I complain about some stuff since I can.
The shootout is a joke so I’m glad people are treating it like a joke. Why does it exist? Apparently, cause Americans hate ties? But okay, it’s entertaining, whatever. But we can all agree a shootout win isn’t the same as a normal win. Is there any reason the NHL doesn’t award 3 points for a win and 2 points for a shootout win? That way you conserve the number of points in a game (good) and make a regulation win worthwhile. What’s the reason the NHL isn’t doing it?
A lot of people who look at NHL fancystats are doing it in a very unfancy way. Take PDO (on-ice save% + on-ice shooting% for your team). Someone (I’m assuming that someone was PDO) realised that over time, this should average out to 1.000. It does. So when it deviates from 1.000, people take it to be a measure of “luck” and predict “regression” will arrive and bring the number back to 1.000. But, the first is an overassumption (just because most streaks are indistinguishable patternwise from random behaviour doesn’t mean they are all random behaviour) and the second is gambler’s fallacy.
Geez Laura Miller, you’re writing an article about science and about debunking some dubious claims. You should not be writing stuff like this:
…a trait possessed by about 35 percent of the world’s population — and growing, since the gene determining it is dominant.
Whether the gene is dominant or not has nothing to do with whether its expression is growing or not. Come on! This is basic stuff, get it right.
I recently read a book on the philosophy of science (this one) where I learned a lot, but I was irritated by the book’s completely punting on the question of Occam’s Razor. (Occam’s Razor is the claim that the simplest viable explanation is the most likely to be correct). Basically all Godfrey-Smith says on the topic is that it’s strange, and that “simple” is hard to define. But to me it seems like the central question in science. Apart from its use for doing things like inferring phylogenetic relationships (principle of parsimony), and evaluating new theories (the ones with the least assumptions are the most promising) it also seems to me to be the assumption underlying all scientific claims. What do we mean when we say that we “know” something to be the case in science (or even not in science, as long as we go by observation)? It’s not that we can’t think of an alternative explanation for observations. When I look at a clock, I assume that the time is what the clock says it is. Whereas it could be a clock that moves backwards. I could have developed a weird optical aberration in my eye that changes numbers. I could be subconsciously doubling the numbers in my head and not realizing it. G_d could be playing a trick on me. The clock could be relaying not only time information but time information and also information on the stock market and the next Canucks game. And yet I don’t even consider these possibilities. These are some pretty terrible examples, by the way, I realise that. I’ll try to think of better ones (help?). In any case, you could make a probability argument for some of these things, but not for all of them. So why do we do this? After all, the world is really complicated. Picking the simplest possible explanation presumably has a really terrible track record in science since everything is super not-simple. And yet it’s what we do and it somehow seems intuitively obvious to me that this is the correct way to go. Why?
By the way, Rudolf Carnap is an awesome name. I know this isn’t a complaint.