stdbrouw

Your experience also points to the limits of monitoring and subsequent behavioral change, though. I mean, yeah, it might prompt you to start your washing machine a bit earlier or a bit later to align with high production by your solar panels... but how much consumption can you really move around like that, and how many energy hogs can you just decide to not use? If you notice high energy use while cooking, are you going to start eating more salads instead? Across Europe electricity meters are being replaced by smart meters and people are really hyping up the advantages of being able to continuously monitor your energy usage, but I think the jury is still out whether it'll actually lead to significant energy savings.

Ultimately the biggest wins are when big appliances and heating/cooling respond to self-production or take advantage of times when electricity is cheap (if you're on a per-hour or per-day dynamic contract), whether that's with a simple timer like the one you installed, a relay that shuts down heating when you're cooking or something fancier like a Fronius Ohmpilot [1] that tweaks heating power to exactly match PV (over)production.

[1] https://www.fronius.com/en/solar-energy/installers-partners/...

Actually, no, people don't generally copy-paste parts of each others' literature review sections.

I mean, I guess you could call Claudine Gay a cargo culter because she used a particular method without fully understanding its assumptions, but if you set the bar that high then almost nothing is going to pass. If you've ever been in academia, you know Jose Ortega y Gasset was right when he wrote: "it is necessary to insist upon this extraordinary but undeniable fact: experimental science has progressed thanks in great part to the work of men astoundingly mediocre, and even less than mediocre. That is to say, modern science, the root and symbol of our actual civilization, finds a place for the intellectually commonplace man and allows him to work therein with success." Most of us, most of the time, work with "recipes" that we think will produce good science, and often that works, except when it doesn't.

As for Neri Oxman, the kind of word salad she spit out is definitely reminiscent of cargo cult behavior, but the motivation is different: if you're in a cult, you're a believer, whereas it's not clear to me that Oxman believes her own nonsense, and in any event she seems to tend to borrow stuff from other fields and does not presume to contribute to those fields. Consequently, her research doesn't really qualify as the kind of mimicking without understanding that Feynmen was talking about.

So I guess I must ask: is Sven Schnieders (the author) cargo culting about cargo cult science?

Thanks for taking the time to answer, fascinating!

I think it makes perfect sense. Modern EPDM or fiber-reinforced bitumen roofs are very unlikely to leak before it's time to replace or renovate them after 25 years or so. They provide ample space for solar panels and for a heat pump, all out of sight. No need for gutters. They are easier to insulate. You don't end up with a bunch of barely usable attic space you don't need (even worse if we're talking about a hip roof instead of a gable roof.) It just takes a bit of effort to not make them look like a cheap cardboard box, whereas a house with a nice sloped roof has instant appeal.

I dunno, I'm not a stickler for details. For example, I know that windows with many small panes ("muntins") are not cost-effective today and partially defeat the benefits of double and triple glazing, and the alternatives (fake bars that are glued on or between the glass) look shockingly bad... but any kind of partition at all (casement, fixed pane + tilt-turn pane, etc.) already adds visual interest and does not look out of place, as e.g. is evident in this photo of Georgian architecture with windows that I'm guessing are not historically accurate: https://upload.wikimedia.org/wikipedia/commons/7/78/Baggot_S...

Yeah, I think the building you show is a pretty good example of what Twu would like to see: one big mass, made pretty with consoles, balconies and cornices. You can do a more modern style too, though. I especially like Persian-inspired modern brickwork [1] [2].

What do you consider to be the defining features of Stalinist architecture though? The Wikipedia page portrays it as more or less "whatever happened to be built during Stalin's reign, whatever the style".

[1] https://www.yankodesign.com/2021/12/28/creative-brick-design...

[2] https://www.dezeen.com/2016/05/10/kahrizak-residential-housi...

At first glance, these look nice to me? Perhaps they look silly up close. I'm not sure what you mean by impractical though, I don't consider tall and narrow to be essential to the style, though many of the 18th and 19th century townhouses in the style surely were.

I feel we should bring back Georgian architecture [1]: on the one hand it is quite plain and so it fits well with modern sensibilities (and modern budgets that don't allow for a lot of faff, and the lack of skilled labor), but because of its strong emphasis on symmetry and a modest amount of ornament on doors, windows and railings, it looks vastly better than the lime rendered boxes of today. It also looks great with flat roofs, and flat roofs are here to stay -- why bother with nice roof tiles when you have to cover them with solar panels anyway.

[1] https://en.wikipedia.org/wiki/Georgian_architecture

Ironically, we're also losing a lot of efficiency because cities are demanding setbacks, bumps, differences in height etc. for bigger buildings to avoid ending up with drab rectangular blocks, but that leads to more heat loss due to more surface exposed to the environment, higher use of materials and a higher risk of water ingress because there's a ton of weak points where different surfaces connect. (It also just doesn't look very good.)

Ornament might be an alternative: just make massed buildings look nice instead of demanding that they are broken up into a bunch of lego blocks.

A good thread by Alfred Twu here: https://twitter.com/alfred_twu/status/1715230266531475702

I also wonder if the trend away from ornament has something to do with the trend towards fancier materials and bolder colors -- wasserstrich bricks, wood facades, decorative plasters in bathrooms, highly polished pigmented concrete floors, aluminium window frames in unusual colors, etc. People need their interestingness fix :-)

Just a priori this does not seem very likely to me. If you have budget that needs spending, it is key to spend it inconspicuously because as soon as a higher-up finds out that you are just wasting it to maintain your yearly allocation, they will still conclude that you need less next year (and that perhaps you ought to be fired.)

But geothermal heat pumps are interesting in cold climates because air-air and air-water heat pumps (the most common and affordable types) are less efficient below freezing, so I feel like that confirms rather than dispels the myth!

Also wonderful in cheese, adds a little umami but also just looks great: https://static.colruytgroup.com/images/500x500/std.lang.all/...

A lot of stuff has to be reported at source these days, e.g. in Belgium where I live, Airbnb reports your income from hosting to the revenue service, your bank reports how much you paid on your mortgage (deductible), if you're unemployed then the benefits agency will communicate that, so will child care (deductible), and so on. And as the parent said, you can always add/edit amounts as necessary, which indeed will often be necessary for things like equity.

> Fractional factorial is also most effective when the covariates are orthogonal and independent, which is rarely the case. We can get around this by projecting high dimensional covariates into lower dimensional space using PCA, which guarantees orthogonality, but this also seems to not be done so much.

I guess you're right in that if there were very many interactions between variables, then the aliasing of a fractional design would be an incredible nuisance, but on the other hand if there were no dependence between any of the variables at all then there would be no point to any sort of factorial design as you could just test each variable sequentially.

> Just wondering if anyone is using fractional factorial designs in real life? (or optimal designs like D-optimal designs)

I have to admit, I have never ever seen anyone calculate a D-optimal or G-optimal design for anything -- that said, I only graduated as a statistician 6 or 7 years ago. As you say, there might be some use to it for continuous variables but for factors, given that latin squares (etc.) are known to be optimal so you can just grab that off the shelf.

Let's say you're testing different formulations for cake but because you don't have time to do the full cartesian product of all possible variations over sugar content x fat content x type of flour, you just pick a handful, and it turns out that every time you're testing a high-sugar variation you are also testing spelt flour -- that's an alias because if you really like these particular cakes there's no way to know whether it was the spelt or the sugar that did it.

What experimental design brings to the table is a disciplined way to figure out what variables will be aliased and to make sure that these aliases are mostly harmless, either with the help of subject matter knowledge (you can taste sugar so it's fine if it's aliased with other things) or mathematically (let's try to avoid aliasing sugar content with fat content, but instead alias sugar content with the interaction effect of sugar, fat and flour type because that higher-order interaction is unlikely to matter over and above the first-order effects.)

Fractional designs in particular are typically used in agriculture and industrial settings, places where you want to try to optimize over very many factors at once but cannot afford to test every variation. It is not common in web analytics because (1) we usually assume that one particular change to a site or app will be independent of another change elsewhere, so there is no need to test them concurrently to see if a particular combo stands out and (2) if we did want to test combinations of variables concurrently, there's usually enough users or visitors to just test the entire grid and not worry about picking a selection of variations.

The tradeoff is that while there are many estimators of the expected value that are more efficient (have less variance) than the sample mean for various distributions, they typically introduce bias.

If every bar reflects the same proportion of the population then that would mean every bar of the histogram would have exactly the same height, because the y axis on a histogram is density or proportion.

I don't get how you can be so confident about a concept you only vaguely remember from a class you took long ago, and that even after reading the relevant article on wikipedia you conclude that therefore wikipedia must be wrong rather than thinking "hmm, wonder whether I'm misremembering things here..."

> What are some well-known distributions that appear to have outliers? A log-normal distribution maybe?

All of them. Which is why outside of a handful of contexts, the consensus in statistical modeling nowadays seems to be not to worry about outliers unless the values are completely unreasonable or there are a suspicious amount of them.

As to your bootstrapping idea, why not bootstrap the entire distribution, why only the tails? If you only bootstrap the tails but allow draws from the entire empirical distribution then you are changing the underlying distribution.

I googled around a bit for legal cases and for advice aimed at Dutch companies, but because the law is so recent there's not much. My hunch (as someone who is perfectly clueless about legal matters) is that a restaurant or a shop might find it hard to argue that replacing one full-time worker with two part-time workers is impossible (given that the work is already organized into shifts) but perhaps companies that employ knowledge workers could successfully argue that you can't hire two part-time coders to do the work of one full-timer because the additional coordination costs in terms of additional meetings, additional equipment (i.e. laptop), perhaps the need for additional management personnel, and so on, are an unreasonable demand on the company. I could also see demands to reduce work to 4/5 be rejected when few people at work share your job description, because finding another employee willing to work 1/5 is really hard.

There's also the power dynamic to think of: are you really going to ask your boss to change your hours if you think that he/she will react badly and only grudgingly allow it?

Or maybe I'm completely off base... but anyhow, I guess I just wanted to point out that it's not clear to me whether this law has teeth.

I mean, I don't mind, you can attribute whatever to whatever and yeah, a crib full of plush toys doesn't seem like such a smart idea... but the question is what a doctor should do, how we want them to behave in the face of limited evidence.

Not an expert but it seems to me you'd be making the same kind of error that the article denounces: attaching equal value to direct evidence on the one hand and inference to the best explanation on the other. If a physician does not observe tell-tale signs of suffocation in an infant, then it is not their role to say "Well, statistically speaking, or logically speaking, pretty good chance it's accidental suffocation isn't it? I'll jot this down as Accidental Suffocation Syndrome" or alternatively "I know in my heart that this is accidental suffocation but let's just call it SIDS for the benefit of the parents" but rather they should simply conclude "there's not enough evidence, therefore this death remains unexplained".

Don't know about the pods, but for a while we used DIY dishwashing powder with soda crystals as the main ingredient, and after a while you could definitely notice slight scratching of glassware, and any printing would fade pretty heavily. Not sure if that's a chemical process or simply the abrasiveness of the powder.

Glass is a little bit better at keeping oxygen out than plastic, dark glass keeps UV rays at bay, and thick glass bottles can better withstand the pressure of a second fermentation in the bottle, as with champagne and abbey beers. It also looks nice. These are advantages for beer and wine, but mostly irrelevant for soda, water and juices.

Scepticism is great, but that's perhaps not how I would describe praising a quote from an Ig Nobel prize winner that discards the entire field of personality psychology just because it happens to align with what you already believe.

The thesis of the article seems to be that "the world has changed" so that Wirecutter and similarly rigorous product review sites are less appealing than they used to be but... I don't see it? Products now are mostly "good enough" so that careful research is not necessary, and ten years ago that wasn't true? People now are interested in personal recommendations via social media and YouTube, and ten years ago that didn't happen? People now impulse buy, and ten years ago they didn't? There are so many reviews now that it's a commodity, as if ten years ago that wasn't exactly the problem that Wirecutter hoped to solve by promising to be the only product recommendation you'd need to read?

I think Wirecutter has always been a somewhat niche offering that appeals to a slightly neurotic, slightly nerdy audience that doesn't want to be fooled by marketing – like most of us here on HN, I guess. If they can keep up the quality of the reviews, I think they'll be around for a long time.

The paper you link does not mention anything about the need to try every potential treatment (and no treatment) on every person in a study, and it does not mention anything about any "inherent bias" in RCTs -- these were the two claims you made -- so I'm not quite sure why you bring it up. Given that you have brought it up: I like Deaton's work a lot and know the arguments well, but you have to take into account that in this paper they are trying to provide some counterweight to a hype cycle that suddenly led a bunch of social scientists to argue that observational studies are stupid and should just be gotten rid of entirely and replaced by randomized trials, which for many reasons is not a good idea. That hardly implies that we can never know anything about whether psychotherapy works until we can simulate the brain, that's just a totally bizarre conclusion to draw from that paper.

That's not how it works, though.

First, if it is not possible to blind the patient to the intervention and/or provide a meaningful placebo (which also happens with surgery and drugs with prominent side-effects) that's not a big deal, it simply changes the interpretation from "the effect of a particular regime of psychotherapy as such" to "the effect of a particular regime of psychotherapy, including expectation effects". Those expectation effects also exist in real life, if you go to a therapist you really really hope you'll get better and may be motivated to make a few changes in your life independently of the treatment, so if anything the trial becomes more realistic. Note that you'd still use a control arm so that if people get better over time without treatment, you're not exaggerating the effects of the intervention.

Second, almost every randomized controlled trial compares an experimental group against a control group, you are assigned to a group and don't move between them, and what that means is that we can only compare whether the treatment works on average, not whether it worked for one person in particular because, as you say, we don't know what would have happened (counterfactually) if that one specific person had been randomly assigned to the other group. But again, this hardly matters, because we can still know how well a treatment works on average. What you refer to, an experiment in which every subject receives every treatment, is known as a crossover design and while that design has its advantages, it's not that common because it is harder to pull off and while it can be more efficient (need fewer patients to attain an estimate with a given precision) it is not "less biased" or inherently more trustworthy.

The point is that we can verify whether something works (with a randomized controlled trial) independently of knowing why it works, and psychotherapy interventions are commonly evaluated with RCTs.

It took almost a century before we figured out how photo negatives worked chemically. Didn’t stop anyone from taking photographs though. Mechanistic explanations are great but not a requirement for good science.