors to flight around the World to assassinate descendants of the powerful people who could have acted against climate change, a poor merge of Mad Max (Oz, of course!) and The Road . Very binary. It took me 50 pages to realise the story was set in the Northwestern part of Australia. And another 50 page to give up! I also started The Witcher Season 4 a few days before leaving, with a new actor William Hemsworth replacing Henry Cavill, “as charismatic as a bollard with a wig”, a season that proves a complete disaster, in part but only in part because the original story itself deteriorates at this point. I stopped early, once I reached a dreadful anime of the vampyre’s background.

minimax, maximin or plain

Posted in Kids, R with tags , , , on June 1, 2020 by xi'an

A simple riddle from The Riddler on choosing between the maximum between two minima of two throws of an N-face dice, the minimum between two maxima of two throws of an N-face dice, and a single throw. Since maximin is always less than minimax, second choice is always worse than first and a stochastic domination version of the argument shows the single throw should stand in the middle.

more concentration, everywhere

Posted in R, Statistics with tags , , , , , , , , , , on January 25, 2019 by xi'an

Although it may sound like an excessive notion of optimality, one can hope at obtaining an estimator δ of a unidimensional parameter θ that is always closer to θ that any other parameter. In distribution if not almost surely, meaning the cdf of (δ-θ) is steeper than for other estimators enjoying the same cdf at zero (for instance ½ to make them all median-unbiased). When I saw this question on X validated, I thought of the Cauchy location example, where there is no uniformly optimal estimator, albeit a large collection of unbiased ones. But a simulation experiment shows that the MLE does better than the competition. At least than three (above) four of them (since I tried the Pitman estimator via Christian Henning’s smoothmest R package). The differences to the MLE empirical cd make it clearer below (with tomato for a score correction, gold for the Pitman estimator, sienna for the 38% trimmed mean, and blue for the median):I wonder at a general theory along these lines. There is a vague similarity with Pitman nearness or closeness but without the paradoxes induced by this criterion. More in the spirit of stochastic dominance, which may be achievable for location invariant and mean unbiased estimators…